From patchwork Tue Dec 28 20:11:22 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Sunil Pandey X-Patchwork-Id: 1573820 Return-Path: X-Original-To: incoming@patchwork.ozlabs.org Delivered-To: patchwork-incoming@bilbo.ozlabs.org Authentication-Results: bilbo.ozlabs.org; dkim=pass (1024-bit key; secure) header.d=sourceware.org header.i=@sourceware.org header.a=rsa-sha256 header.s=default header.b=YHmLzcbm; dkim-atps=neutral Authentication-Results: ozlabs.org; spf=pass (sender SPF authorized) smtp.mailfrom=sourceware.org (client-ip=2620:52:3:1:0:246e:9693:128c; helo=sourceware.org; envelope-from=libc-alpha-bounces+incoming=patchwork.ozlabs.org@sourceware.org; receiver=) Received: from sourceware.org (server2.sourceware.org [IPv6:2620:52:3:1:0:246e:9693:128c]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (4096 bits) server-digest SHA256) (No client certificate requested) by bilbo.ozlabs.org (Postfix) with ESMTPS id 4JNmBh46DHz9sVq for ; Wed, 29 Dec 2021 07:21:00 +1100 (AEDT) Received: from server2.sourceware.org (localhost [IPv6:::1]) by sourceware.org (Postfix) with ESMTP id 7349F3858417 for ; Tue, 28 Dec 2021 20:20:58 +0000 (GMT) DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 7349F3858417 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=sourceware.org; s=default; t=1640722858; bh=3f+z+8TDmwZSVFX2xUiz5F8KIec2zl4C4FEUn55bsXM=; h=To:Subject:Date:In-Reply-To:References:List-Id:List-Unsubscribe: List-Archive:List-Post:List-Help:List-Subscribe:From:Reply-To:Cc: From; b=YHmLzcbm6TpwG6CSqSSw8NBXL7z/gXBFkD4OtV3Q4qXFs3anyHJmHFRiuVdhzVrcE ULa6ZZrJ+AqcjwzB1em+5+WP6132+jewP9AQdSaeOk8Mx1vYf/80viNFvEJhq55/No kUGlinorEbo9QraRdW86N3ZYQML8mCKdL8LGzI6g= X-Original-To: libc-alpha@sourceware.org Delivered-To: libc-alpha@sourceware.org Received: from mga09.intel.com (mga09.intel.com [134.134.136.24]) by sourceware.org (Postfix) with ESMTPS id CF6A0385842A for ; Tue, 28 Dec 2021 20:11:35 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.1 sourceware.org CF6A0385842A X-IronPort-AV: E=McAfee;i="6200,9189,10211"; a="241218099" X-IronPort-AV: E=Sophos;i="5.88,242,1635231600"; d="scan'208";a="241218099" Received: from fmsmga001.fm.intel.com ([10.253.24.23]) by orsmga102.jf.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 28 Dec 2021 12:11:32 -0800 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.88,242,1635231600"; d="scan'208";a="666095348" Received: from scymds01.sc.intel.com ([10.148.94.138]) by fmsmga001.fm.intel.com with ESMTP; 28 Dec 2021 12:11:32 -0800 Received: from gskx-1.sc.intel.com (gskx-1.sc.intel.com [172.25.149.211]) by scymds01.sc.intel.com with ESMTP id 1BSKBUse016522; Tue, 28 Dec 2021 12:11:31 -0800 To: libc-alpha@sourceware.org Subject: [PATCH v4 10/18] x86-64: Add vector atan2/atan2f implementation to libmvec Date: Tue, 28 Dec 2021 12:11:22 -0800 Message-Id: <20211228201130.737370-11-skpgkp2@gmail.com> X-Mailer: git-send-email 2.31.1 In-Reply-To: <20211228201130.737370-1-skpgkp2@gmail.com> References: <20211228201130.737370-1-skpgkp2@gmail.com> MIME-Version: 1.0 X-Spam-Status: No, score=-7.1 required=5.0 tests=BAYES_00, DKIM_ADSP_CUSTOM_MED, FORGED_GMAIL_RCVD, FREEMAIL_ENVFROM_END_DIGIT, FREEMAIL_FROM, GIT_PATCH_0, HK_RANDOM_ENVFROM, HK_RANDOM_FROM, KAM_DMARC_NONE, KAM_DMARC_STATUS, KAM_SHORT, KAM_STOCKGEN, NML_ADSP_CUSTOM_MED, RCVD_IN_MSPIKE_H3, RCVD_IN_MSPIKE_WL, SPF_HELO_NONE, SPF_SOFTFAIL, TXREP autolearn=ham autolearn_force=no version=3.4.4 X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on server2.sourceware.org X-BeenThere: libc-alpha@sourceware.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Libc-alpha mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-Patchwork-Original-From: Sunil K Pandey via Libc-alpha From: Sunil Pandey Reply-To: Sunil K Pandey Cc: andrey.kolesov@intel.com, marius.cornea@intel.com Errors-To: libc-alpha-bounces+incoming=patchwork.ozlabs.org@sourceware.org Sender: "Libc-alpha" Implement vectorized atan2/atan2f containing SSE, AVX, AVX2 and AVX512 versions for libmvec as per vector ABI. It also contains accuracy and ABI tests for vector atan2/atan2f with regenerated ulps. --- bits/libm-simd-decl-stubs.h | 11 + math/bits/mathcalls.h | 2 +- .../unix/sysv/linux/x86_64/libmvec.abilist | 8 + sysdeps/x86/fpu/bits/math-vector.h | 4 + .../x86/fpu/finclude/math-vector-fortran.h | 4 + sysdeps/x86_64/fpu/Makeconfig | 1 + sysdeps/x86_64/fpu/Versions | 2 + sysdeps/x86_64/fpu/libm-test-ulps | 20 + .../fpu/multiarch/svml_d_atan22_core-sse2.S | 20 + .../x86_64/fpu/multiarch/svml_d_atan22_core.c | 28 ++ .../fpu/multiarch/svml_d_atan22_core_sse4.S | 471 +++++++++++++++++ .../fpu/multiarch/svml_d_atan24_core-sse.S | 20 + .../x86_64/fpu/multiarch/svml_d_atan24_core.c | 28 ++ .../fpu/multiarch/svml_d_atan24_core_avx2.S | 451 +++++++++++++++++ .../fpu/multiarch/svml_d_atan28_core-avx2.S | 20 + .../x86_64/fpu/multiarch/svml_d_atan28_core.c | 28 ++ .../fpu/multiarch/svml_d_atan28_core_avx512.S | 475 ++++++++++++++++++ .../fpu/multiarch/svml_s_atan2f16_core-avx2.S | 20 + .../fpu/multiarch/svml_s_atan2f16_core.c | 28 ++ .../multiarch/svml_s_atan2f16_core_avx512.S | 399 +++++++++++++++ .../fpu/multiarch/svml_s_atan2f4_core-sse2.S | 20 + .../fpu/multiarch/svml_s_atan2f4_core.c | 28 ++ .../fpu/multiarch/svml_s_atan2f4_core_sse4.S | 384 ++++++++++++++ .../fpu/multiarch/svml_s_atan2f8_core-sse.S | 20 + .../fpu/multiarch/svml_s_atan2f8_core.c | 28 ++ .../fpu/multiarch/svml_s_atan2f8_core_avx2.S | 362 +++++++++++++ sysdeps/x86_64/fpu/svml_d_atan22_core.S | 29 ++ sysdeps/x86_64/fpu/svml_d_atan24_core.S | 29 ++ sysdeps/x86_64/fpu/svml_d_atan24_core_avx.S | 25 + sysdeps/x86_64/fpu/svml_d_atan28_core.S | 25 + sysdeps/x86_64/fpu/svml_s_atan2f16_core.S | 25 + sysdeps/x86_64/fpu/svml_s_atan2f4_core.S | 29 ++ sysdeps/x86_64/fpu/svml_s_atan2f8_core.S | 29 ++ sysdeps/x86_64/fpu/svml_s_atan2f8_core_avx.S | 25 + .../fpu/test-double-libmvec-atan2-avx.c | 1 + .../fpu/test-double-libmvec-atan2-avx2.c | 1 + .../fpu/test-double-libmvec-atan2-avx512f.c | 1 + .../x86_64/fpu/test-double-libmvec-atan2.c | 3 + .../x86_64/fpu/test-double-vlen2-wrappers.c | 1 + .../fpu/test-double-vlen4-avx2-wrappers.c | 1 + .../x86_64/fpu/test-double-vlen4-wrappers.c | 1 + .../x86_64/fpu/test-double-vlen8-wrappers.c | 1 + .../fpu/test-float-libmvec-atan2f-avx.c | 1 + .../fpu/test-float-libmvec-atan2f-avx2.c | 1 + .../fpu/test-float-libmvec-atan2f-avx512f.c | 1 + .../x86_64/fpu/test-float-libmvec-atan2f.c | 3 + .../x86_64/fpu/test-float-vlen16-wrappers.c | 1 + .../x86_64/fpu/test-float-vlen4-wrappers.c | 1 + .../fpu/test-float-vlen8-avx2-wrappers.c | 1 + .../x86_64/fpu/test-float-vlen8-wrappers.c | 1 + 50 files changed, 3117 insertions(+), 1 deletion(-) create mode 100644 sysdeps/x86_64/fpu/multiarch/svml_d_atan22_core-sse2.S create mode 100644 sysdeps/x86_64/fpu/multiarch/svml_d_atan22_core.c create mode 100644 sysdeps/x86_64/fpu/multiarch/svml_d_atan22_core_sse4.S create mode 100644 sysdeps/x86_64/fpu/multiarch/svml_d_atan24_core-sse.S create mode 100644 sysdeps/x86_64/fpu/multiarch/svml_d_atan24_core.c create mode 100644 sysdeps/x86_64/fpu/multiarch/svml_d_atan24_core_avx2.S create mode 100644 sysdeps/x86_64/fpu/multiarch/svml_d_atan28_core-avx2.S create mode 100644 sysdeps/x86_64/fpu/multiarch/svml_d_atan28_core.c create mode 100644 sysdeps/x86_64/fpu/multiarch/svml_d_atan28_core_avx512.S create mode 100644 sysdeps/x86_64/fpu/multiarch/svml_s_atan2f16_core-avx2.S create mode 100644 sysdeps/x86_64/fpu/multiarch/svml_s_atan2f16_core.c create mode 100644 sysdeps/x86_64/fpu/multiarch/svml_s_atan2f16_core_avx512.S create mode 100644 sysdeps/x86_64/fpu/multiarch/svml_s_atan2f4_core-sse2.S create mode 100644 sysdeps/x86_64/fpu/multiarch/svml_s_atan2f4_core.c create mode 100644 sysdeps/x86_64/fpu/multiarch/svml_s_atan2f4_core_sse4.S create mode 100644 sysdeps/x86_64/fpu/multiarch/svml_s_atan2f8_core-sse.S create mode 100644 sysdeps/x86_64/fpu/multiarch/svml_s_atan2f8_core.c create mode 100644 sysdeps/x86_64/fpu/multiarch/svml_s_atan2f8_core_avx2.S create mode 100644 sysdeps/x86_64/fpu/svml_d_atan22_core.S create mode 100644 sysdeps/x86_64/fpu/svml_d_atan24_core.S create mode 100644 sysdeps/x86_64/fpu/svml_d_atan24_core_avx.S create mode 100644 sysdeps/x86_64/fpu/svml_d_atan28_core.S create mode 100644 sysdeps/x86_64/fpu/svml_s_atan2f16_core.S create mode 100644 sysdeps/x86_64/fpu/svml_s_atan2f4_core.S create mode 100644 sysdeps/x86_64/fpu/svml_s_atan2f8_core.S create mode 100644 sysdeps/x86_64/fpu/svml_s_atan2f8_core_avx.S create mode 100644 sysdeps/x86_64/fpu/test-double-libmvec-atan2-avx.c create mode 100644 sysdeps/x86_64/fpu/test-double-libmvec-atan2-avx2.c create mode 100644 sysdeps/x86_64/fpu/test-double-libmvec-atan2-avx512f.c create mode 100644 sysdeps/x86_64/fpu/test-double-libmvec-atan2.c create mode 100644 sysdeps/x86_64/fpu/test-float-libmvec-atan2f-avx.c create mode 100644 sysdeps/x86_64/fpu/test-float-libmvec-atan2f-avx2.c create mode 100644 sysdeps/x86_64/fpu/test-float-libmvec-atan2f-avx512f.c create mode 100644 sysdeps/x86_64/fpu/test-float-libmvec-atan2f.c diff --git a/bits/libm-simd-decl-stubs.h b/bits/libm-simd-decl-stubs.h index 7f1304ed1d..31878bf4ed 100644 --- a/bits/libm-simd-decl-stubs.h +++ b/bits/libm-simd-decl-stubs.h @@ -208,4 +208,15 @@ #define __DECL_SIMD_cbrtf32x #define __DECL_SIMD_cbrtf64x #define __DECL_SIMD_cbrtf128x + +#define __DECL_SIMD_atan2 +#define __DECL_SIMD_atan2f +#define __DECL_SIMD_atan2l +#define __DECL_SIMD_atan2f16 +#define __DECL_SIMD_atan2f32 +#define __DECL_SIMD_atan2f64 +#define __DECL_SIMD_atan2f128 +#define __DECL_SIMD_atan2f32x +#define __DECL_SIMD_atan2f64x +#define __DECL_SIMD_atan2f128x #endif diff --git a/math/bits/mathcalls.h b/math/bits/mathcalls.h index 26d18f0135..1bd4911993 100644 --- a/math/bits/mathcalls.h +++ b/math/bits/mathcalls.h @@ -56,7 +56,7 @@ __MATHCALL_VEC (asin,, (_Mdouble_ __x)); /* Arc tangent of X. */ __MATHCALL_VEC (atan,, (_Mdouble_ __x)); /* Arc tangent of Y/X. */ -__MATHCALL (atan2,, (_Mdouble_ __y, _Mdouble_ __x)); +__MATHCALL_VEC (atan2,, (_Mdouble_ __y, _Mdouble_ __x)); /* Cosine of X. */ __MATHCALL_VEC (cos,, (_Mdouble_ __x)); diff --git a/sysdeps/unix/sysv/linux/x86_64/libmvec.abilist b/sysdeps/unix/sysv/linux/x86_64/libmvec.abilist index a6558d9810..2b3b8d3886 100644 --- a/sysdeps/unix/sysv/linux/x86_64/libmvec.abilist +++ b/sysdeps/unix/sysv/linux/x86_64/libmvec.abilist @@ -55,6 +55,7 @@ GLIBC_2.35 _ZGVbN2v_exp10 F GLIBC_2.35 _ZGVbN2v_exp2 F GLIBC_2.35 _ZGVbN2v_expm1 F GLIBC_2.35 _ZGVbN2v_sinh F +GLIBC_2.35 _ZGVbN2vv_atan2 F GLIBC_2.35 _ZGVbN2vv_hypot F GLIBC_2.35 _ZGVbN4v_acosf F GLIBC_2.35 _ZGVbN4v_asinf F @@ -65,6 +66,7 @@ GLIBC_2.35 _ZGVbN4v_exp10f F GLIBC_2.35 _ZGVbN4v_exp2f F GLIBC_2.35 _ZGVbN4v_expm1f F GLIBC_2.35 _ZGVbN4v_sinhf F +GLIBC_2.35 _ZGVbN4vv_atan2f F GLIBC_2.35 _ZGVbN4vv_hypotf F GLIBC_2.35 _ZGVcN4v_acos F GLIBC_2.35 _ZGVcN4v_asin F @@ -75,6 +77,7 @@ GLIBC_2.35 _ZGVcN4v_exp10 F GLIBC_2.35 _ZGVcN4v_exp2 F GLIBC_2.35 _ZGVcN4v_expm1 F GLIBC_2.35 _ZGVcN4v_sinh F +GLIBC_2.35 _ZGVcN4vv_atan2 F GLIBC_2.35 _ZGVcN4vv_hypot F GLIBC_2.35 _ZGVcN8v_acosf F GLIBC_2.35 _ZGVcN8v_asinf F @@ -85,6 +88,7 @@ GLIBC_2.35 _ZGVcN8v_exp10f F GLIBC_2.35 _ZGVcN8v_exp2f F GLIBC_2.35 _ZGVcN8v_expm1f F GLIBC_2.35 _ZGVcN8v_sinhf F +GLIBC_2.35 _ZGVcN8vv_atan2f F GLIBC_2.35 _ZGVcN8vv_hypotf F GLIBC_2.35 _ZGVdN4v_acos F GLIBC_2.35 _ZGVdN4v_asin F @@ -95,6 +99,7 @@ GLIBC_2.35 _ZGVdN4v_exp10 F GLIBC_2.35 _ZGVdN4v_exp2 F GLIBC_2.35 _ZGVdN4v_expm1 F GLIBC_2.35 _ZGVdN4v_sinh F +GLIBC_2.35 _ZGVdN4vv_atan2 F GLIBC_2.35 _ZGVdN4vv_hypot F GLIBC_2.35 _ZGVdN8v_acosf F GLIBC_2.35 _ZGVdN8v_asinf F @@ -105,6 +110,7 @@ GLIBC_2.35 _ZGVdN8v_exp10f F GLIBC_2.35 _ZGVdN8v_exp2f F GLIBC_2.35 _ZGVdN8v_expm1f F GLIBC_2.35 _ZGVdN8v_sinhf F +GLIBC_2.35 _ZGVdN8vv_atan2f F GLIBC_2.35 _ZGVdN8vv_hypotf F GLIBC_2.35 _ZGVeN16v_acosf F GLIBC_2.35 _ZGVeN16v_asinf F @@ -115,6 +121,7 @@ GLIBC_2.35 _ZGVeN16v_exp10f F GLIBC_2.35 _ZGVeN16v_exp2f F GLIBC_2.35 _ZGVeN16v_expm1f F GLIBC_2.35 _ZGVeN16v_sinhf F +GLIBC_2.35 _ZGVeN16vv_atan2f F GLIBC_2.35 _ZGVeN16vv_hypotf F GLIBC_2.35 _ZGVeN8v_acos F GLIBC_2.35 _ZGVeN8v_asin F @@ -125,4 +132,5 @@ GLIBC_2.35 _ZGVeN8v_exp10 F GLIBC_2.35 _ZGVeN8v_exp2 F GLIBC_2.35 _ZGVeN8v_expm1 F GLIBC_2.35 _ZGVeN8v_sinh F +GLIBC_2.35 _ZGVeN8vv_atan2 F GLIBC_2.35 _ZGVeN8vv_hypot F diff --git a/sysdeps/x86/fpu/bits/math-vector.h b/sysdeps/x86/fpu/bits/math-vector.h index dcd45934ab..62f2890ab3 100644 --- a/sysdeps/x86/fpu/bits/math-vector.h +++ b/sysdeps/x86/fpu/bits/math-vector.h @@ -98,6 +98,10 @@ # define __DECL_SIMD_cbrt __DECL_SIMD_x86_64 # undef __DECL_SIMD_cbrtf # define __DECL_SIMD_cbrtf __DECL_SIMD_x86_64 +# undef __DECL_SIMD_atan2 +# define __DECL_SIMD_atan2 __DECL_SIMD_x86_64 +# undef __DECL_SIMD_atan2f +# define __DECL_SIMD_atan2f __DECL_SIMD_x86_64 # endif #endif diff --git a/sysdeps/x86/fpu/finclude/math-vector-fortran.h b/sysdeps/x86/fpu/finclude/math-vector-fortran.h index dfb5f13ea3..2269b74d50 100644 --- a/sysdeps/x86/fpu/finclude/math-vector-fortran.h +++ b/sysdeps/x86/fpu/finclude/math-vector-fortran.h @@ -48,6 +48,8 @@ !GCC$ builtin (sinhf) attributes simd (notinbranch) if('x86_64') !GCC$ builtin (cbrt) attributes simd (notinbranch) if('x86_64') !GCC$ builtin (cbrtf) attributes simd (notinbranch) if('x86_64') +!GCC$ builtin (atan2) attributes simd (notinbranch) if('x86_64') +!GCC$ builtin (atan2f) attributes simd (notinbranch) if('x86_64') !GCC$ builtin (cos) attributes simd (notinbranch) if('x32') !GCC$ builtin (cosf) attributes simd (notinbranch) if('x32') @@ -81,3 +83,5 @@ !GCC$ builtin (sinhf) attributes simd (notinbranch) if('x32') !GCC$ builtin (cbrt) attributes simd (notinbranch) if('x32') !GCC$ builtin (cbrtf) attributes simd (notinbranch) if('x32') +!GCC$ builtin (atan2) attributes simd (notinbranch) if('x32') +!GCC$ builtin (atan2f) attributes simd (notinbranch) if('x32') diff --git a/sysdeps/x86_64/fpu/Makeconfig b/sysdeps/x86_64/fpu/Makeconfig index dde737c0d6..96a40856fa 100644 --- a/sysdeps/x86_64/fpu/Makeconfig +++ b/sysdeps/x86_64/fpu/Makeconfig @@ -25,6 +25,7 @@ libmvec-funcs = \ acos \ asin \ atan \ + atan2 \ cbrt \ cos \ cosh \ diff --git a/sysdeps/x86_64/fpu/Versions b/sysdeps/x86_64/fpu/Versions index b70aeb3e2f..f58c98eb45 100644 --- a/sysdeps/x86_64/fpu/Versions +++ b/sysdeps/x86_64/fpu/Versions @@ -23,6 +23,7 @@ libmvec { _ZGVbN2v_exp2; _ZGVcN4v_exp2; _ZGVdN4v_exp2; _ZGVeN8v_exp2; _ZGVbN2v_expm1; _ZGVcN4v_expm1; _ZGVdN4v_expm1; _ZGVeN8v_expm1; _ZGVbN2v_sinh; _ZGVcN4v_sinh; _ZGVdN4v_sinh; _ZGVeN8v_sinh; + _ZGVbN2vv_atan2; _ZGVcN4vv_atan2; _ZGVdN4vv_atan2; _ZGVeN8vv_atan2; _ZGVbN2vv_hypot; _ZGVcN4vv_hypot; _ZGVdN4vv_hypot; _ZGVeN8vv_hypot; _ZGVbN4v_acosf; _ZGVcN8v_acosf; _ZGVdN8v_acosf; _ZGVeN16v_acosf; _ZGVbN4v_asinf; _ZGVcN8v_asinf; _ZGVdN8v_asinf; _ZGVeN16v_asinf; @@ -33,6 +34,7 @@ libmvec { _ZGVbN4v_exp2f; _ZGVcN8v_exp2f; _ZGVdN8v_exp2f; _ZGVeN16v_exp2f; _ZGVbN4v_expm1f; _ZGVcN8v_expm1f; _ZGVdN8v_expm1f; _ZGVeN16v_expm1f; _ZGVbN4v_sinhf; _ZGVcN8v_sinhf; _ZGVdN8v_sinhf; _ZGVeN16v_sinhf; + _ZGVbN4vv_atan2f; _ZGVcN8vv_atan2f; _ZGVdN8vv_atan2f; _ZGVeN16vv_atan2f; _ZGVbN4vv_hypotf; _ZGVcN8vv_hypotf; _ZGVdN8vv_hypotf; _ZGVeN16vv_hypotf; } } diff --git a/sysdeps/x86_64/fpu/libm-test-ulps b/sysdeps/x86_64/fpu/libm-test-ulps index e039a993df..6f59c61756 100644 --- a/sysdeps/x86_64/fpu/libm-test-ulps +++ b/sysdeps/x86_64/fpu/libm-test-ulps @@ -166,6 +166,26 @@ float: 2 float128: 2 ldouble: 1 +Function: "atan2_vlen16": +float: 2 + +Function: "atan2_vlen2": +double: 1 + +Function: "atan2_vlen4": +double: 1 +float: 2 + +Function: "atan2_vlen4_avx2": +double: 1 + +Function: "atan2_vlen8": +double: 1 +float: 2 + +Function: "atan2_vlen8_avx2": +float: 2 + Function: "atan_downward": double: 1 float: 2 diff --git a/sysdeps/x86_64/fpu/multiarch/svml_d_atan22_core-sse2.S b/sysdeps/x86_64/fpu/multiarch/svml_d_atan22_core-sse2.S new file mode 100644 index 0000000000..6c3ad05a6c --- /dev/null +++ b/sysdeps/x86_64/fpu/multiarch/svml_d_atan22_core-sse2.S @@ -0,0 +1,20 @@ +/* SSE2 version of vectorized atan2. + Copyright (C) 2021 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, see + . */ + +#define _ZGVbN2vv_atan2 _ZGVbN2vv_atan2_sse2 +#include "../svml_d_atan22_core.S" diff --git a/sysdeps/x86_64/fpu/multiarch/svml_d_atan22_core.c b/sysdeps/x86_64/fpu/multiarch/svml_d_atan22_core.c new file mode 100644 index 0000000000..43f1ee7f33 --- /dev/null +++ b/sysdeps/x86_64/fpu/multiarch/svml_d_atan22_core.c @@ -0,0 +1,28 @@ +/* Multiple versions of vectorized atan2, vector length is 2. + Copyright (C) 2021 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, see + . */ + +#define SYMBOL_NAME _ZGVbN2vv_atan2 +#include "ifunc-mathvec-sse4_1.h" + +libc_ifunc_redirected (REDIRECT_NAME, SYMBOL_NAME, IFUNC_SELECTOR ()); + +#ifdef SHARED +__hidden_ver1 (_ZGVbN2vv_atan2, __GI__ZGVbN2vv_atan2, + __redirect__ZGVbN2vv_atan2) + __attribute__ ((visibility ("hidden"))); +#endif diff --git a/sysdeps/x86_64/fpu/multiarch/svml_d_atan22_core_sse4.S b/sysdeps/x86_64/fpu/multiarch/svml_d_atan22_core_sse4.S new file mode 100644 index 0000000000..5c0d0fd17f --- /dev/null +++ b/sysdeps/x86_64/fpu/multiarch/svml_d_atan22_core_sse4.S @@ -0,0 +1,471 @@ +/* Function atan2 vectorized with SSE4. + Copyright (C) 2021 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, see + https://www.gnu.org/licenses/. */ + +/* + * ALGORITHM DESCRIPTION: + * For 0.0 <= x <= 7.0/16.0: atan(x) = atan(0.0) + atan(s), where s=(x-0.0)/(1.0+0.0*x) + * For 7.0/16.0 <= x <= 11.0/16.0: atan(x) = atan(0.5) + atan(s), where s=(x-0.5)/(1.0+0.5*x) + * For 11.0/16.0 <= x <= 19.0/16.0: atan(x) = atan(1.0) + atan(s), where s=(x-1.0)/(1.0+1.0*x) + * For 19.0/16.0 <= x <= 39.0/16.0: atan(x) = atan(1.5) + atan(s), where s=(x-1.5)/(1.0+1.5*x) + * For 39.0/16.0 <= x <= inf : atan(x) = atan(inf) + atan(s), where s=-1.0/x + * Where atan(s) ~= s+s^3*Poly11(s^2) on interval |s|<7.0/0.16. + * + * + */ + +/* Offsets for data table __svml_datan2_data_internal + */ +#define dPI 0 +#define dPIO2 16 +#define dA19 32 +#define dA18 48 +#define dA17 64 +#define dA16 80 +#define dA15 96 +#define dA14 112 +#define dA13 128 +#define dA12 144 +#define dA11 160 +#define dA10 176 +#define dA09 192 +#define dA08 208 +#define dA07 224 +#define dA06 240 +#define dA05 256 +#define dA04 272 +#define dA03 288 +#define dA02 304 +#define dA01 320 +#define dA00 336 +#define dSIGN_MASK 352 +#define iCHK_WORK_SUB 368 +#define iCHK_WORK_CMP 384 +#define dABS_MASK 400 +#define dZERO 416 + +#include + + .text + .section .text.sse4,"ax",@progbits +ENTRY(_ZGVbN2vv_atan2_sse4) + subq $88, %rsp + cfi_def_cfa_offset(96) + movaps %xmm0, %xmm8 + +/* + * #define NO_VECTOR_ZERO_ATAN2_ARGS + * Declarations + * Variables + * Constants + * The end of declarations + * Implementation + * Get r0~=1/B + * Cannot be replaced by VQRCP(D, dR0, dB); + * Argument Absolute values + */ + movups dABS_MASK+__svml_datan2_data_internal(%rip), %xmm4 + movaps %xmm1, %xmm9 + movaps %xmm4, %xmm1 + andps %xmm8, %xmm4 + andps %xmm9, %xmm1 + movaps %xmm4, %xmm2 + cmpnltpd %xmm1, %xmm2 + +/* Argument signs */ + movups dSIGN_MASK+__svml_datan2_data_internal(%rip), %xmm3 + movaps %xmm2, %xmm0 + movups dPIO2+__svml_datan2_data_internal(%rip), %xmm5 + movaps %xmm3, %xmm7 + movaps %xmm3, %xmm6 + +/* + * 1) If yx then a=-x, b=y, PIO2=Pi/2 + */ + orps %xmm1, %xmm3 + movaps %xmm2, %xmm10 + andps %xmm2, %xmm5 + andnps %xmm4, %xmm0 + andps %xmm2, %xmm3 + andnps %xmm1, %xmm10 + andps %xmm4, %xmm2 + orps %xmm3, %xmm0 + orps %xmm2, %xmm10 + divpd %xmm10, %xmm0 + movq iCHK_WORK_SUB+__svml_datan2_data_internal(%rip), %xmm11 + +/* if x<0, dPI = Pi, else dPI =0 */ + movaps %xmm9, %xmm3 + +/* Check if y and x are on main path. */ + pshufd $221, %xmm1, %xmm12 + andps %xmm9, %xmm7 + psubd %xmm11, %xmm12 + andps %xmm8, %xmm6 + movq iCHK_WORK_CMP+__svml_datan2_data_internal(%rip), %xmm13 + xorl %edx, %edx + movups %xmm4, 16(%rsp) + xorl %eax, %eax + pshufd $221, %xmm4, %xmm14 + movdqa %xmm12, %xmm4 + pcmpgtd %xmm13, %xmm4 + pcmpeqd %xmm13, %xmm12 + por %xmm12, %xmm4 + +/* Polynomial. */ + movaps %xmm0, %xmm12 + mulpd %xmm0, %xmm12 + cmplepd dZERO+__svml_datan2_data_internal(%rip), %xmm3 + psubd %xmm11, %xmm14 + movdqa %xmm14, %xmm15 + pcmpeqd %xmm13, %xmm14 + pcmpgtd %xmm13, %xmm15 + por %xmm14, %xmm15 + movaps %xmm12, %xmm14 + mulpd %xmm12, %xmm14 + por %xmm15, %xmm4 + movaps %xmm14, %xmm15 + mulpd %xmm14, %xmm15 + movmskps %xmm4, %ecx + movups %xmm10, (%rsp) + movups dA19+__svml_datan2_data_internal(%rip), %xmm10 + mulpd %xmm15, %xmm10 + movups dA18+__svml_datan2_data_internal(%rip), %xmm13 + movups dA17+__svml_datan2_data_internal(%rip), %xmm11 + addpd dA15+__svml_datan2_data_internal(%rip), %xmm10 + mulpd %xmm15, %xmm13 + mulpd %xmm15, %xmm11 + mulpd %xmm15, %xmm10 + addpd dA14+__svml_datan2_data_internal(%rip), %xmm13 + addpd dA13+__svml_datan2_data_internal(%rip), %xmm11 + addpd dA11+__svml_datan2_data_internal(%rip), %xmm10 + mulpd %xmm15, %xmm13 + mulpd %xmm15, %xmm11 + mulpd %xmm15, %xmm10 + addpd dA10+__svml_datan2_data_internal(%rip), %xmm13 + addpd dA09+__svml_datan2_data_internal(%rip), %xmm11 + addpd dA07+__svml_datan2_data_internal(%rip), %xmm10 + mulpd %xmm15, %xmm13 + mulpd %xmm15, %xmm11 + mulpd %xmm15, %xmm10 + addpd dA06+__svml_datan2_data_internal(%rip), %xmm13 + addpd dA05+__svml_datan2_data_internal(%rip), %xmm11 + addpd dA03+__svml_datan2_data_internal(%rip), %xmm10 + mulpd %xmm15, %xmm13 + mulpd %xmm15, %xmm11 + mulpd %xmm12, %xmm10 + addpd dA02+__svml_datan2_data_internal(%rip), %xmm13 + addpd dA01+__svml_datan2_data_internal(%rip), %xmm11 + addpd %xmm10, %xmm13 + mulpd %xmm11, %xmm12 + mulpd %xmm13, %xmm14 + movups dA16+__svml_datan2_data_internal(%rip), %xmm2 + mulpd %xmm15, %xmm2 + addpd dA12+__svml_datan2_data_internal(%rip), %xmm2 + mulpd %xmm15, %xmm2 + addpd dA08+__svml_datan2_data_internal(%rip), %xmm2 + mulpd %xmm15, %xmm2 + addpd dA04+__svml_datan2_data_internal(%rip), %xmm2 + +/* A00=1.0, account for it later VQFMA(D, dP4, dP4, dR8, dA00); */ + mulpd %xmm2, %xmm15 + addpd %xmm12, %xmm15 + addpd %xmm14, %xmm15 + +/* + * Reconstruction. + * dP=(R+R*dP) + dPIO2 + */ + mulpd %xmm0, %xmm15 + addpd %xmm15, %xmm0 + addpd %xmm5, %xmm0 + andps __svml_datan2_data_internal(%rip), %xmm3 + orps %xmm7, %xmm0 + addpd %xmm3, %xmm0 + +/* Special branch for fast (vector) processing of zero arguments */ + movups 16(%rsp), %xmm11 + orps %xmm6, %xmm0 + testb $3, %cl + +/* Go to auxilary branch */ + jne L(AUX_BRANCH) + # LOE rbx rbp r12 r13 r14 r15 eax edx xmm0 xmm1 xmm3 xmm4 xmm5 xmm6 xmm7 xmm8 xmm9 xmm11 + +/* Return from auxilary branch + * for out of main path inputs + */ + +L(AUX_BRANCH_RETURN): +/* + * Special branch for fast (vector) processing of zero arguments + * The end of implementation + */ + testl %edx, %edx + +/* Go to special inputs processing branch */ + jne L(SPECIAL_VALUES_BRANCH) + # LOE rbx rbp r12 r13 r14 r15 eax edx xmm0 xmm8 xmm9 + +/* Restore registers + * and exit the function + */ + +L(EXIT): + addq $88, %rsp + cfi_def_cfa_offset(8) + ret + cfi_def_cfa_offset(96) + +/* Branch to process + * special inputs + */ + +L(SPECIAL_VALUES_BRANCH): + movups %xmm8, 32(%rsp) + movups %xmm9, 48(%rsp) + movups %xmm0, 64(%rsp) + # LOE rbx rbp r12 r13 r14 r15 eax edx xmm0 + + movq %r12, 16(%rsp) + cfi_offset(12, -80) + movl %eax, %r12d + movq %r13, 8(%rsp) + cfi_offset(13, -88) + movl %edx, %r13d + movq %r14, (%rsp) + cfi_offset(14, -96) + # LOE rbx rbp r15 r12d r13d + +/* Range mask + * bits check + */ + +L(RANGEMASK_CHECK): + btl %r12d, %r13d + +/* Call scalar math function */ + jc L(SCALAR_MATH_CALL) + # LOE rbx rbp r15 r12d r13d + +/* Special inputs + * processing loop + */ + +L(SPECIAL_VALUES_LOOP): + incl %r12d + cmpl $2, %r12d + +/* Check bits in range mask */ + jl L(RANGEMASK_CHECK) + # LOE rbx rbp r15 r12d r13d + + movq 16(%rsp), %r12 + cfi_restore(12) + movq 8(%rsp), %r13 + cfi_restore(13) + movq (%rsp), %r14 + cfi_restore(14) + movups 64(%rsp), %xmm0 + +/* Go to exit */ + jmp L(EXIT) + cfi_offset(12, -80) + cfi_offset(13, -88) + cfi_offset(14, -96) + # LOE rbx rbp r12 r13 r14 r15 xmm0 + +/* Scalar math fucntion call + * to process special input + */ + +L(SCALAR_MATH_CALL): + movl %r12d, %r14d + movsd 32(%rsp,%r14,8), %xmm0 + movsd 48(%rsp,%r14,8), %xmm1 + call atan2@PLT + # LOE rbx rbp r14 r15 r12d r13d xmm0 + + movsd %xmm0, 64(%rsp,%r14,8) + +/* Process special inputs in loop */ + jmp L(SPECIAL_VALUES_LOOP) + cfi_restore(12) + cfi_restore(13) + cfi_restore(14) + # LOE rbx rbp r15 r12d r13d + +/* Auxilary branch + * for out of main path inputs + */ + +L(AUX_BRANCH): +/* Check if at least on of Y or Y is zero: iAXAYZERO */ + movups dZERO+__svml_datan2_data_internal(%rip), %xmm2 + +/* Check if both X & Y are not NaNs: iXYnotNAN */ + movaps %xmm9, %xmm12 + movaps %xmm8, %xmm10 + cmpordpd %xmm9, %xmm12 + cmpordpd %xmm8, %xmm10 + cmpeqpd %xmm2, %xmm1 + cmpeqpd %xmm2, %xmm11 + andps %xmm10, %xmm12 + orps %xmm11, %xmm1 + pshufd $221, %xmm1, %xmm1 + pshufd $221, %xmm12, %xmm11 + +/* Check if at least on of Y or Y is zero and not NaN: iAXAYZEROnotNAN */ + pand %xmm11, %xmm1 + +/* Exclude from previous callout mask zero (and not NaN) arguments */ + movdqa %xmm1, %xmm13 + pandn %xmm4, %xmm13 + +/* + * Path for zero arguments (at least one of both) + * Check if both args are zeros (den. is zero) + */ + movups (%rsp), %xmm4 + cmpeqpd %xmm2, %xmm4 + +/* Go to callout */ + movmskps %xmm13, %edx + +/* Set sPIO2 to zero if den. is zero */ + movaps %xmm4, %xmm15 + andps %xmm2, %xmm4 + andnps %xmm5, %xmm15 + andl $3, %edx + orps %xmm4, %xmm15 + pshufd $221, %xmm9, %xmm5 + orps %xmm7, %xmm15 + +/* Res = sign(Y)*(X<0)?(PIO2+PI):PIO2 */ + pshufd $221, %xmm2, %xmm7 + pcmpgtd %xmm5, %xmm7 + pshufd $80, %xmm7, %xmm14 + andps %xmm3, %xmm14 + addpd %xmm14, %xmm15 + +/* Merge results from main and spec path */ + pshufd $80, %xmm1, %xmm3 + orps %xmm6, %xmm15 + movdqa %xmm3, %xmm6 + andps %xmm3, %xmm15 + andnps %xmm0, %xmm6 + movaps %xmm6, %xmm0 + orps %xmm15, %xmm0 + +/* Return to main vector processing path */ + jmp L(AUX_BRANCH_RETURN) + # LOE rbx rbp r12 r13 r14 r15 eax edx xmm0 xmm8 xmm9 +END(_ZGVbN2vv_atan2_sse4) + + .section .rodata, "a" + .align 16 + +#ifdef __svml_datan2_data_internal_typedef +typedef unsigned int VUINT32; +typedef struct { + __declspec(align(16)) VUINT32 dPI[2][2]; + __declspec(align(16)) VUINT32 dPIO2[2][2]; + __declspec(align(16)) VUINT32 dA19[2][2]; + __declspec(align(16)) VUINT32 dA18[2][2]; + __declspec(align(16)) VUINT32 dA17[2][2]; + __declspec(align(16)) VUINT32 dA16[2][2]; + __declspec(align(16)) VUINT32 dA15[2][2]; + __declspec(align(16)) VUINT32 dA14[2][2]; + __declspec(align(16)) VUINT32 dA13[2][2]; + __declspec(align(16)) VUINT32 dA12[2][2]; + __declspec(align(16)) VUINT32 dA11[2][2]; + __declspec(align(16)) VUINT32 dA10[2][2]; + __declspec(align(16)) VUINT32 dA09[2][2]; + __declspec(align(16)) VUINT32 dA08[2][2]; + __declspec(align(16)) VUINT32 dA07[2][2]; + __declspec(align(16)) VUINT32 dA06[2][2]; + __declspec(align(16)) VUINT32 dA05[2][2]; + __declspec(align(16)) VUINT32 dA04[2][2]; + __declspec(align(16)) VUINT32 dA03[2][2]; + __declspec(align(16)) VUINT32 dA02[2][2]; + __declspec(align(16)) VUINT32 dA01[2][2]; + __declspec(align(16)) VUINT32 dA00[2][2]; + __declspec(align(16)) VUINT32 dSIGN_MASK[2][2]; + __declspec(align(16)) VUINT32 iCHK_WORK_SUB[4][1]; + __declspec(align(16)) VUINT32 iCHK_WORK_CMP[4][1]; + __declspec(align(16)) VUINT32 dABS_MASK[2][2]; + __declspec(align(16)) VUINT32 dZERO[2][2]; +} __svml_datan2_data_internal; +#endif +__svml_datan2_data_internal: + .quad 0x400921FB54442D18, 0x400921FB54442D18 //dPI + .align 16 + .quad 0x3FF921FB54442D18, 0x3FF921FB54442D18 //dPIO2 + .align 16 + .quad 0xBEF4FDB537ABC7A3, 0xBEF4FDB537ABC7A3 // dA19 + .align 16 + .quad 0x3F2CED0A36665209, 0x3F2CED0A36665209 // dA18 + .align 16 + .quad 0xBF52E67C93954C23, 0xBF52E67C93954C23 // dA17 + .align 16 + .quad 0x3F6F5A1DAE82AFB3, 0x3F6F5A1DAE82AFB3 // dA16 + .align 16 + .quad 0xBF82B2EC618E4BAD, 0xBF82B2EC618E4BAD // dA15 + .align 16 + .quad 0x3F914F4C661116A5, 0x3F914F4C661116A5 // dA14 + .align 16 + .quad 0xBF9A5E83B081F69C, 0xBF9A5E83B081F69C // dA13 + .align 16 + .quad 0x3FA169980CB6AD4F, 0x3FA169980CB6AD4F // dA12 + .align 16 + .quad 0xBFA4EFA2E563C1BC, 0xBFA4EFA2E563C1BC // dA11 + .align 16 + .quad 0x3FA7EC0FBC50683B, 0x3FA7EC0FBC50683B // dA10 + .align 16 + .quad 0xBFAAD261EAA09954, 0xBFAAD261EAA09954 // dA09 + .align 16 + .quad 0x3FAE1749BD612DCF, 0x3FAE1749BD612DCF // dA08 + .align 16 + .quad 0xBFB11084009435E0, 0xBFB11084009435E0 // dA07 + .align 16 + .quad 0x3FB3B12A49295651, 0x3FB3B12A49295651 // dA06 + .align 16 + .quad 0xBFB745D009BADA94, 0xBFB745D009BADA94 // dA05 + .align 16 + .quad 0x3FBC71C707F7D5B5, 0x3FBC71C707F7D5B5 // dA04 + .align 16 + .quad 0xBFC2492491EE55C7, 0xBFC2492491EE55C7 // dA03 + .align 16 + .quad 0x3FC999999997EE34, 0x3FC999999997EE34 // dA02 + .align 16 + .quad 0xBFD55555555553C5, 0xBFD55555555553C5 // dA01 + .align 16 + .quad 0x3FF0000000000000, 0x3FF0000000000000 // dA00 + .align 16 + .quad 0x8000000000000000, 0x8000000000000000 //dSIGN_MASK + .align 16 + .long 0x80300000, 0x80300000, 0x80300000, 0x80300000 //iCHK_WORK_SUB + .align 16 + .long 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000 //iCHK_WORK_CMP + .align 16 + .quad 0x7fffffffffffffff, 0x7fffffffffffffff //dABS_MASK + .align 16 + .quad 0x0000000000000000, 0x0000000000000000 //dZERO + .align 16 + .type __svml_datan2_data_internal,@object + .size __svml_datan2_data_internal,.-__svml_datan2_data_internal diff --git a/sysdeps/x86_64/fpu/multiarch/svml_d_atan24_core-sse.S b/sysdeps/x86_64/fpu/multiarch/svml_d_atan24_core-sse.S new file mode 100644 index 0000000000..0db843a088 --- /dev/null +++ b/sysdeps/x86_64/fpu/multiarch/svml_d_atan24_core-sse.S @@ -0,0 +1,20 @@ +/* SSE version of vectorized atan2. + Copyright (C) 2021 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, see + . */ + +#define _ZGVdN4vv_atan2 _ZGVdN4vv_atan2_sse_wrapper +#include "../svml_d_atan24_core.S" diff --git a/sysdeps/x86_64/fpu/multiarch/svml_d_atan24_core.c b/sysdeps/x86_64/fpu/multiarch/svml_d_atan24_core.c new file mode 100644 index 0000000000..c2e2611584 --- /dev/null +++ b/sysdeps/x86_64/fpu/multiarch/svml_d_atan24_core.c @@ -0,0 +1,28 @@ +/* Multiple versions of vectorized atan2, vector length is 4. + Copyright (C) 2021 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, see + . */ + +#define SYMBOL_NAME _ZGVdN4vv_atan2 +#include "ifunc-mathvec-avx2.h" + +libc_ifunc_redirected (REDIRECT_NAME, SYMBOL_NAME, IFUNC_SELECTOR ()); + +#ifdef SHARED +__hidden_ver1 (_ZGVdN4vv_atan2, __GI__ZGVdN4vv_atan2, + __redirect__ZGVdN4vv_atan2) + __attribute__ ((visibility ("hidden"))); +#endif diff --git a/sysdeps/x86_64/fpu/multiarch/svml_d_atan24_core_avx2.S b/sysdeps/x86_64/fpu/multiarch/svml_d_atan24_core_avx2.S new file mode 100644 index 0000000000..cdf780715b --- /dev/null +++ b/sysdeps/x86_64/fpu/multiarch/svml_d_atan24_core_avx2.S @@ -0,0 +1,451 @@ +/* Function atan2 vectorized with AVX2. + Copyright (C) 2021 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, see + https://www.gnu.org/licenses/. */ + +/* + * ALGORITHM DESCRIPTION: + * For 0.0 <= x <= 7.0/16.0: atan(x) = atan(0.0) + atan(s), where s=(x-0.0)/(1.0+0.0*x) + * For 7.0/16.0 <= x <= 11.0/16.0: atan(x) = atan(0.5) + atan(s), where s=(x-0.5)/(1.0+0.5*x) + * For 11.0/16.0 <= x <= 19.0/16.0: atan(x) = atan(1.0) + atan(s), where s=(x-1.0)/(1.0+1.0*x) + * For 19.0/16.0 <= x <= 39.0/16.0: atan(x) = atan(1.5) + atan(s), where s=(x-1.5)/(1.0+1.5*x) + * For 39.0/16.0 <= x <= inf : atan(x) = atan(inf) + atan(s), where s=-1.0/x + * Where atan(s) ~= s+s^3*Poly11(s^2) on interval |s|<7.0/0.16. + * + * + */ + +/* Offsets for data table __svml_datan2_data_internal + */ +#define dPI 0 +#define dPIO2 32 +#define dA19 64 +#define dA18 96 +#define dA17 128 +#define dA16 160 +#define dA15 192 +#define dA14 224 +#define dA13 256 +#define dA12 288 +#define dA11 320 +#define dA10 352 +#define dA09 384 +#define dA08 416 +#define dA07 448 +#define dA06 480 +#define dA05 512 +#define dA04 544 +#define dA03 576 +#define dA02 608 +#define dA01 640 +#define dA00 672 +#define dSIGN_MASK 704 +#define iCHK_WORK_SUB 736 +#define iCHK_WORK_CMP 768 +#define dABS_MASK 800 +#define dZERO 832 + +#include + + .text + .section .text.avx2,"ax",@progbits +ENTRY(_ZGVdN4vv_atan2_avx2) + pushq %rbp + cfi_def_cfa_offset(16) + movq %rsp, %rbp + cfi_def_cfa(6, 16) + cfi_offset(6, -16) + andq $-32, %rsp + subq $128, %rsp + xorl %edx, %edx + +/* + * #define NO_VECTOR_ZERO_ATAN2_ARGS + * Declarations + * Variables + * Constants + * The end of declarations + * Implementation + * Get r0~=1/B + * Cannot be replaced by VQRCP(D, dR0, dB); + * Argument Absolute values + */ + vmovupd dABS_MASK+__svml_datan2_data_internal(%rip), %ymm5 + +/* Argument signs */ + vmovupd dSIGN_MASK+__svml_datan2_data_internal(%rip), %ymm4 + vmovups iCHK_WORK_SUB+__svml_datan2_data_internal(%rip), %xmm13 + vmovupd %ymm0, (%rsp) + vmovapd %ymm1, %ymm8 + vandpd %ymm5, %ymm8, %ymm2 + vandpd %ymm5, %ymm0, %ymm1 + vcmpnlt_uqpd %ymm2, %ymm1, %ymm15 + +/* + * 1) If yx then a=-x, b=y, PIO2=Pi/2 + */ + vorpd %ymm4, %ymm2, %ymm6 + vblendvpd %ymm15, %ymm6, %ymm1, %ymm3 + vblendvpd %ymm15, %ymm1, %ymm2, %ymm6 + vdivpd %ymm6, %ymm3, %ymm14 + vmovups iCHK_WORK_CMP+__svml_datan2_data_internal(%rip), %xmm3 + vmovupd %ymm6, 32(%rsp) + vandpd %ymm4, %ymm0, %ymm7 + vandpd %ymm4, %ymm8, %ymm5 + vandpd dPIO2+__svml_datan2_data_internal(%rip), %ymm15, %ymm4 + +/* Check if y and x are on main path. */ + vextractf128 $1, %ymm2, %xmm9 + vextractf128 $1, %ymm1, %xmm10 + vshufps $221, %xmm9, %xmm2, %xmm11 + vshufps $221, %xmm10, %xmm1, %xmm12 + vpsubd %xmm13, %xmm11, %xmm0 + vpsubd %xmm13, %xmm12, %xmm9 + vpcmpgtd %xmm3, %xmm0, %xmm15 + vpcmpeqd %xmm3, %xmm0, %xmm6 + vpcmpgtd %xmm3, %xmm9, %xmm10 + vpcmpeqd %xmm3, %xmm9, %xmm3 + vpor %xmm6, %xmm15, %xmm11 + vpor %xmm3, %xmm10, %xmm12 + +/* Polynomial. */ + vmulpd %ymm14, %ymm14, %ymm10 + vpor %xmm12, %xmm11, %xmm3 + vmovupd dA18+__svml_datan2_data_internal(%rip), %ymm9 + vmovupd dA17+__svml_datan2_data_internal(%rip), %ymm12 + vmovupd dA16+__svml_datan2_data_internal(%rip), %ymm15 + vmulpd %ymm10, %ymm10, %ymm11 + +/* if x<0, dPI = Pi, else dPI =0 */ + vcmple_oqpd dZERO+__svml_datan2_data_internal(%rip), %ymm8, %ymm13 + vmovmskps %xmm3, %eax + vmulpd %ymm11, %ymm11, %ymm0 + vandpd __svml_datan2_data_internal(%rip), %ymm13, %ymm6 + vmovupd dA19+__svml_datan2_data_internal(%rip), %ymm13 + vfmadd213pd dA14+__svml_datan2_data_internal(%rip), %ymm0, %ymm9 + vfmadd213pd dA13+__svml_datan2_data_internal(%rip), %ymm0, %ymm12 + vfmadd213pd dA12+__svml_datan2_data_internal(%rip), %ymm0, %ymm15 + vfmadd213pd dA15+__svml_datan2_data_internal(%rip), %ymm0, %ymm13 + vfmadd213pd dA10+__svml_datan2_data_internal(%rip), %ymm0, %ymm9 + vfmadd213pd dA09+__svml_datan2_data_internal(%rip), %ymm0, %ymm12 + vfmadd213pd dA08+__svml_datan2_data_internal(%rip), %ymm0, %ymm15 + vfmadd213pd dA11+__svml_datan2_data_internal(%rip), %ymm0, %ymm13 + vfmadd213pd dA06+__svml_datan2_data_internal(%rip), %ymm0, %ymm9 + vfmadd213pd dA05+__svml_datan2_data_internal(%rip), %ymm0, %ymm12 + vfmadd213pd dA04+__svml_datan2_data_internal(%rip), %ymm0, %ymm15 + vfmadd213pd dA07+__svml_datan2_data_internal(%rip), %ymm0, %ymm13 + vfmadd213pd dA02+__svml_datan2_data_internal(%rip), %ymm0, %ymm9 + vfmadd213pd dA01+__svml_datan2_data_internal(%rip), %ymm0, %ymm12 + vfmadd213pd dA03+__svml_datan2_data_internal(%rip), %ymm0, %ymm13 + +/* A00=1.0, account for it later VQFMA(D, dP4, dP4, dR8, dA00); */ + vmulpd %ymm15, %ymm0, %ymm0 + vfmadd213pd %ymm9, %ymm10, %ymm13 + vfmadd213pd %ymm0, %ymm10, %ymm12 + vfmadd213pd %ymm12, %ymm11, %ymm13 + +/* + * Reconstruction. + * dP=(R+R*dP) + dPIO2 + */ + vfmadd213pd %ymm14, %ymm14, %ymm13 + vaddpd %ymm13, %ymm4, %ymm14 + vorpd %ymm5, %ymm14, %ymm0 + vaddpd %ymm0, %ymm6, %ymm9 + vorpd %ymm7, %ymm9, %ymm0 + +/* Special branch for fast (vector) processing of zero arguments */ + testl %eax, %eax + +/* Go to auxilary branch */ + jne L(AUX_BRANCH) + # LOE rbx r12 r13 r14 r15 edx xmm3 ymm0 ymm1 ymm2 ymm4 ymm5 ymm6 ymm7 ymm8 + +/* Return from auxilary branch + * for out of main path inputs + */ + +L(AUX_BRANCH_RETURN): +/* + * Special branch for fast (vector) processing of zero arguments + * The end of implementation + */ + testl %edx, %edx + +/* Go to special inputs processing branch */ + jne L(SPECIAL_VALUES_BRANCH) + # LOE rbx r12 r13 r14 r15 edx ymm0 ymm8 + +/* Restore registers + * and exit the function + */ + +L(EXIT): + movq %rbp, %rsp + popq %rbp + cfi_def_cfa(7, 8) + cfi_restore(6) + ret + cfi_def_cfa(6, 16) + cfi_offset(6, -16) + +/* Branch to process + * special inputs + */ + +L(SPECIAL_VALUES_BRANCH): + vmovupd (%rsp), %ymm1 + vmovupd %ymm8, 64(%rsp) + vmovupd %ymm0, 96(%rsp) + vmovupd %ymm1, 32(%rsp) + # LOE rbx r12 r13 r14 r15 edx ymm0 + + xorl %eax, %eax + # LOE rbx r12 r13 r14 r15 eax edx + + vzeroupper + movq %r12, 16(%rsp) + /* DW_CFA_expression: r12 (r12) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -112; DW_OP_plus) */ + .cfi_escape 0x10, 0x0c, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x90, 0xff, 0xff, 0xff, 0x22 + movl %eax, %r12d + movq %r13, 8(%rsp) + /* DW_CFA_expression: r13 (r13) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -120; DW_OP_plus) */ + .cfi_escape 0x10, 0x0d, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x88, 0xff, 0xff, 0xff, 0x22 + movl %edx, %r13d + movq %r14, (%rsp) + /* DW_CFA_expression: r14 (r14) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -128; DW_OP_plus) */ + .cfi_escape 0x10, 0x0e, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x80, 0xff, 0xff, 0xff, 0x22 + # LOE rbx r15 r12d r13d + +/* Range mask + * bits check + */ + +L(RANGEMASK_CHECK): + btl %r12d, %r13d + +/* Call scalar math function */ + jc L(SCALAR_MATH_CALL) + # LOE rbx r15 r12d r13d + +/* Special inputs + * processing loop + */ + +L(SPECIAL_VALUES_LOOP): + incl %r12d + cmpl $4, %r12d + +/* Check bits in range mask */ + jl L(RANGEMASK_CHECK) + # LOE rbx r15 r12d r13d + + movq 16(%rsp), %r12 + cfi_restore(12) + movq 8(%rsp), %r13 + cfi_restore(13) + movq (%rsp), %r14 + cfi_restore(14) + vmovupd 96(%rsp), %ymm0 + +/* Go to exit */ + jmp L(EXIT) + /* DW_CFA_expression: r12 (r12) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -112; DW_OP_plus) */ + .cfi_escape 0x10, 0x0c, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x90, 0xff, 0xff, 0xff, 0x22 + /* DW_CFA_expression: r13 (r13) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -120; DW_OP_plus) */ + .cfi_escape 0x10, 0x0d, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x88, 0xff, 0xff, 0xff, 0x22 + /* DW_CFA_expression: r14 (r14) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -128; DW_OP_plus) */ + .cfi_escape 0x10, 0x0e, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x80, 0xff, 0xff, 0xff, 0x22 + # LOE rbx r12 r13 r14 r15 ymm0 + +/* Scalar math fucntion call + * to process special input + */ + +L(SCALAR_MATH_CALL): + movl %r12d, %r14d + movsd 32(%rsp,%r14,8), %xmm0 + movsd 64(%rsp,%r14,8), %xmm1 + call atan2@PLT + # LOE rbx r14 r15 r12d r13d xmm0 + + movsd %xmm0, 96(%rsp,%r14,8) + +/* Process special inputs in loop */ + jmp L(SPECIAL_VALUES_LOOP) + cfi_restore(12) + cfi_restore(13) + cfi_restore(14) + # LOE rbx r15 r12d r13d + +/* Auxilary branch + * for out of main path inputs + */ + +L(AUX_BRANCH): + vmovupd (%rsp), %ymm11 + +/* Check if at least on of Y or Y is zero: iAXAYZERO */ + vmovupd dZERO+__svml_datan2_data_internal(%rip), %ymm10 + +/* Check if both X & Y are not NaNs: iXYnotNAN */ + vcmpordpd %ymm8, %ymm8, %ymm12 + vcmpordpd %ymm11, %ymm11, %ymm13 + vcmpeqpd %ymm10, %ymm2, %ymm2 + vcmpeqpd %ymm10, %ymm1, %ymm1 + vandpd %ymm13, %ymm12, %ymm14 + vorpd %ymm1, %ymm2, %ymm2 + vextractf128 $1, %ymm14, %xmm15 + vextractf128 $1, %ymm2, %xmm11 + vshufps $221, %xmm15, %xmm14, %xmm9 + vshufps $221, %xmm11, %xmm2, %xmm12 + +/* + * Path for zero arguments (at least one of both) + * Check if both args are zeros (den. is zero) + */ + vcmpeqpd 32(%rsp), %ymm10, %ymm2 + +/* Check if at least on of Y or Y is zero and not NaN: iAXAYZEROnotNAN */ + vpand %xmm9, %xmm12, %xmm1 + +/* Exclude from previous callout mask zero (and not NaN) arguments */ + vpandn %xmm3, %xmm1, %xmm3 + +/* Go to callout */ + vmovmskps %xmm3, %edx + +/* Set sPIO2 to zero if den. is zero */ + vblendvpd %ymm2, %ymm10, %ymm4, %ymm4 + vorpd %ymm5, %ymm4, %ymm5 + +/* Res = sign(Y)*(X<0)?(PIO2+PI):PIO2 */ + vextractf128 $1, %ymm10, %xmm2 + vextractf128 $1, %ymm8, %xmm3 + vshufps $221, %xmm2, %xmm10, %xmm4 + vshufps $221, %xmm3, %xmm8, %xmm9 + vpcmpgtd %xmm9, %xmm4, %xmm12 + vpshufd $80, %xmm12, %xmm11 + vpshufd $250, %xmm12, %xmm13 + vinsertf128 $1, %xmm13, %ymm11, %ymm14 + vandpd %ymm6, %ymm14, %ymm6 + vaddpd %ymm6, %ymm5, %ymm2 + vorpd %ymm7, %ymm2, %ymm2 + +/* Merge results from main and spec path */ + vpshufd $80, %xmm1, %xmm7 + vpshufd $250, %xmm1, %xmm1 + vinsertf128 $1, %xmm1, %ymm7, %ymm3 + vblendvpd %ymm3, %ymm2, %ymm0, %ymm0 + +/* Return to main vector processing path */ + jmp L(AUX_BRANCH_RETURN) + # LOE rbx r12 r13 r14 r15 edx ymm0 ymm8 +END(_ZGVdN4vv_atan2_avx2) + + .section .rodata, "a" + .align 32 + +#ifdef __svml_datan2_data_internal_typedef +typedef unsigned int VUINT32; +typedef struct { + __declspec(align(32)) VUINT32 dPI[4][2]; + __declspec(align(32)) VUINT32 dPIO2[4][2]; + __declspec(align(32)) VUINT32 dA19[4][2]; + __declspec(align(32)) VUINT32 dA18[4][2]; + __declspec(align(32)) VUINT32 dA17[4][2]; + __declspec(align(32)) VUINT32 dA16[4][2]; + __declspec(align(32)) VUINT32 dA15[4][2]; + __declspec(align(32)) VUINT32 dA14[4][2]; + __declspec(align(32)) VUINT32 dA13[4][2]; + __declspec(align(32)) VUINT32 dA12[4][2]; + __declspec(align(32)) VUINT32 dA11[4][2]; + __declspec(align(32)) VUINT32 dA10[4][2]; + __declspec(align(32)) VUINT32 dA09[4][2]; + __declspec(align(32)) VUINT32 dA08[4][2]; + __declspec(align(32)) VUINT32 dA07[4][2]; + __declspec(align(32)) VUINT32 dA06[4][2]; + __declspec(align(32)) VUINT32 dA05[4][2]; + __declspec(align(32)) VUINT32 dA04[4][2]; + __declspec(align(32)) VUINT32 dA03[4][2]; + __declspec(align(32)) VUINT32 dA02[4][2]; + __declspec(align(32)) VUINT32 dA01[4][2]; + __declspec(align(32)) VUINT32 dA00[4][2]; + __declspec(align(32)) VUINT32 dSIGN_MASK[4][2]; + __declspec(align(32)) VUINT32 iCHK_WORK_SUB[8][1]; + __declspec(align(32)) VUINT32 iCHK_WORK_CMP[8][1]; + __declspec(align(32)) VUINT32 dABS_MASK[4][2]; + __declspec(align(32)) VUINT32 dZERO[4][2]; +} __svml_datan2_data_internal; +#endif +__svml_datan2_data_internal: + .quad 0x400921FB54442D18, 0x400921FB54442D18, 0x400921FB54442D18, 0x400921FB54442D18 //dPI + .align 32 + .quad 0x3FF921FB54442D18, 0x3FF921FB54442D18, 0x3FF921FB54442D18, 0x3FF921FB54442D18 //dPIO2 + .align 32 + .quad 0xBEF4FDB537ABC7A3, 0xBEF4FDB537ABC7A3, 0xBEF4FDB537ABC7A3, 0xBEF4FDB537ABC7A3 // dA19 + .align 32 + .quad 0x3F2CED0A36665209, 0x3F2CED0A36665209, 0x3F2CED0A36665209, 0x3F2CED0A36665209 // dA18 + .align 32 + .quad 0xBF52E67C93954C23, 0xBF52E67C93954C23, 0xBF52E67C93954C23, 0xBF52E67C93954C23 // dA17 + .align 32 + .quad 0x3F6F5A1DAE82AFB3, 0x3F6F5A1DAE82AFB3, 0x3F6F5A1DAE82AFB3, 0x3F6F5A1DAE82AFB3 // dA16 + .align 32 + .quad 0xBF82B2EC618E4BAD, 0xBF82B2EC618E4BAD, 0xBF82B2EC618E4BAD, 0xBF82B2EC618E4BAD // dA15 + .align 32 + .quad 0x3F914F4C661116A5, 0x3F914F4C661116A5, 0x3F914F4C661116A5, 0x3F914F4C661116A5 // dA14 + .align 32 + .quad 0xBF9A5E83B081F69C, 0xBF9A5E83B081F69C, 0xBF9A5E83B081F69C, 0xBF9A5E83B081F69C // dA13 + .align 32 + .quad 0x3FA169980CB6AD4F, 0x3FA169980CB6AD4F, 0x3FA169980CB6AD4F, 0x3FA169980CB6AD4F // dA12 + .align 32 + .quad 0xBFA4EFA2E563C1BC, 0xBFA4EFA2E563C1BC, 0xBFA4EFA2E563C1BC, 0xBFA4EFA2E563C1BC // dA11 + .align 32 + .quad 0x3FA7EC0FBC50683B, 0x3FA7EC0FBC50683B, 0x3FA7EC0FBC50683B, 0x3FA7EC0FBC50683B // dA10 + .align 32 + .quad 0xBFAAD261EAA09954, 0xBFAAD261EAA09954, 0xBFAAD261EAA09954, 0xBFAAD261EAA09954 // dA09 + .align 32 + .quad 0x3FAE1749BD612DCF, 0x3FAE1749BD612DCF, 0x3FAE1749BD612DCF, 0x3FAE1749BD612DCF // dA08 + .align 32 + .quad 0xBFB11084009435E0, 0xBFB11084009435E0, 0xBFB11084009435E0, 0xBFB11084009435E0 // dA07 + .align 32 + .quad 0x3FB3B12A49295651, 0x3FB3B12A49295651, 0x3FB3B12A49295651, 0x3FB3B12A49295651 // dA06 + .align 32 + .quad 0xBFB745D009BADA94, 0xBFB745D009BADA94, 0xBFB745D009BADA94, 0xBFB745D009BADA94 // dA05 + .align 32 + .quad 0x3FBC71C707F7D5B5, 0x3FBC71C707F7D5B5, 0x3FBC71C707F7D5B5, 0x3FBC71C707F7D5B5 // dA04 + .align 32 + .quad 0xBFC2492491EE55C7, 0xBFC2492491EE55C7, 0xBFC2492491EE55C7, 0xBFC2492491EE55C7 // dA03 + .align 32 + .quad 0x3FC999999997EE34, 0x3FC999999997EE34, 0x3FC999999997EE34, 0x3FC999999997EE34 // dA02 + .align 32 + .quad 0xBFD55555555553C5, 0xBFD55555555553C5, 0xBFD55555555553C5, 0xBFD55555555553C5 // dA01 + .align 32 + .quad 0x3FF0000000000000, 0x3FF0000000000000, 0x3FF0000000000000, 0x3FF0000000000000 // dA00 + .align 32 + .quad 0x8000000000000000, 0x8000000000000000, 0x8000000000000000, 0x8000000000000000 //dSIGN_MASK + .align 32 + .long 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000 //iCHK_WORK_SUB + .align 32 + .long 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000 //iCHK_WORK_CMP + .align 32 + .quad 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff //dABS_MASK + .align 32 + .quad 0x0000000000000000, 0x0000000000000000, 0x0000000000000000, 0x0000000000000000 //dZERO + .align 32 + .type __svml_datan2_data_internal,@object + .size __svml_datan2_data_internal,.-__svml_datan2_data_internal diff --git a/sysdeps/x86_64/fpu/multiarch/svml_d_atan28_core-avx2.S b/sysdeps/x86_64/fpu/multiarch/svml_d_atan28_core-avx2.S new file mode 100644 index 0000000000..a8d34a6143 --- /dev/null +++ b/sysdeps/x86_64/fpu/multiarch/svml_d_atan28_core-avx2.S @@ -0,0 +1,20 @@ +/* AVX2 version of vectorized atan2. + Copyright (C) 2021 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, see + . */ + +#define _ZGVeN8vv_atan2 _ZGVeN8vv_atan2_avx2_wrapper +#include "../svml_d_atan28_core.S" diff --git a/sysdeps/x86_64/fpu/multiarch/svml_d_atan28_core.c b/sysdeps/x86_64/fpu/multiarch/svml_d_atan28_core.c new file mode 100644 index 0000000000..a0897e9cf0 --- /dev/null +++ b/sysdeps/x86_64/fpu/multiarch/svml_d_atan28_core.c @@ -0,0 +1,28 @@ +/* Multiple versions of vectorized atan2, vector length is 8. + Copyright (C) 2021 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, see + . */ + +#define SYMBOL_NAME _ZGVeN8vv_atan2 +#include "ifunc-mathvec-avx512-skx.h" + +libc_ifunc_redirected (REDIRECT_NAME, SYMBOL_NAME, IFUNC_SELECTOR ()); + +#ifdef SHARED +__hidden_ver1 (_ZGVeN8vv_atan2, __GI__ZGVeN8vv_atan2, + __redirect__ZGVeN8vv_atan2) + __attribute__ ((visibility ("hidden"))); +#endif diff --git a/sysdeps/x86_64/fpu/multiarch/svml_d_atan28_core_avx512.S b/sysdeps/x86_64/fpu/multiarch/svml_d_atan28_core_avx512.S new file mode 100644 index 0000000000..6d18f5f757 --- /dev/null +++ b/sysdeps/x86_64/fpu/multiarch/svml_d_atan28_core_avx512.S @@ -0,0 +1,475 @@ +/* Function atan2 vectorized with AVX-512. + Copyright (C) 2021 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, see + https://www.gnu.org/licenses/. */ + +/* + * ALGORITHM DESCRIPTION: + * For 0.0 <= x <= 7.0/16.0: atan(x) = atan(0.0) + atan(s), where s=(x-0.0)/(1.0+0.0*x) + * For 7.0/16.0 <= x <= 11.0/16.0: atan(x) = atan(0.5) + atan(s), where s=(x-0.5)/(1.0+0.5*x) + * For 11.0/16.0 <= x <= 19.0/16.0: atan(x) = atan(1.0) + atan(s), where s=(x-1.0)/(1.0+1.0*x) + * For 19.0/16.0 <= x <= 39.0/16.0: atan(x) = atan(1.5) + atan(s), where s=(x-1.5)/(1.0+1.5*x) + * For 39.0/16.0 <= x <= inf : atan(x) = atan(inf) + atan(s), where s=-1.0/x + * Where atan(s) ~= s+s^3*Poly11(s^2) on interval |s|<7.0/0.16. + * + * + */ + +/* Offsets for data table __svml_datan2_data_internal + */ +#define dPI 0 +#define dPIO2 64 +#define dA19 128 +#define dA18 192 +#define dA17 256 +#define dA16 320 +#define dA15 384 +#define dA14 448 +#define dA13 512 +#define dA12 576 +#define dA11 640 +#define dA10 704 +#define dA09 768 +#define dA08 832 +#define dA07 896 +#define dA06 960 +#define dA05 1024 +#define dA04 1088 +#define dA03 1152 +#define dA02 1216 +#define dA01 1280 +#define dA00 1344 +#define dSIGN_MASK 1408 +#define iCHK_WORK_SUB 1472 +#define iCHK_WORK_CMP 1536 +#define dABS_MASK 1600 +#define dZERO 1664 + +#include + + .text + .section .text.evex512,"ax",@progbits +ENTRY(_ZGVeN8vv_atan2_skx) + pushq %rbp + cfi_def_cfa_offset(16) + movq %rsp, %rbp + cfi_def_cfa(6, 16) + cfi_offset(6, -16) + andq $-64, %rsp + subq $256, %rsp + xorl %edx, %edx + +/* + * #define NO_VECTOR_ZERO_ATAN2_ARGS + * Declarations + * Variables + * Constants + * The end of declarations + * Implementation + * Get r0~=1/B + * Cannot be replaced by VQRCP(D, dR0, dB); + * Argument Absolute values + */ + vmovups dABS_MASK+__svml_datan2_data_internal(%rip), %zmm4 + +/* Argument signs */ + vmovups dSIGN_MASK+__svml_datan2_data_internal(%rip), %zmm6 + +/* + * 1) If yx then a=-x, b=y, PIO2=Pi/2 + */ + vmovups dPIO2+__svml_datan2_data_internal(%rip), %zmm3 + vandpd %zmm4, %zmm0, %zmm11 + vmovaps %zmm1, %zmm7 + vandpd %zmm4, %zmm7, %zmm2 + vandpd %zmm6, %zmm7, %zmm5 + vandpd %zmm6, %zmm0, %zmm4 + vorpd %zmm6, %zmm2, %zmm12 + vcmppd $17, {sae}, %zmm2, %zmm11, %k1 + vmovdqu iCHK_WORK_CMP+__svml_datan2_data_internal(%rip), %ymm6 + vmovups %zmm11, 64(%rsp) + +/* Check if y and x are on main path. */ + vpsrlq $32, %zmm2, %zmm9 + vblendmpd %zmm11, %zmm12, %zmm13{%k1} + vblendmpd %zmm2, %zmm11, %zmm15{%k1} + vpsrlq $32, %zmm11, %zmm8 + vmovdqu iCHK_WORK_SUB+__svml_datan2_data_internal(%rip), %ymm12 + vdivpd {rn-sae}, %zmm15, %zmm13, %zmm1 + vmovups %zmm15, (%rsp) + vpmovqd %zmm9, %ymm14 + vpmovqd %zmm8, %ymm10 + vxorpd %zmm3, %zmm3, %zmm3{%k1} + vpsubd %ymm12, %ymm14, %ymm13 + vpsubd %ymm12, %ymm10, %ymm9 + +/* Polynomial. */ + vmulpd {rn-sae}, %zmm1, %zmm1, %zmm12 + vpcmpgtd %ymm6, %ymm13, %ymm15 + vpcmpeqd %ymm6, %ymm13, %ymm11 + vmulpd {rn-sae}, %zmm12, %zmm12, %zmm13 + vpor %ymm11, %ymm15, %ymm8 + vmovups dA19+__svml_datan2_data_internal(%rip), %zmm11 + vmovups dA15+__svml_datan2_data_internal(%rip), %zmm15 + vpcmpgtd %ymm6, %ymm9, %ymm14 + vpcmpeqd %ymm6, %ymm9, %ymm6 + vpor %ymm6, %ymm14, %ymm10 + vmulpd {rn-sae}, %zmm13, %zmm13, %zmm14 + vmovups dA18+__svml_datan2_data_internal(%rip), %zmm9 + vpor %ymm10, %ymm8, %ymm6 + vmovups dA17+__svml_datan2_data_internal(%rip), %zmm10 + vfmadd231pd {rn-sae}, %zmm14, %zmm11, %zmm15 + vmovups dA14+__svml_datan2_data_internal(%rip), %zmm11 + vmovups dA12+__svml_datan2_data_internal(%rip), %zmm8 + vfmadd231pd {rn-sae}, %zmm14, %zmm9, %zmm11 + vmovups dA13+__svml_datan2_data_internal(%rip), %zmm9 + vfmadd231pd {rn-sae}, %zmm14, %zmm10, %zmm9 + vmovups dA16+__svml_datan2_data_internal(%rip), %zmm10 + vfmadd231pd {rn-sae}, %zmm14, %zmm10, %zmm8 + vmovups dA11+__svml_datan2_data_internal(%rip), %zmm10 + vfmadd213pd {rn-sae}, %zmm10, %zmm14, %zmm15 + vmovups dA10+__svml_datan2_data_internal(%rip), %zmm10 + vfmadd213pd {rn-sae}, %zmm10, %zmm14, %zmm11 + vmovups dA09+__svml_datan2_data_internal(%rip), %zmm10 + vfmadd213pd {rn-sae}, %zmm10, %zmm14, %zmm9 + vmovups dA08+__svml_datan2_data_internal(%rip), %zmm10 + vfmadd213pd {rn-sae}, %zmm10, %zmm14, %zmm8 + vmovups dA07+__svml_datan2_data_internal(%rip), %zmm10 + vfmadd213pd {rn-sae}, %zmm10, %zmm14, %zmm15 + vmovups dA06+__svml_datan2_data_internal(%rip), %zmm10 + vfmadd213pd {rn-sae}, %zmm10, %zmm14, %zmm11 + vmovups dA05+__svml_datan2_data_internal(%rip), %zmm10 + vfmadd213pd {rn-sae}, %zmm10, %zmm14, %zmm9 + vmovups dA04+__svml_datan2_data_internal(%rip), %zmm10 + vfmadd213pd {rn-sae}, %zmm10, %zmm14, %zmm8 + vmovups dA03+__svml_datan2_data_internal(%rip), %zmm10 + +/* A00=1.0, account for it later VQFMA(D, dP4, dP4, dR8, dA00); */ + vmulpd {rn-sae}, %zmm14, %zmm8, %zmm8 + vfmadd213pd {rn-sae}, %zmm10, %zmm14, %zmm15 + vmovups dA02+__svml_datan2_data_internal(%rip), %zmm10 + vfmadd213pd {rn-sae}, %zmm10, %zmm14, %zmm11 + vmovups dA01+__svml_datan2_data_internal(%rip), %zmm10 + vfmadd213pd {rn-sae}, %zmm11, %zmm12, %zmm15 + vfmadd213pd {rn-sae}, %zmm10, %zmm14, %zmm9 + vfmadd213pd {rn-sae}, %zmm8, %zmm12, %zmm9 + vmovups __svml_datan2_data_internal(%rip), %zmm8 + vfmadd213pd {rn-sae}, %zmm9, %zmm13, %zmm15 + +/* + * Reconstruction. + * dP=(R+R*dP) + dPIO2 + */ + vfmadd213pd {rn-sae}, %zmm1, %zmm1, %zmm15 + vaddpd {rn-sae}, %zmm3, %zmm15, %zmm1 + vorpd %zmm5, %zmm1, %zmm9 + +/* if x<0, dPI = Pi, else dPI =0 */ + vmovups dZERO+__svml_datan2_data_internal(%rip), %zmm1 + vcmppd $18, {sae}, %zmm1, %zmm7, %k2 + vaddpd {rn-sae}, %zmm8, %zmm9, %zmm9{%k2} + vmovmskps %ymm6, %eax + vorpd %zmm4, %zmm9, %zmm11 + +/* Special branch for fast (vector) processing of zero arguments */ + vmovups 64(%rsp), %zmm9 + testl %eax, %eax + +/* Go to auxilary branch */ + jne L(AUX_BRANCH) + # LOE rbx r12 r13 r14 r15 edx ymm6 zmm0 zmm2 zmm3 zmm4 zmm5 zmm7 zmm9 zmm11 + +/* Return from auxilary branch + * for out of main path inputs + */ + +L(AUX_BRANCH_RETURN): +/* + * Special branch for fast (vector) processing of zero arguments + * The end of implementation + */ + testl %edx, %edx + +/* Go to special inputs processing branch */ + jne L(SPECIAL_VALUES_BRANCH) + # LOE rbx r12 r13 r14 r15 edx zmm0 zmm7 zmm11 + +/* Restore registers + * and exit the function + */ + +L(EXIT): + vmovaps %zmm11, %zmm0 + movq %rbp, %rsp + popq %rbp + cfi_def_cfa(7, 8) + cfi_restore(6) + ret + cfi_def_cfa(6, 16) + cfi_offset(6, -16) + +/* Branch to process + * special inputs + */ + +L(SPECIAL_VALUES_BRANCH): + vmovups %zmm0, 64(%rsp) + vmovups %zmm7, 128(%rsp) + vmovups %zmm11, 192(%rsp) + # LOE rbx r12 r13 r14 r15 edx zmm11 + + xorl %eax, %eax + # LOE rbx r12 r13 r14 r15 eax edx + + vzeroupper + movq %r12, 16(%rsp) + /* DW_CFA_expression: r12 (r12) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -240; DW_OP_plus) */ + .cfi_escape 0x10, 0x0c, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x10, 0xff, 0xff, 0xff, 0x22 + movl %eax, %r12d + movq %r13, 8(%rsp) + /* DW_CFA_expression: r13 (r13) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -248; DW_OP_plus) */ + .cfi_escape 0x10, 0x0d, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x08, 0xff, 0xff, 0xff, 0x22 + movl %edx, %r13d + movq %r14, (%rsp) + /* DW_CFA_expression: r14 (r14) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -256; DW_OP_plus) */ + .cfi_escape 0x10, 0x0e, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x00, 0xff, 0xff, 0xff, 0x22 + # LOE rbx r15 r12d r13d + +/* Range mask + * bits check + */ + +L(RANGEMASK_CHECK): + btl %r12d, %r13d + +/* Call scalar math function */ + jc L(SCALAR_MATH_CALL) + # LOE rbx r15 r12d r13d + +/* Special inputs + * processing loop + */ + +L(SPECIAL_VALUES_LOOP): + incl %r12d + cmpl $8, %r12d + +/* Check bits in range mask */ + jl L(RANGEMASK_CHECK) + # LOE rbx r15 r12d r13d + + movq 16(%rsp), %r12 + cfi_restore(12) + movq 8(%rsp), %r13 + cfi_restore(13) + movq (%rsp), %r14 + cfi_restore(14) + vmovups 192(%rsp), %zmm11 + +/* Go to exit */ + jmp L(EXIT) + /* DW_CFA_expression: r12 (r12) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -240; DW_OP_plus) */ + .cfi_escape 0x10, 0x0c, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x10, 0xff, 0xff, 0xff, 0x22 + /* DW_CFA_expression: r13 (r13) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -248; DW_OP_plus) */ + .cfi_escape 0x10, 0x0d, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x08, 0xff, 0xff, 0xff, 0x22 + /* DW_CFA_expression: r14 (r14) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -256; DW_OP_plus) */ + .cfi_escape 0x10, 0x0e, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x00, 0xff, 0xff, 0xff, 0x22 + # LOE rbx r12 r13 r14 r15 zmm11 + +/* Scalar math fucntion call + * to process special input + */ + +L(SCALAR_MATH_CALL): + movl %r12d, %r14d + movsd 64(%rsp,%r14,8), %xmm0 + movsd 128(%rsp,%r14,8), %xmm1 + call atan2@PLT + # LOE rbx r14 r15 r12d r13d xmm0 + + movsd %xmm0, 192(%rsp,%r14,8) + +/* Process special inputs in loop */ + jmp L(SPECIAL_VALUES_LOOP) + cfi_restore(12) + cfi_restore(13) + cfi_restore(14) + # LOE rbx r15 r12d r13d + +/* Auxilary branch + * for out of main path inputs + */ + +L(AUX_BRANCH): +/* Check if at least on of Y or Y is zero: iAXAYZERO */ + vmovups dZERO+__svml_datan2_data_internal(%rip), %zmm8 + +/* Check if both X & Y are not NaNs: iXYnotNAN */ + vcmppd $3, {sae}, %zmm7, %zmm7, %k1 + vcmppd $3, {sae}, %zmm0, %zmm0, %k2 + vcmppd $4, {sae}, %zmm8, %zmm2, %k3 + vcmppd $4, {sae}, %zmm8, %zmm9, %k4 + +/* Res = sign(Y)*(X<0)?(PIO2+PI):PIO2 */ + vpcmpgtq %zmm7, %zmm8, %k6 + vpternlogd $0xff, %zmm1, %zmm1, %zmm10 + vmovaps %zmm10, %zmm15 + vmovaps %zmm10, %zmm12 + vmovaps %zmm10, %zmm13 + vpandnq %zmm2, %zmm2, %zmm15{%k3} + vmovaps %zmm10, %zmm2 + vpandnq %zmm7, %zmm7, %zmm12{%k1} + vpandnq %zmm0, %zmm0, %zmm13{%k2} + vpandnq %zmm9, %zmm9, %zmm2{%k4} + vandpd %zmm13, %zmm12, %zmm14 + vorpd %zmm2, %zmm15, %zmm9 + vpsrlq $32, %zmm14, %zmm1 + vpsrlq $32, %zmm9, %zmm2 + vpmovqd %zmm1, %ymm1 + vpmovqd %zmm2, %ymm9 + +/* Check if at least on of Y or Y is zero and not NaN: iAXAYZEROnotNAN */ + vpand %ymm1, %ymm9, %ymm2 + +/* + * Path for zero arguments (at least one of both) + * Check if both args are zeros (den. is zero) + */ + vmovups (%rsp), %zmm1 + +/* Exclude from previous callout mask zero (and not NaN) arguments */ + vpandn %ymm6, %ymm2, %ymm6 + vcmppd $4, {sae}, %zmm8, %zmm1, %k5 + +/* Go to callout */ + vmovmskps %ymm6, %edx + vpandnq %zmm1, %zmm1, %zmm10{%k5} + +/* Set sPIO2 to zero if den. is zero */ + vpandnq %zmm3, %zmm10, %zmm3 + vpandq %zmm10, %zmm8, %zmm1 + vporq %zmm1, %zmm3, %zmm3 + vorpd %zmm5, %zmm3, %zmm1 + vmovups __svml_datan2_data_internal(%rip), %zmm5 + vaddpd {rn-sae}, %zmm5, %zmm1, %zmm1{%k6} + vorpd %zmm4, %zmm1, %zmm1 + +/* Merge results from main and spec path */ + vpmovzxdq %ymm2, %zmm4 + vpsllq $32, %zmm4, %zmm2 + vpord %zmm2, %zmm4, %zmm3 + vpandnq %zmm11, %zmm3, %zmm11 + vpandq %zmm3, %zmm1, %zmm1 + vporq %zmm1, %zmm11, %zmm11 + +/* Return to main vector processing path */ + jmp L(AUX_BRANCH_RETURN) + # LOE rbx r12 r13 r14 r15 edx zmm0 zmm7 zmm11 +END(_ZGVeN8vv_atan2_skx) + + .section .rodata, "a" + .align 64 + +#ifdef __svml_datan2_data_internal_typedef +typedef unsigned int VUINT32; +typedef struct { + __declspec(align(64)) VUINT32 dPI[8][2]; + __declspec(align(64)) VUINT32 dPIO2[8][2]; + __declspec(align(64)) VUINT32 dA19[8][2]; + __declspec(align(64)) VUINT32 dA18[8][2]; + __declspec(align(64)) VUINT32 dA17[8][2]; + __declspec(align(64)) VUINT32 dA16[8][2]; + __declspec(align(64)) VUINT32 dA15[8][2]; + __declspec(align(64)) VUINT32 dA14[8][2]; + __declspec(align(64)) VUINT32 dA13[8][2]; + __declspec(align(64)) VUINT32 dA12[8][2]; + __declspec(align(64)) VUINT32 dA11[8][2]; + __declspec(align(64)) VUINT32 dA10[8][2]; + __declspec(align(64)) VUINT32 dA09[8][2]; + __declspec(align(64)) VUINT32 dA08[8][2]; + __declspec(align(64)) VUINT32 dA07[8][2]; + __declspec(align(64)) VUINT32 dA06[8][2]; + __declspec(align(64)) VUINT32 dA05[8][2]; + __declspec(align(64)) VUINT32 dA04[8][2]; + __declspec(align(64)) VUINT32 dA03[8][2]; + __declspec(align(64)) VUINT32 dA02[8][2]; + __declspec(align(64)) VUINT32 dA01[8][2]; + __declspec(align(64)) VUINT32 dA00[8][2]; + __declspec(align(64)) VUINT32 dSIGN_MASK[8][2]; + __declspec(align(64)) VUINT32 iCHK_WORK_SUB[16][1]; + __declspec(align(64)) VUINT32 iCHK_WORK_CMP[16][1]; + __declspec(align(64)) VUINT32 dABS_MASK[8][2]; + __declspec(align(64)) VUINT32 dZERO[8][2]; +} __svml_datan2_data_internal; +#endif +__svml_datan2_data_internal: + .quad 0x400921FB54442D18, 0x400921FB54442D18, 0x400921FB54442D18, 0x400921FB54442D18, 0x400921FB54442D18, 0x400921FB54442D18, 0x400921FB54442D18, 0x400921FB54442D18 //dPI + .align 64 + .quad 0x3FF921FB54442D18, 0x3FF921FB54442D18, 0x3FF921FB54442D18, 0x3FF921FB54442D18, 0x3FF921FB54442D18, 0x3FF921FB54442D18, 0x3FF921FB54442D18, 0x3FF921FB54442D18 //dPIO2 + .align 64 + .quad 0xBEF4FDB537ABC7A3, 0xBEF4FDB537ABC7A3, 0xBEF4FDB537ABC7A3, 0xBEF4FDB537ABC7A3, 0xBEF4FDB537ABC7A3, 0xBEF4FDB537ABC7A3, 0xBEF4FDB537ABC7A3, 0xBEF4FDB537ABC7A3 // dA19 + .align 64 + .quad 0x3F2CED0A36665209, 0x3F2CED0A36665209, 0x3F2CED0A36665209, 0x3F2CED0A36665209, 0x3F2CED0A36665209, 0x3F2CED0A36665209, 0x3F2CED0A36665209, 0x3F2CED0A36665209 // dA18 + .align 64 + .quad 0xBF52E67C93954C23, 0xBF52E67C93954C23, 0xBF52E67C93954C23, 0xBF52E67C93954C23, 0xBF52E67C93954C23, 0xBF52E67C93954C23, 0xBF52E67C93954C23, 0xBF52E67C93954C23 // dA17 + .align 64 + .quad 0x3F6F5A1DAE82AFB3, 0x3F6F5A1DAE82AFB3, 0x3F6F5A1DAE82AFB3, 0x3F6F5A1DAE82AFB3, 0x3F6F5A1DAE82AFB3, 0x3F6F5A1DAE82AFB3, 0x3F6F5A1DAE82AFB3, 0x3F6F5A1DAE82AFB3 // dA16 + .align 64 + .quad 0xBF82B2EC618E4BAD, 0xBF82B2EC618E4BAD, 0xBF82B2EC618E4BAD, 0xBF82B2EC618E4BAD, 0xBF82B2EC618E4BAD, 0xBF82B2EC618E4BAD, 0xBF82B2EC618E4BAD, 0xBF82B2EC618E4BAD // dA15 + .align 64 + .quad 0x3F914F4C661116A5, 0x3F914F4C661116A5, 0x3F914F4C661116A5, 0x3F914F4C661116A5, 0x3F914F4C661116A5, 0x3F914F4C661116A5, 0x3F914F4C661116A5, 0x3F914F4C661116A5 // dA14 + .align 64 + .quad 0xBF9A5E83B081F69C, 0xBF9A5E83B081F69C, 0xBF9A5E83B081F69C, 0xBF9A5E83B081F69C, 0xBF9A5E83B081F69C, 0xBF9A5E83B081F69C, 0xBF9A5E83B081F69C, 0xBF9A5E83B081F69C // dA13 + .align 64 + .quad 0x3FA169980CB6AD4F, 0x3FA169980CB6AD4F, 0x3FA169980CB6AD4F, 0x3FA169980CB6AD4F, 0x3FA169980CB6AD4F, 0x3FA169980CB6AD4F, 0x3FA169980CB6AD4F, 0x3FA169980CB6AD4F // dA12 + .align 64 + .quad 0xBFA4EFA2E563C1BC, 0xBFA4EFA2E563C1BC, 0xBFA4EFA2E563C1BC, 0xBFA4EFA2E563C1BC, 0xBFA4EFA2E563C1BC, 0xBFA4EFA2E563C1BC, 0xBFA4EFA2E563C1BC, 0xBFA4EFA2E563C1BC // dA11 + .align 64 + .quad 0x3FA7EC0FBC50683B, 0x3FA7EC0FBC50683B, 0x3FA7EC0FBC50683B, 0x3FA7EC0FBC50683B, 0x3FA7EC0FBC50683B, 0x3FA7EC0FBC50683B, 0x3FA7EC0FBC50683B, 0x3FA7EC0FBC50683B // dA10 + .align 64 + .quad 0xBFAAD261EAA09954, 0xBFAAD261EAA09954, 0xBFAAD261EAA09954, 0xBFAAD261EAA09954, 0xBFAAD261EAA09954, 0xBFAAD261EAA09954, 0xBFAAD261EAA09954, 0xBFAAD261EAA09954 // dA09 + .align 64 + .quad 0x3FAE1749BD612DCF, 0x3FAE1749BD612DCF, 0x3FAE1749BD612DCF, 0x3FAE1749BD612DCF, 0x3FAE1749BD612DCF, 0x3FAE1749BD612DCF, 0x3FAE1749BD612DCF, 0x3FAE1749BD612DCF // dA08 + .align 64 + .quad 0xBFB11084009435E0, 0xBFB11084009435E0, 0xBFB11084009435E0, 0xBFB11084009435E0, 0xBFB11084009435E0, 0xBFB11084009435E0, 0xBFB11084009435E0, 0xBFB11084009435E0 // dA07 + .align 64 + .quad 0x3FB3B12A49295651, 0x3FB3B12A49295651, 0x3FB3B12A49295651, 0x3FB3B12A49295651, 0x3FB3B12A49295651, 0x3FB3B12A49295651, 0x3FB3B12A49295651, 0x3FB3B12A49295651 // dA06 + .align 64 + .quad 0xBFB745D009BADA94, 0xBFB745D009BADA94, 0xBFB745D009BADA94, 0xBFB745D009BADA94, 0xBFB745D009BADA94, 0xBFB745D009BADA94, 0xBFB745D009BADA94, 0xBFB745D009BADA94 // dA05 + .align 64 + .quad 0x3FBC71C707F7D5B5, 0x3FBC71C707F7D5B5, 0x3FBC71C707F7D5B5, 0x3FBC71C707F7D5B5, 0x3FBC71C707F7D5B5, 0x3FBC71C707F7D5B5, 0x3FBC71C707F7D5B5, 0x3FBC71C707F7D5B5 // dA04 + .align 64 + .quad 0xBFC2492491EE55C7, 0xBFC2492491EE55C7, 0xBFC2492491EE55C7, 0xBFC2492491EE55C7, 0xBFC2492491EE55C7, 0xBFC2492491EE55C7, 0xBFC2492491EE55C7, 0xBFC2492491EE55C7 // dA03 + .align 64 + .quad 0x3FC999999997EE34, 0x3FC999999997EE34, 0x3FC999999997EE34, 0x3FC999999997EE34, 0x3FC999999997EE34, 0x3FC999999997EE34, 0x3FC999999997EE34, 0x3FC999999997EE34 // dA02 + .align 64 + .quad 0xBFD55555555553C5, 0xBFD55555555553C5, 0xBFD55555555553C5, 0xBFD55555555553C5, 0xBFD55555555553C5, 0xBFD55555555553C5, 0xBFD55555555553C5, 0xBFD55555555553C5 // dA01 + .align 64 + .quad 0x3FF0000000000000, 0x3FF0000000000000, 0x3FF0000000000000, 0x3FF0000000000000, 0x3FF0000000000000, 0x3FF0000000000000, 0x3FF0000000000000, 0x3FF0000000000000 // dA00 + .align 64 + .quad 0x8000000000000000, 0x8000000000000000, 0x8000000000000000, 0x8000000000000000, 0x8000000000000000, 0x8000000000000000, 0x8000000000000000, 0x8000000000000000 //dSIGN_MASK + .align 64 + .long 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000 //iCHK_WORK_SUB + .align 64 + .long 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000 //iCHK_WORK_CMP + .align 64 + .quad 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff //dABS_MASK + .align 64 + .quad 0x0000000000000000, 0x0000000000000000, 0x0000000000000000, 0x0000000000000000, 0x0000000000000000, 0x0000000000000000, 0x0000000000000000, 0x0000000000000000 //dZERO + .align 64 + .type __svml_datan2_data_internal,@object + .size __svml_datan2_data_internal,.-__svml_datan2_data_internal diff --git a/sysdeps/x86_64/fpu/multiarch/svml_s_atan2f16_core-avx2.S b/sysdeps/x86_64/fpu/multiarch/svml_s_atan2f16_core-avx2.S new file mode 100644 index 0000000000..a2a76e8bfd --- /dev/null +++ b/sysdeps/x86_64/fpu/multiarch/svml_s_atan2f16_core-avx2.S @@ -0,0 +1,20 @@ +/* AVX2 version of vectorized atan2f. + Copyright (C) 2021 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, see + . */ + +#define _ZGVeN16vv_atan2f _ZGVeN16vv_atan2f_avx2_wrapper +#include "../svml_s_atan2f16_core.S" diff --git a/sysdeps/x86_64/fpu/multiarch/svml_s_atan2f16_core.c b/sysdeps/x86_64/fpu/multiarch/svml_s_atan2f16_core.c new file mode 100644 index 0000000000..6fa806414d --- /dev/null +++ b/sysdeps/x86_64/fpu/multiarch/svml_s_atan2f16_core.c @@ -0,0 +1,28 @@ +/* Multiple versions of vectorized atan2f, vector length is 16. + Copyright (C) 2021 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, see + . */ + +#define SYMBOL_NAME _ZGVeN16vv_atan2f +#include "ifunc-mathvec-avx512-skx.h" + +libc_ifunc_redirected (REDIRECT_NAME, SYMBOL_NAME, IFUNC_SELECTOR ()); + +#ifdef SHARED +__hidden_ver1 (_ZGVeN16vv_atan2f, __GI__ZGVeN16vv_atan2f, + __redirect__ZGVeN16vv_atan2f) + __attribute__ ((visibility ("hidden"))); +#endif diff --git a/sysdeps/x86_64/fpu/multiarch/svml_s_atan2f16_core_avx512.S b/sysdeps/x86_64/fpu/multiarch/svml_s_atan2f16_core_avx512.S new file mode 100644 index 0000000000..f3477cc8e6 --- /dev/null +++ b/sysdeps/x86_64/fpu/multiarch/svml_s_atan2f16_core_avx512.S @@ -0,0 +1,399 @@ +/* Function atan2f vectorized with AVX-512. + Copyright (C) 2021 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, see + https://www.gnu.org/licenses/. */ + +/* + * ALGORITHM DESCRIPTION: + * For 0.0 <= x <= 7.0/16.0: atan(x) = atan(0.0) + atan(s), where s=(x-0.0)/(1.0+0.0*x) + * For 7.0/16.0 <= x <= 11.0/16.0: atan(x) = atan(0.5) + atan(s), where s=(x-0.5)/(1.0+0.5*x) + * For 11.0/16.0 <= x <= 19.0/16.0: atan(x) = atan(1.0) + atan(s), where s=(x-1.0)/(1.0+1.0*x) + * For 19.0/16.0 <= x <= 39.0/16.0: atan(x) = atan(1.5) + atan(s), where s=(x-1.5)/(1.0+1.5*x) + * For 39.0/16.0 <= x <= inf : atan(x) = atan(inf) + atan(s), where s=-1.0/x + * Where atan(s) ~= s+s^3*Poly11(s^2) on interval |s|<7.0/0.16. + * + * + */ + +/* Offsets for data table __svml_satan2_data_internal + */ +#define sZERO 0 +#define sONE 64 +#define sSIGN_MASK 128 +#define sABS_MASK 192 +#define sPIO2 256 +#define sPI 320 +#define sPC8 384 +#define sPC7 448 +#define sPC6 512 +#define sPC5 576 +#define sPC4 640 +#define sPC3 704 +#define sPC2 768 +#define sPC1 832 +#define sPC0 896 +#define iCHK_WORK_SUB 960 +#define iCHK_WORK_CMP 1024 + +#include + + .text + .section .text.exex512,"ax",@progbits +ENTRY(_ZGVeN16vv_atan2f_skx) + pushq %rbp + cfi_def_cfa_offset(16) + movq %rsp, %rbp + cfi_def_cfa(6, 16) + cfi_offset(6, -16) + andq $-64, %rsp + subq $256, %rsp + xorl %edx, %edx + +/* + * #define NO_VECTOR_ZERO_ATAN2_ARGS + * Declarations + * Variables + * Constants + * The end of declarations + * Implementation + * Arguments signs + */ + vmovups sABS_MASK+__svml_satan2_data_internal(%rip), %zmm6 + vmovups sONE+__svml_satan2_data_internal(%rip), %zmm3 + +/* Testing on working interval. */ + vmovups iCHK_WORK_SUB+__svml_satan2_data_internal(%rip), %zmm9 + vmovups iCHK_WORK_CMP+__svml_satan2_data_internal(%rip), %zmm14 + +/* + * 1) If yx then a=-x, b=y, PIO2=Pi/2 + */ + vmovups sPIO2+__svml_satan2_data_internal(%rip), %zmm4 + vpternlogd $255, %zmm13, %zmm13, %zmm13 + vmovaps %zmm1, %zmm8 + vandps %zmm6, %zmm8, %zmm2 + vandps %zmm6, %zmm0, %zmm1 + vorps sSIGN_MASK+__svml_satan2_data_internal(%rip), %zmm2, %zmm5 + vpsubd %zmm9, %zmm2, %zmm10 + vpsubd %zmm9, %zmm1, %zmm12 + vxorps %zmm2, %zmm8, %zmm7 + vxorps %zmm1, %zmm0, %zmm6 + vcmpps $17, {sae}, %zmm2, %zmm1, %k1 + vpcmpgtd %zmm10, %zmm14, %k2 + vpcmpgtd %zmm12, %zmm14, %k3 + vmovups sPC6+__svml_satan2_data_internal(%rip), %zmm14 + vblendmps %zmm1, %zmm5, %zmm11{%k1} + vblendmps %zmm2, %zmm1, %zmm5{%k1} + vxorps %zmm4, %zmm4, %zmm4{%k1} + +/* + * Division a/b. + * Enabled when FMA is available and + * performance is better with NR iteration + */ + vrcp14ps %zmm5, %zmm15 + vfnmadd231ps {rn-sae}, %zmm5, %zmm15, %zmm3 + vfmadd213ps {rn-sae}, %zmm15, %zmm3, %zmm15 + vmulps {rn-sae}, %zmm15, %zmm11, %zmm3 + vfnmadd231ps {rn-sae}, %zmm5, %zmm3, %zmm11 + vfmadd213ps {rn-sae}, %zmm3, %zmm11, %zmm15 + vmovups sPC8+__svml_satan2_data_internal(%rip), %zmm11 + vpternlogd $255, %zmm3, %zmm3, %zmm3 + +/* Polynomial. */ + vmulps {rn-sae}, %zmm15, %zmm15, %zmm9 + vpandnd %zmm10, %zmm10, %zmm13{%k2} + vmulps {rn-sae}, %zmm9, %zmm9, %zmm10 + vfmadd231ps {rn-sae}, %zmm10, %zmm11, %zmm14 + vmovups sPC5+__svml_satan2_data_internal(%rip), %zmm11 + vpandnd %zmm12, %zmm12, %zmm3{%k3} + vpord %zmm3, %zmm13, %zmm3 + vmovups sPC4+__svml_satan2_data_internal(%rip), %zmm13 + vmovups sPC7+__svml_satan2_data_internal(%rip), %zmm12 + vptestmd %zmm3, %zmm3, %k0 + vfmadd213ps {rn-sae}, %zmm13, %zmm10, %zmm14 + vfmadd231ps {rn-sae}, %zmm10, %zmm12, %zmm11 + vmovups sPC3+__svml_satan2_data_internal(%rip), %zmm12 + vmovups sPC2+__svml_satan2_data_internal(%rip), %zmm13 + +/* Special branch for fast (vector) processing of zero arguments */ + kortestw %k0, %k0 + vfmadd213ps {rn-sae}, %zmm12, %zmm10, %zmm11 + vmovups sPC1+__svml_satan2_data_internal(%rip), %zmm12 + vfmadd213ps {rn-sae}, %zmm13, %zmm10, %zmm14 + vmovups sPC0+__svml_satan2_data_internal(%rip), %zmm13 + vfmadd213ps {rn-sae}, %zmm12, %zmm10, %zmm11 + vfmadd213ps {rn-sae}, %zmm13, %zmm10, %zmm14 + vfmadd213ps {rn-sae}, %zmm14, %zmm9, %zmm11 + +/* Reconstruction. */ + vfmadd213ps {rn-sae}, %zmm4, %zmm15, %zmm11 + +/* if x<0, sPI = Pi, else sPI =0 */ + vmovups __svml_satan2_data_internal(%rip), %zmm15 + vorps %zmm7, %zmm11, %zmm9 + vcmpps $18, {sae}, %zmm15, %zmm8, %k4 + vmovups sPI+__svml_satan2_data_internal(%rip), %zmm11 + vaddps {rn-sae}, %zmm11, %zmm9, %zmm9{%k4} + vorps %zmm6, %zmm9, %zmm10 + +/* Go to auxilary branch */ + jne L(AUX_BRANCH) + # LOE rbx r12 r13 r14 r15 edx zmm0 zmm1 zmm2 zmm3 zmm4 zmm5 zmm6 zmm7 zmm8 zmm10 zmm11 + +/* Return from auxilary branch + * for out of main path inputs + */ + +L(AUX_BRANCH_RETURN): +/* + * Special branch for fast (vector) processing of zero arguments + * The end of implementation + */ + testl %edx, %edx + +/* Go to special inputs processing branch */ + jne L(SPECIAL_VALUES_BRANCH) + # LOE rbx r12 r13 r14 r15 edx zmm0 zmm8 zmm10 + +/* Restore registers + * and exit the function + */ + +L(EXIT): + vmovaps %zmm10, %zmm0 + movq %rbp, %rsp + popq %rbp + cfi_def_cfa(7, 8) + cfi_restore(6) + ret + cfi_def_cfa(6, 16) + cfi_offset(6, -16) + +/* Branch to process + * special inputs + */ + +L(SPECIAL_VALUES_BRANCH): + vmovups %zmm0, 64(%rsp) + vmovups %zmm8, 128(%rsp) + vmovups %zmm10, 192(%rsp) + # LOE rbx r12 r13 r14 r15 edx zmm10 + + xorl %eax, %eax + # LOE rbx r12 r13 r14 r15 eax edx + + vzeroupper + movq %r12, 16(%rsp) + /* DW_CFA_expression: r12 (r12) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -240; DW_OP_plus) */ + .cfi_escape 0x10, 0x0c, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x10, 0xff, 0xff, 0xff, 0x22 + movl %eax, %r12d + movq %r13, 8(%rsp) + /* DW_CFA_expression: r13 (r13) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -248; DW_OP_plus) */ + .cfi_escape 0x10, 0x0d, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x08, 0xff, 0xff, 0xff, 0x22 + movl %edx, %r13d + movq %r14, (%rsp) + /* DW_CFA_expression: r14 (r14) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -256; DW_OP_plus) */ + .cfi_escape 0x10, 0x0e, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x00, 0xff, 0xff, 0xff, 0x22 + # LOE rbx r15 r12d r13d + +/* Range mask + * bits check + */ + +L(RANGEMASK_CHECK): + btl %r12d, %r13d + +/* Call scalar math function */ + jc L(SCALAR_MATH_CALL) + # LOE rbx r15 r12d r13d + +/* Special inputs + * processing loop + */ + +L(SPECIAL_VALUES_LOOP): + incl %r12d + cmpl $16, %r12d + +/* Check bits in range mask */ + jl L(RANGEMASK_CHECK) + # LOE rbx r15 r12d r13d + + movq 16(%rsp), %r12 + cfi_restore(12) + movq 8(%rsp), %r13 + cfi_restore(13) + movq (%rsp), %r14 + cfi_restore(14) + vmovups 192(%rsp), %zmm10 + +/* Go to exit */ + jmp L(EXIT) + /* DW_CFA_expression: r12 (r12) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -240; DW_OP_plus) */ + .cfi_escape 0x10, 0x0c, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x10, 0xff, 0xff, 0xff, 0x22 + /* DW_CFA_expression: r13 (r13) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -248; DW_OP_plus) */ + .cfi_escape 0x10, 0x0d, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x08, 0xff, 0xff, 0xff, 0x22 + /* DW_CFA_expression: r14 (r14) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -64; DW_OP_and; DW_OP_const4s: -256; DW_OP_plus) */ + .cfi_escape 0x10, 0x0e, 0x0e, 0x38, 0x1c, 0x0d, 0xc0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x00, 0xff, 0xff, 0xff, 0x22 + # LOE rbx r12 r13 r14 r15 zmm10 + +/* Scalar math fucntion call + * to process special input + */ + +L(SCALAR_MATH_CALL): + movl %r12d, %r14d + movss 64(%rsp,%r14,4), %xmm0 + movss 128(%rsp,%r14,4), %xmm1 + call atan2f@PLT + # LOE rbx r14 r15 r12d r13d xmm0 + + movss %xmm0, 192(%rsp,%r14,4) + +/* Process special inputs in loop */ + jmp L(SPECIAL_VALUES_LOOP) + cfi_restore(12) + cfi_restore(13) + cfi_restore(14) + # LOE rbx r15 r12d r13d + +/* Auxilary branch + * for out of main path inputs + */ + +L(AUX_BRANCH): +/* Check if at least on of Y or Y is zero: iAXAYZERO */ + vmovups __svml_satan2_data_internal(%rip), %zmm9 + +/* Check if both X & Y are not NaNs: iXYnotNAN */ + vcmpps $3, {sae}, %zmm8, %zmm8, %k1 + vcmpps $3, {sae}, %zmm0, %zmm0, %k2 + vpcmpd $4, %zmm9, %zmm2, %k3 + vpcmpd $4, %zmm9, %zmm1, %k4 + +/* + * Path for zero arguments (at least one of both) + * Check if both args are zeros (den. is zero) + */ + vcmpps $4, {sae}, %zmm9, %zmm5, %k5 + +/* Res = sign(Y)*(X<0)?(PIO2+PI):PIO2 */ + vpcmpgtd %zmm8, %zmm9, %k6 + vpternlogd $255, %zmm14, %zmm14, %zmm14 + vpternlogd $255, %zmm12, %zmm12, %zmm12 + vpternlogd $255, %zmm13, %zmm13, %zmm13 + vpandnd %zmm2, %zmm2, %zmm14{%k3} + vpternlogd $255, %zmm2, %zmm2, %zmm2 + vpandnd %zmm1, %zmm1, %zmm2{%k4} + vpord %zmm2, %zmm14, %zmm15 + vpternlogd $255, %zmm2, %zmm2, %zmm2 + vpandnd %zmm5, %zmm5, %zmm2{%k5} + +/* Set sPIO2 to zero if den. is zero */ + vpandnd %zmm4, %zmm2, %zmm4 + vpandd %zmm2, %zmm9, %zmm5 + vpord %zmm5, %zmm4, %zmm2 + vorps %zmm7, %zmm2, %zmm7 + vaddps {rn-sae}, %zmm11, %zmm7, %zmm7{%k6} + vorps %zmm6, %zmm7, %zmm6 + vpandnd %zmm8, %zmm8, %zmm12{%k1} + vpandnd %zmm0, %zmm0, %zmm13{%k2} + vandps %zmm13, %zmm12, %zmm12 + +/* Check if at least on of Y or Y is zero and not NaN: iAXAYZEROnotNAN */ + vpandd %zmm12, %zmm15, %zmm1 + +/* Exclude from previous callout mask zero (and not NaN) arguments */ + vpandnd %zmm3, %zmm1, %zmm3 + +/* Go to callout */ + vptestmd %zmm3, %zmm3, %k0 + kmovw %k0, %edx + +/* Merge results from main and spec path */ + vpandnd %zmm10, %zmm1, %zmm10 + vpandd %zmm1, %zmm6, %zmm11 + vpord %zmm11, %zmm10, %zmm10 + +/* Return to main vector processing path */ + jmp L(AUX_BRANCH_RETURN) + # LOE rbx r12 r13 r14 r15 edx zmm0 zmm8 zmm10 +END(_ZGVeN16vv_atan2f_skx) + + .section .rodata, "a" + .align 64 + +#ifdef __svml_satan2_data_internal_typedef +typedef unsigned int VUINT32; +typedef struct { + __declspec(align(64)) VUINT32 sZERO[16][1]; + __declspec(align(64)) VUINT32 sONE[16][1]; + __declspec(align(64)) VUINT32 sSIGN_MASK[16][1]; + __declspec(align(64)) VUINT32 sABS_MASK[16][1]; + __declspec(align(64)) VUINT32 sPIO2[16][1]; + __declspec(align(64)) VUINT32 sPI[16][1]; + __declspec(align(64)) VUINT32 sPC8[16][1]; + __declspec(align(64)) VUINT32 sPC7[16][1]; + __declspec(align(64)) VUINT32 sPC6[16][1]; + __declspec(align(64)) VUINT32 sPC5[16][1]; + __declspec(align(64)) VUINT32 sPC4[16][1]; + __declspec(align(64)) VUINT32 sPC3[16][1]; + __declspec(align(64)) VUINT32 sPC2[16][1]; + __declspec(align(64)) VUINT32 sPC1[16][1]; + __declspec(align(64)) VUINT32 sPC0[16][1]; + __declspec(align(64)) VUINT32 iCHK_WORK_SUB[16][1]; + __declspec(align(64)) VUINT32 iCHK_WORK_CMP[16][1]; +} __svml_satan2_data_internal; +#endif +__svml_satan2_data_internal: + .long 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000 // sZERO + .align 64 + .long 0x3f800000, 0x3f800000, 0x3f800000, 0x3f800000, 0x3f800000, 0x3f800000, 0x3f800000, 0x3f800000, 0x3f800000, 0x3f800000, 0x3f800000, 0x3f800000, 0x3f800000, 0x3f800000, 0x3f800000, 0x3f800000 // sONE + .align 64 + .long 0x80000000, 0x80000000, 0x80000000, 0x80000000, 0x80000000, 0x80000000, 0x80000000, 0x80000000, 0x80000000, 0x80000000, 0x80000000, 0x80000000, 0x80000000, 0x80000000, 0x80000000, 0x80000000 // sSIGN_MASK + .align 64 + .long 0x7FFFFFFF, 0x7FFFFFFF, 0x7FFFFFFF, 0x7FFFFFFF, 0x7FFFFFFF, 0x7FFFFFFF, 0x7FFFFFFF, 0x7FFFFFFF, 0x7FFFFFFF, 0x7FFFFFFF, 0x7FFFFFFF, 0x7FFFFFFF, 0x7FFFFFFF, 0x7FFFFFFF, 0x7FFFFFFF, 0x7FFFFFFF // sABS_MASK + .align 64 + .long 0x3FC90FDB, 0x3FC90FDB, 0x3FC90FDB, 0x3FC90FDB, 0x3FC90FDB, 0x3FC90FDB, 0x3FC90FDB, 0x3FC90FDB, 0x3FC90FDB, 0x3FC90FDB, 0x3FC90FDB, 0x3FC90FDB, 0x3FC90FDB, 0x3FC90FDB, 0x3FC90FDB, 0x3FC90FDB // sPIO2 + .align 64 + .long 0x40490FDB, 0x40490FDB, 0x40490FDB, 0x40490FDB, 0x40490FDB, 0x40490FDB, 0x40490FDB, 0x40490FDB, 0x40490FDB, 0x40490FDB, 0x40490FDB, 0x40490FDB, 0x40490FDB, 0x40490FDB, 0x40490FDB, 0x40490FDB // sPI + .align 64 + .long 0x3B322CC0, 0x3B322CC0, 0x3B322CC0, 0x3B322CC0, 0x3B322CC0, 0x3B322CC0, 0x3B322CC0, 0x3B322CC0, 0x3B322CC0, 0x3B322CC0, 0x3B322CC0, 0x3B322CC0, 0x3B322CC0, 0x3B322CC0, 0x3B322CC0, 0x3B322CC0 // sA08 + .align 64 + .long 0xBC7F2631, 0xBC7F2631, 0xBC7F2631, 0xBC7F2631, 0xBC7F2631, 0xBC7F2631, 0xBC7F2631, 0xBC7F2631, 0xBC7F2631, 0xBC7F2631, 0xBC7F2631, 0xBC7F2631, 0xBC7F2631, 0xBC7F2631, 0xBC7F2631, 0xBC7F2631 // sA07 + .align 64 + .long 0x3D2BC384, 0x3D2BC384, 0x3D2BC384, 0x3D2BC384, 0x3D2BC384, 0x3D2BC384, 0x3D2BC384, 0x3D2BC384, 0x3D2BC384, 0x3D2BC384, 0x3D2BC384, 0x3D2BC384, 0x3D2BC384, 0x3D2BC384, 0x3D2BC384, 0x3D2BC384 // sA06 + .align 64 + .long 0xBD987629, 0xBD987629, 0xBD987629, 0xBD987629, 0xBD987629, 0xBD987629, 0xBD987629, 0xBD987629, 0xBD987629, 0xBD987629, 0xBD987629, 0xBD987629, 0xBD987629, 0xBD987629, 0xBD987629, 0xBD987629 // sA05 + .align 64 + .long 0x3DD96474, 0x3DD96474, 0x3DD96474, 0x3DD96474, 0x3DD96474, 0x3DD96474, 0x3DD96474, 0x3DD96474, 0x3DD96474, 0x3DD96474, 0x3DD96474, 0x3DD96474, 0x3DD96474, 0x3DD96474, 0x3DD96474, 0x3DD96474 // sA04 + .align 64 + .long 0xBE1161F8, 0xBE1161F8, 0xBE1161F8, 0xBE1161F8, 0xBE1161F8, 0xBE1161F8, 0xBE1161F8, 0xBE1161F8, 0xBE1161F8, 0xBE1161F8, 0xBE1161F8, 0xBE1161F8, 0xBE1161F8, 0xBE1161F8, 0xBE1161F8, 0xBE1161F8 // sA03 + .align 64 + .long 0x3E4CB79F, 0x3E4CB79F, 0x3E4CB79F, 0x3E4CB79F, 0x3E4CB79F, 0x3E4CB79F, 0x3E4CB79F, 0x3E4CB79F, 0x3E4CB79F, 0x3E4CB79F, 0x3E4CB79F, 0x3E4CB79F, 0x3E4CB79F, 0x3E4CB79F, 0x3E4CB79F, 0x3E4CB79F // sA02 + .align 64 + .long 0xBEAAAA49, 0xBEAAAA49, 0xBEAAAA49, 0xBEAAAA49, 0xBEAAAA49, 0xBEAAAA49, 0xBEAAAA49, 0xBEAAAA49, 0xBEAAAA49, 0xBEAAAA49, 0xBEAAAA49, 0xBEAAAA49, 0xBEAAAA49, 0xBEAAAA49, 0xBEAAAA49, 0xBEAAAA49 // sA01 + .align 64 + .long 0x3F800000, 0x3F800000, 0x3F800000, 0x3F800000, 0x3F800000, 0x3F800000, 0x3F800000, 0x3F800000, 0x3F800000, 0x3F800000, 0x3F800000, 0x3F800000, 0x3F800000, 0x3F800000, 0x3F800000, 0x3F800000 // sA00 + .align 64 + .long 0x81000000, 0x81000000, 0x81000000, 0x81000000, 0x81000000, 0x81000000, 0x81000000, 0x81000000, 0x81000000, 0x81000000, 0x81000000, 0x81000000, 0x81000000, 0x81000000, 0x81000000, 0x81000000 //iCHK_WORK_SUB + .align 64 + .long 0xFC000000, 0xFC000000, 0xFC000000, 0xFC000000, 0xFC000000, 0xFC000000, 0xFC000000, 0xFC000000, 0xFC000000, 0xFC000000, 0xFC000000, 0xFC000000, 0xFC000000, 0xFC000000, 0xFC000000, 0xFC000000 //iCHK_WORK_CMP + .align 64 + .type __svml_satan2_data_internal,@object + .size __svml_satan2_data_internal,.-__svml_satan2_data_internal diff --git a/sysdeps/x86_64/fpu/multiarch/svml_s_atan2f4_core-sse2.S b/sysdeps/x86_64/fpu/multiarch/svml_s_atan2f4_core-sse2.S new file mode 100644 index 0000000000..d1a67facf1 --- /dev/null +++ b/sysdeps/x86_64/fpu/multiarch/svml_s_atan2f4_core-sse2.S @@ -0,0 +1,20 @@ +/* SSE2 version of vectorized atan2f. + Copyright (C) 2021 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, see + . */ + +#define _ZGVbN4vv_atan2f _ZGVbN4vv_atan2f_sse2 +#include "../svml_s_atan2f4_core.S" diff --git a/sysdeps/x86_64/fpu/multiarch/svml_s_atan2f4_core.c b/sysdeps/x86_64/fpu/multiarch/svml_s_atan2f4_core.c new file mode 100644 index 0000000000..ee882b0557 --- /dev/null +++ b/sysdeps/x86_64/fpu/multiarch/svml_s_atan2f4_core.c @@ -0,0 +1,28 @@ +/* Multiple versions of vectorized atan2f, vector length is 4. + Copyright (C) 2021 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, see + . */ + +#define SYMBOL_NAME _ZGVbN4vv_atan2f +#include "ifunc-mathvec-sse4_1.h" + +libc_ifunc_redirected (REDIRECT_NAME, SYMBOL_NAME, IFUNC_SELECTOR ()); + +#ifdef SHARED +__hidden_ver1 (_ZGVbN4vv_atan2f, __GI__ZGVbN4vv_atan2f, + __redirect__ZGVbN4vv_atan2f) + __attribute__ ((visibility ("hidden"))); +#endif diff --git a/sysdeps/x86_64/fpu/multiarch/svml_s_atan2f4_core_sse4.S b/sysdeps/x86_64/fpu/multiarch/svml_s_atan2f4_core_sse4.S new file mode 100644 index 0000000000..e4fbe82501 --- /dev/null +++ b/sysdeps/x86_64/fpu/multiarch/svml_s_atan2f4_core_sse4.S @@ -0,0 +1,384 @@ +/* Function atan2f vectorized with SSE4. + Copyright (C) 2021 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, see + https://www.gnu.org/licenses/. */ + +/* + * ALGORITHM DESCRIPTION: + * For 0.0 <= x <= 7.0/16.0: atan(x) = atan(0.0) + atan(s), where s=(x-0.0)/(1.0+0.0*x) + * For 7.0/16.0 <= x <= 11.0/16.0: atan(x) = atan(0.5) + atan(s), where s=(x-0.5)/(1.0+0.5*x) + * For 11.0/16.0 <= x <= 19.0/16.0: atan(x) = atan(1.0) + atan(s), where s=(x-1.0)/(1.0+1.0*x) + * For 19.0/16.0 <= x <= 39.0/16.0: atan(x) = atan(1.5) + atan(s), where s=(x-1.5)/(1.0+1.5*x) + * For 39.0/16.0 <= x <= inf : atan(x) = atan(inf) + atan(s), where s=-1.0/x + * Where atan(s) ~= s+s^3*Poly11(s^2) on interval |s|<7.0/0.16. + * + * + */ + +/* Offsets for data table __svml_satan2_data_internal + */ +#define sZERO 0 +#define sSIGN_MASK 16 +#define sABS_MASK 32 +#define sPIO2 48 +#define sPI 64 +#define sPC8 80 +#define sPC7 96 +#define sPC6 112 +#define sPC5 128 +#define sPC4 144 +#define sPC3 160 +#define sPC2 176 +#define sPC1 192 +#define sPC0 208 +#define iCHK_WORK_SUB 224 +#define iCHK_WORK_CMP 240 + +#include + + .text + .section .text.sse4,"ax",@progbits +ENTRY(_ZGVbN4vv_atan2f_sse4) + subq $88, %rsp + cfi_def_cfa_offset(96) + movaps %xmm0, %xmm12 + +/* + * #define NO_VECTOR_ZERO_ATAN2_ARGS + * Declarations + * Variables + * Constants + * The end of declarations + * Implementation + * Arguments signs + */ + movups sABS_MASK+__svml_satan2_data_internal(%rip), %xmm10 + movaps %xmm1, %xmm13 + movaps %xmm10, %xmm11 + andps %xmm12, %xmm10 + andps %xmm13, %xmm11 + movaps %xmm10, %xmm7 + cmpltps %xmm11, %xmm7 + +/* + * 1) If yx then a=-x, b=y, PIO2=Pi/2 + */ + movups sSIGN_MASK+__svml_satan2_data_internal(%rip), %xmm6 + movaps %xmm7, %xmm0 + orps %xmm11, %xmm6 + movaps %xmm10, %xmm4 + andnps %xmm6, %xmm0 + movaps %xmm7, %xmm6 + movaps %xmm11, %xmm5 + andps %xmm7, %xmm4 + andnps %xmm10, %xmm6 + andps %xmm7, %xmm5 + orps %xmm4, %xmm0 + orps %xmm5, %xmm6 + +/* Division a/b. */ + divps %xmm6, %xmm0 + +/* Testing on working interval. */ + movdqu iCHK_WORK_SUB+__svml_satan2_data_internal(%rip), %xmm14 + movaps %xmm11, %xmm15 + movaps %xmm10, %xmm3 + psubd %xmm14, %xmm15 + psubd %xmm14, %xmm3 + movdqa %xmm15, %xmm1 + movdqu iCHK_WORK_CMP+__svml_satan2_data_internal(%rip), %xmm2 + movdqa %xmm3, %xmm14 + pcmpgtd %xmm2, %xmm1 + pcmpeqd %xmm2, %xmm15 + pcmpgtd %xmm2, %xmm14 + pcmpeqd %xmm2, %xmm3 + por %xmm15, %xmm1 + por %xmm3, %xmm14 + por %xmm14, %xmm1 + +/* Polynomial. */ + movaps %xmm0, %xmm14 + mulps %xmm0, %xmm14 + movaps %xmm13, %xmm4 + movmskps %xmm1, %ecx + movaps %xmm14, %xmm15 + movaps %xmm11, %xmm9 + mulps %xmm14, %xmm15 + pxor %xmm13, %xmm9 + movups sPC8+__svml_satan2_data_internal(%rip), %xmm2 + movaps %xmm10, %xmm8 + mulps %xmm15, %xmm2 + pxor %xmm12, %xmm8 + movups sPC7+__svml_satan2_data_internal(%rip), %xmm3 + xorl %edx, %edx + mulps %xmm15, %xmm3 + addps sPC6+__svml_satan2_data_internal(%rip), %xmm2 + mulps %xmm15, %xmm2 + addps sPC5+__svml_satan2_data_internal(%rip), %xmm3 + mulps %xmm15, %xmm3 + addps sPC4+__svml_satan2_data_internal(%rip), %xmm2 + mulps %xmm15, %xmm2 + addps sPC3+__svml_satan2_data_internal(%rip), %xmm3 + mulps %xmm15, %xmm3 + addps sPC2+__svml_satan2_data_internal(%rip), %xmm2 + mulps %xmm2, %xmm15 + addps sPC1+__svml_satan2_data_internal(%rip), %xmm3 + mulps %xmm3, %xmm14 + addps sPC0+__svml_satan2_data_internal(%rip), %xmm15 + +/* if x<0, sPI = Pi, else sPI =0 */ + movups __svml_satan2_data_internal(%rip), %xmm5 + xorl %eax, %eax + andnps sPIO2+__svml_satan2_data_internal(%rip), %xmm7 + addps %xmm14, %xmm15 + cmpleps %xmm5, %xmm4 + +/* Reconstruction. */ + mulps %xmm15, %xmm0 + andps sPI+__svml_satan2_data_internal(%rip), %xmm4 + addps %xmm7, %xmm0 + orps %xmm9, %xmm0 + addps %xmm4, %xmm0 + orps %xmm8, %xmm0 + +/* Special branch for fast (vector) processing of zero arguments */ + testl %ecx, %ecx + +/* Go to auxilary branch */ + jne L(AUX_BRANCH) + # LOE rbx rbp r12 r13 r14 r15 eax edx xmm0 xmm1 xmm4 xmm5 xmm6 xmm7 xmm8 xmm9 xmm10 xmm11 xmm12 xmm13 + +/* Return from auxilary branch + * for out of main path inputs + */ + +L(AUX_BRANCH_RETURN): +/* + * Special branch for fast (vector) processing of zero arguments + * The end of implementation + */ + testl %edx, %edx + +/* Go to special inputs processing branch */ + jne L(SPECIAL_VALUES_BRANCH) + # LOE rbx rbp r12 r13 r14 r15 eax edx xmm0 xmm12 xmm13 + +/* Restore registers + * and exit the function + */ + +L(EXIT): + addq $88, %rsp + cfi_def_cfa_offset(8) + ret + cfi_def_cfa_offset(96) + +/* Branch to process + * special inputs + */ + +L(SPECIAL_VALUES_BRANCH): + movups %xmm12, 32(%rsp) + movups %xmm13, 48(%rsp) + movups %xmm0, 64(%rsp) + # LOE rbx rbp r12 r13 r14 r15 eax edx xmm0 + + movq %r12, 16(%rsp) + cfi_offset(12, -80) + movl %eax, %r12d + movq %r13, 8(%rsp) + cfi_offset(13, -88) + movl %edx, %r13d + movq %r14, (%rsp) + cfi_offset(14, -96) + # LOE rbx rbp r15 r12d r13d + +/* Range mask + * bits check + */ + +L(RANGEMASK_CHECK): + btl %r12d, %r13d + +/* Call scalar math function */ + jc L(SCALAR_MATH_CALL) + # LOE rbx rbp r15 r12d r13d + +/* Special inputs + * processing loop + */ + +L(SPECIAL_VALUES_LOOP): + incl %r12d + cmpl $4, %r12d + +/* Check bits in range mask */ + jl L(RANGEMASK_CHECK) + # LOE rbx rbp r15 r12d r13d + + movq 16(%rsp), %r12 + cfi_restore(12) + movq 8(%rsp), %r13 + cfi_restore(13) + movq (%rsp), %r14 + cfi_restore(14) + movups 64(%rsp), %xmm0 + +/* Go to exit */ + jmp L(EXIT) + cfi_offset(12, -80) + cfi_offset(13, -88) + cfi_offset(14, -96) + # LOE rbx rbp r12 r13 r14 r15 xmm0 + +/* Scalar math fucntion call + * to process special input + */ + +L(SCALAR_MATH_CALL): + movl %r12d, %r14d + movss 32(%rsp,%r14,4), %xmm0 + movss 48(%rsp,%r14,4), %xmm1 + call atan2f@PLT + # LOE rbx rbp r14 r15 r12d r13d xmm0 + + movss %xmm0, 64(%rsp,%r14,4) + +/* Process special inputs in loop */ + jmp L(SPECIAL_VALUES_LOOP) + cfi_restore(12) + cfi_restore(13) + cfi_restore(14) + # LOE rbx rbp r15 r12d r13d + +/* Auxilary branch + * for out of main path inputs + */ + +L(AUX_BRANCH): +/* Check if both X & Y are not NaNs: iXYnotNAN */ + movaps %xmm13, %xmm3 + movaps %xmm12, %xmm2 + cmpordps %xmm13, %xmm3 + cmpordps %xmm12, %xmm2 + +/* + * Path for zero arguments (at least one of both) + * Check if both args are zeros (den. is zero) + */ + cmpeqps %xmm5, %xmm6 + +/* Check if at least on of Y or Y is zero: iAXAYZERO */ + pcmpeqd %xmm5, %xmm11 + pcmpeqd %xmm5, %xmm10 + andps %xmm2, %xmm3 + por %xmm10, %xmm11 + +/* Check if at least on of Y or Y is zero and not NaN: iAXAYZEROnotNAN */ + andps %xmm3, %xmm11 + +/* Exclude from previous callout mask zero (and not NaN) arguments */ + movaps %xmm11, %xmm10 + pandn %xmm1, %xmm10 + +/* Set sPIO2 to zero if den. is zero */ + movaps %xmm6, %xmm1 + andnps %xmm7, %xmm1 + andps %xmm5, %xmm6 + orps %xmm6, %xmm1 + +/* Res = sign(Y)*(X<0)?(PIO2+PI):PIO2 */ + pcmpgtd %xmm13, %xmm5 + orps %xmm9, %xmm1 + andps %xmm4, %xmm5 + +/* Merge results from main and spec path */ + movaps %xmm11, %xmm4 + addps %xmm5, %xmm1 + +/* Go to callout */ + movmskps %xmm10, %edx + orps %xmm8, %xmm1 + andnps %xmm0, %xmm4 + andps %xmm11, %xmm1 + movaps %xmm4, %xmm0 + orps %xmm1, %xmm0 + +/* Return to main vector processing path */ + jmp L(AUX_BRANCH_RETURN) + # LOE rbx rbp r12 r13 r14 r15 eax edx xmm0 xmm12 xmm13 +END(_ZGVbN4vv_atan2f_sse4) + + .section .rodata, "a" + .align 16 + +#ifdef __svml_satan2_data_internal_typedef +typedef unsigned int VUINT32; +typedef struct { + __declspec(align(16)) VUINT32 sZERO[4][1]; + __declspec(align(16)) VUINT32 sSIGN_MASK[4][1]; + __declspec(align(16)) VUINT32 sABS_MASK[4][1]; + __declspec(align(16)) VUINT32 sPIO2[4][1]; + __declspec(align(16)) VUINT32 sPI[4][1]; + __declspec(align(16)) VUINT32 sPC8[4][1]; + __declspec(align(16)) VUINT32 sPC7[4][1]; + __declspec(align(16)) VUINT32 sPC6[4][1]; + __declspec(align(16)) VUINT32 sPC5[4][1]; + __declspec(align(16)) VUINT32 sPC4[4][1]; + __declspec(align(16)) VUINT32 sPC3[4][1]; + __declspec(align(16)) VUINT32 sPC2[4][1]; + __declspec(align(16)) VUINT32 sPC1[4][1]; + __declspec(align(16)) VUINT32 sPC0[4][1]; + __declspec(align(16)) VUINT32 iCHK_WORK_SUB[4][1]; + __declspec(align(16)) VUINT32 iCHK_WORK_CMP[4][1]; +} __svml_satan2_data_internal; +#endif +__svml_satan2_data_internal: + .long 0x00000000, 0x00000000, 0x00000000, 0x00000000 // sZERO + .align 16 + .long 0x80000000, 0x80000000, 0x80000000, 0x80000000 // sSIGN_MASK + .align 16 + .long 0x7FFFFFFF, 0x7FFFFFFF, 0x7FFFFFFF, 0x7FFFFFFF // sABS_MASK + .align 16 + .long 0x3FC90FDB, 0x3FC90FDB, 0x3FC90FDB, 0x3FC90FDB // sPIO2 + .align 16 + .long 0x40490FDB, 0x40490FDB, 0x40490FDB, 0x40490FDB // sPI + .align 16 + .long 0x3B322CC0, 0x3B322CC0, 0x3B322CC0, 0x3B322CC0 // sA08 + .align 16 + .long 0xBC7F2631, 0xBC7F2631, 0xBC7F2631, 0xBC7F2631 // sA07 + .align 16 + .long 0x3D2BC384, 0x3D2BC384, 0x3D2BC384, 0x3D2BC384 // sA06 + .align 16 + .long 0xBD987629, 0xBD987629, 0xBD987629, 0xBD987629 // sA05 + .align 16 + .long 0x3DD96474, 0x3DD96474, 0x3DD96474, 0x3DD96474 // sA04 + .align 16 + .long 0xBE1161F8, 0xBE1161F8, 0xBE1161F8, 0xBE1161F8 // sA03 + .align 16 + .long 0x3E4CB79F, 0x3E4CB79F, 0x3E4CB79F, 0x3E4CB79F // sA02 + .align 16 + .long 0xBEAAAA49, 0xBEAAAA49, 0xBEAAAA49, 0xBEAAAA49 // sA01 + .align 16 + .long 0x3F800000, 0x3F800000, 0x3F800000, 0x3F800000 // sA00 + .align 16 + .long 0x81000000, 0x81000000, 0x81000000, 0x81000000 //iCHK_WORK_SUB + .align 16 + .long 0xFC000000, 0xFC000000, 0xFC000000, 0xFC000000 //iCHK_WORK_CMP + .align 16 + .type __svml_satan2_data_internal,@object + .size __svml_satan2_data_internal,.-__svml_satan2_data_internal diff --git a/sysdeps/x86_64/fpu/multiarch/svml_s_atan2f8_core-sse.S b/sysdeps/x86_64/fpu/multiarch/svml_s_atan2f8_core-sse.S new file mode 100644 index 0000000000..21b1d3ff63 --- /dev/null +++ b/sysdeps/x86_64/fpu/multiarch/svml_s_atan2f8_core-sse.S @@ -0,0 +1,20 @@ +/* SSE version of vectorized atan2f. + Copyright (C) 2021 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, see + . */ + +#define _ZGVdN8vv_atan2f _ZGVdN8vv_atan2f_sse_wrapper +#include "../svml_s_atan2f8_core.S" diff --git a/sysdeps/x86_64/fpu/multiarch/svml_s_atan2f8_core.c b/sysdeps/x86_64/fpu/multiarch/svml_s_atan2f8_core.c new file mode 100644 index 0000000000..7e02050983 --- /dev/null +++ b/sysdeps/x86_64/fpu/multiarch/svml_s_atan2f8_core.c @@ -0,0 +1,28 @@ +/* Multiple versions of vectorized sinf, vector length is 8. + Copyright (C) 2021 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, see + . */ + +#define SYMBOL_NAME _ZGVdN8vv_atan2f +#include "ifunc-mathvec-avx2.h" + +libc_ifunc_redirected (REDIRECT_NAME, SYMBOL_NAME, IFUNC_SELECTOR ()); + +#ifdef SHARED +__hidden_ver1 (_ZGVdN8vv_atan2f, __GI__ZGVdN8vv_atan2f, + __redirect__ZGVdN8vv_atan2f) + __attribute__ ((visibility ("hidden"))); +#endif diff --git a/sysdeps/x86_64/fpu/multiarch/svml_s_atan2f8_core_avx2.S b/sysdeps/x86_64/fpu/multiarch/svml_s_atan2f8_core_avx2.S new file mode 100644 index 0000000000..2e6e5eb71c --- /dev/null +++ b/sysdeps/x86_64/fpu/multiarch/svml_s_atan2f8_core_avx2.S @@ -0,0 +1,362 @@ +/* Function atan2f vectorized with AVX2. + Copyright (C) 2021 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, see + https://www.gnu.org/licenses/. */ + +/* + * ALGORITHM DESCRIPTION: + * For 0.0 <= x <= 7.0/16.0: atan(x) = atan(0.0) + atan(s), where s=(x-0.0)/(1.0+0.0*x) + * For 7.0/16.0 <= x <= 11.0/16.0: atan(x) = atan(0.5) + atan(s), where s=(x-0.5)/(1.0+0.5*x) + * For 11.0/16.0 <= x <= 19.0/16.0: atan(x) = atan(1.0) + atan(s), where s=(x-1.0)/(1.0+1.0*x) + * For 19.0/16.0 <= x <= 39.0/16.0: atan(x) = atan(1.5) + atan(s), where s=(x-1.5)/(1.0+1.5*x) + * For 39.0/16.0 <= x <= inf : atan(x) = atan(inf) + atan(s), where s=-1.0/x + * Where atan(s) ~= s+s^3*Poly11(s^2) on interval |s|<7.0/0.16. + * + * + */ + +/* Offsets for data table __svml_satan2_data_internal + */ +#define sZERO 0 +#define sSIGN_MASK 32 +#define sABS_MASK 64 +#define sPIO2 96 +#define sPI 128 +#define sPC8 160 +#define sPC7 192 +#define sPC6 224 +#define sPC5 256 +#define sPC4 288 +#define sPC3 320 +#define sPC2 352 +#define sPC1 384 +#define sPC0 416 +#define iCHK_WORK_SUB 448 +#define iCHK_WORK_CMP 480 + +#include + + .text + .section .text.avx2,"ax",@progbits +ENTRY(_ZGVdN8vv_atan2f_avx2) + pushq %rbp + cfi_def_cfa_offset(16) + movq %rsp, %rbp + cfi_def_cfa(6, 16) + cfi_offset(6, -16) + andq $-32, %rsp + subq $128, %rsp + xorl %edx, %edx + +/* + * #define NO_VECTOR_ZERO_ATAN2_ARGS + * Declarations + * Variables + * Constants + * The end of declarations + * Implementation + * Arguments signs + */ + vmovups sABS_MASK+__svml_satan2_data_internal(%rip), %ymm2 + +/* Testing on working interval. */ + vmovups iCHK_WORK_SUB+__svml_satan2_data_internal(%rip), %ymm15 + vmovups iCHK_WORK_CMP+__svml_satan2_data_internal(%rip), %ymm9 + +/* if x<0, sPI = Pi, else sPI =0 */ + vmovups __svml_satan2_data_internal(%rip), %ymm5 + vmovaps %ymm1, %ymm7 + vandps %ymm2, %ymm7, %ymm13 + vandps %ymm2, %ymm0, %ymm12 + vcmplt_oqps %ymm13, %ymm12, %ymm4 + vcmple_oqps %ymm5, %ymm7, %ymm6 + vpsubd %ymm15, %ymm13, %ymm10 + vpsubd %ymm15, %ymm12, %ymm8 + +/* + * 1) If yx then a=-x, b=y, PIO2=Pi/2 + */ + vorps sSIGN_MASK+__svml_satan2_data_internal(%rip), %ymm13, %ymm3 + vblendvps %ymm4, %ymm12, %ymm3, %ymm14 + vblendvps %ymm4, %ymm13, %ymm12, %ymm3 + +/* Division a/b. */ + vdivps %ymm3, %ymm14, %ymm11 + vpcmpgtd %ymm9, %ymm10, %ymm14 + vpcmpeqd %ymm9, %ymm10, %ymm15 + vpor %ymm15, %ymm14, %ymm10 + vmovups sPC7+__svml_satan2_data_internal(%rip), %ymm15 + vpcmpgtd %ymm9, %ymm8, %ymm14 + vpcmpeqd %ymm9, %ymm8, %ymm8 + vpor %ymm8, %ymm14, %ymm9 + vmovups sPC8+__svml_satan2_data_internal(%rip), %ymm14 + vpor %ymm9, %ymm10, %ymm10 + +/* Polynomial. */ + vmulps %ymm11, %ymm11, %ymm9 + vmulps %ymm9, %ymm9, %ymm8 + vfmadd213ps sPC6+__svml_satan2_data_internal(%rip), %ymm8, %ymm14 + vfmadd213ps sPC5+__svml_satan2_data_internal(%rip), %ymm8, %ymm15 + vfmadd213ps sPC4+__svml_satan2_data_internal(%rip), %ymm8, %ymm14 + vfmadd213ps sPC3+__svml_satan2_data_internal(%rip), %ymm8, %ymm15 + vfmadd213ps sPC2+__svml_satan2_data_internal(%rip), %ymm8, %ymm14 + vfmadd213ps sPC1+__svml_satan2_data_internal(%rip), %ymm8, %ymm15 + vfmadd213ps sPC0+__svml_satan2_data_internal(%rip), %ymm8, %ymm14 + vfmadd213ps %ymm14, %ymm9, %ymm15 + vandnps sPIO2+__svml_satan2_data_internal(%rip), %ymm4, %ymm4 + +/* Reconstruction. */ + vfmadd213ps %ymm4, %ymm11, %ymm15 + vxorps %ymm13, %ymm7, %ymm1 + vandps sPI+__svml_satan2_data_internal(%rip), %ymm6, %ymm6 + vorps %ymm1, %ymm15, %ymm11 + vaddps %ymm11, %ymm6, %ymm8 + vmovmskps %ymm10, %eax + vxorps %ymm12, %ymm0, %ymm2 + vorps %ymm2, %ymm8, %ymm9 + +/* Special branch for fast (vector) processing of zero arguments */ + testl %eax, %eax + +/* Go to auxilary branch */ + jne L(AUX_BRANCH) + # LOE rbx r12 r13 r14 r15 edx ymm0 ymm1 ymm2 ymm3 ymm4 ymm5 ymm6 ymm7 ymm9 ymm10 ymm12 ymm13 + +/* Return from auxilary branch + * for out of main path inputs + */ + +L(AUX_BRANCH_RETURN): +/* + * Special branch for fast (vector) processing of zero arguments + * The end of implementation + */ + testl %edx, %edx + +/* Go to special inputs processing branch */ + jne L(SPECIAL_VALUES_BRANCH) + # LOE rbx r12 r13 r14 r15 edx ymm0 ymm7 ymm9 + +/* Restore registers + * and exit the function + */ + +L(EXIT): + vmovaps %ymm9, %ymm0 + movq %rbp, %rsp + popq %rbp + cfi_def_cfa(7, 8) + cfi_restore(6) + ret + cfi_def_cfa(6, 16) + cfi_offset(6, -16) + +/* Branch to process + * special inputs + */ + +L(SPECIAL_VALUES_BRANCH): + vmovups %ymm0, 32(%rsp) + vmovups %ymm7, 64(%rsp) + vmovups %ymm9, 96(%rsp) + # LOE rbx r12 r13 r14 r15 edx ymm9 + + xorl %eax, %eax + # LOE rbx r12 r13 r14 r15 eax edx + + vzeroupper + movq %r12, 16(%rsp) + /* DW_CFA_expression: r12 (r12) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -112; DW_OP_plus) */ + .cfi_escape 0x10, 0x0c, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x90, 0xff, 0xff, 0xff, 0x22 + movl %eax, %r12d + movq %r13, 8(%rsp) + /* DW_CFA_expression: r13 (r13) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -120; DW_OP_plus) */ + .cfi_escape 0x10, 0x0d, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x88, 0xff, 0xff, 0xff, 0x22 + movl %edx, %r13d + movq %r14, (%rsp) + /* DW_CFA_expression: r14 (r14) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -128; DW_OP_plus) */ + .cfi_escape 0x10, 0x0e, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x80, 0xff, 0xff, 0xff, 0x22 + # LOE rbx r15 r12d r13d + +/* Range mask + * bits check + */ + +L(RANGEMASK_CHECK): + btl %r12d, %r13d + +/* Call scalar math function */ + jc L(SCALAR_MATH_CALL) + # LOE rbx r15 r12d r13d + +/* Special inputs + * processing loop + */ + +L(SPECIAL_VALUES_LOOP): + incl %r12d + cmpl $8, %r12d + +/* Check bits in range mask */ + jl L(RANGEMASK_CHECK) + # LOE rbx r15 r12d r13d + + movq 16(%rsp), %r12 + cfi_restore(12) + movq 8(%rsp), %r13 + cfi_restore(13) + movq (%rsp), %r14 + cfi_restore(14) + vmovups 96(%rsp), %ymm9 + +/* Go to exit */ + jmp L(EXIT) + /* DW_CFA_expression: r12 (r12) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -112; DW_OP_plus) */ + .cfi_escape 0x10, 0x0c, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x90, 0xff, 0xff, 0xff, 0x22 + /* DW_CFA_expression: r13 (r13) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -120; DW_OP_plus) */ + .cfi_escape 0x10, 0x0d, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x88, 0xff, 0xff, 0xff, 0x22 + /* DW_CFA_expression: r14 (r14) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -128; DW_OP_plus) */ + .cfi_escape 0x10, 0x0e, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x80, 0xff, 0xff, 0xff, 0x22 + # LOE rbx r12 r13 r14 r15 ymm9 + +/* Scalar math fucntion call + * to process special input + */ + +L(SCALAR_MATH_CALL): + movl %r12d, %r14d + movss 32(%rsp,%r14,4), %xmm0 + movss 64(%rsp,%r14,4), %xmm1 + call atan2f@PLT + # LOE rbx r14 r15 r12d r13d xmm0 + + movss %xmm0, 96(%rsp,%r14,4) + +/* Process special inputs in loop */ + jmp L(SPECIAL_VALUES_LOOP) + cfi_restore(12) + cfi_restore(13) + cfi_restore(14) + # LOE rbx r15 r12d r13d + +/* Auxilary branch + * for out of main path inputs + */ + +L(AUX_BRANCH): +/* Check if at least on of Y or Y is zero: iAXAYZERO */ + vpcmpeqd %ymm5, %ymm13, %ymm13 + vpcmpeqd %ymm5, %ymm12, %ymm12 + +/* Check if both X & Y are not NaNs: iXYnotNAN */ + vcmpordps %ymm7, %ymm7, %ymm11 + vcmpordps %ymm0, %ymm0, %ymm14 + +/* + * Path for zero arguments (at least one of both) + * Check if both args are zeros (den. is zero) + */ + vcmpeqps %ymm5, %ymm3, %ymm3 + vpor %ymm12, %ymm13, %ymm15 + +/* Set sPIO2 to zero if den. is zero */ + vblendvps %ymm3, %ymm5, %ymm4, %ymm4 + vandps %ymm14, %ymm11, %ymm8 + +/* Check if at least on of Y or Y is zero and not NaN: iAXAYZEROnotNAN */ + vpand %ymm8, %ymm15, %ymm8 + +/* Res = sign(Y)*(X<0)?(PIO2+PI):PIO2 */ + vpcmpgtd %ymm7, %ymm5, %ymm5 + vorps %ymm1, %ymm4, %ymm1 + vandps %ymm6, %ymm5, %ymm6 + vaddps %ymm6, %ymm1, %ymm1 + +/* Exclude from previous callout mask zero (and not NaN) arguments */ + vpandn %ymm10, %ymm8, %ymm10 + vorps %ymm2, %ymm1, %ymm2 + +/* Go to callout */ + vmovmskps %ymm10, %edx + +/* Merge results from main and spec path */ + vblendvps %ymm8, %ymm2, %ymm9, %ymm9 + +/* Return to main vector processing path */ + jmp L(AUX_BRANCH_RETURN) + # LOE rbx r12 r13 r14 r15 edx ymm0 ymm7 ymm9 +END(_ZGVdN8vv_atan2f_avx2) + + .section .rodata, "a" + .align 32 + +#ifdef __svml_satan2_data_internal_typedef +typedef unsigned int VUINT32; +typedef struct { + __declspec(align(32)) VUINT32 sZERO[8][1]; + __declspec(align(32)) VUINT32 sSIGN_MASK[8][1]; + __declspec(align(32)) VUINT32 sABS_MASK[8][1]; + __declspec(align(32)) VUINT32 sPIO2[8][1]; + __declspec(align(32)) VUINT32 sPI[8][1]; + __declspec(align(32)) VUINT32 sPC8[8][1]; + __declspec(align(32)) VUINT32 sPC7[8][1]; + __declspec(align(32)) VUINT32 sPC6[8][1]; + __declspec(align(32)) VUINT32 sPC5[8][1]; + __declspec(align(32)) VUINT32 sPC4[8][1]; + __declspec(align(32)) VUINT32 sPC3[8][1]; + __declspec(align(32)) VUINT32 sPC2[8][1]; + __declspec(align(32)) VUINT32 sPC1[8][1]; + __declspec(align(32)) VUINT32 sPC0[8][1]; + __declspec(align(32)) VUINT32 iCHK_WORK_SUB[8][1]; + __declspec(align(32)) VUINT32 iCHK_WORK_CMP[8][1]; +} __svml_satan2_data_internal; +#endif +__svml_satan2_data_internal: + .long 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000, 0x00000000 // sZERO + .align 32 + .long 0x80000000, 0x80000000, 0x80000000, 0x80000000, 0x80000000, 0x80000000, 0x80000000, 0x80000000 // sSIGN_MASK + .align 32 + .long 0x7FFFFFFF, 0x7FFFFFFF, 0x7FFFFFFF, 0x7FFFFFFF, 0x7FFFFFFF, 0x7FFFFFFF, 0x7FFFFFFF, 0x7FFFFFFF // sABS_MASK + .align 32 + .long 0x3FC90FDB, 0x3FC90FDB, 0x3FC90FDB, 0x3FC90FDB, 0x3FC90FDB, 0x3FC90FDB, 0x3FC90FDB, 0x3FC90FDB // sPIO2 + .align 32 + .long 0x40490FDB, 0x40490FDB, 0x40490FDB, 0x40490FDB, 0x40490FDB, 0x40490FDB, 0x40490FDB, 0x40490FDB // sPI + .align 32 + .long 0x3B322CC0, 0x3B322CC0, 0x3B322CC0, 0x3B322CC0, 0x3B322CC0, 0x3B322CC0, 0x3B322CC0, 0x3B322CC0 // sA08 + .align 32 + .long 0xBC7F2631, 0xBC7F2631, 0xBC7F2631, 0xBC7F2631, 0xBC7F2631, 0xBC7F2631, 0xBC7F2631, 0xBC7F2631 // sA07 + .align 32 + .long 0x3D2BC384, 0x3D2BC384, 0x3D2BC384, 0x3D2BC384, 0x3D2BC384, 0x3D2BC384, 0x3D2BC384, 0x3D2BC384 // sA06 + .align 32 + .long 0xBD987629, 0xBD987629, 0xBD987629, 0xBD987629, 0xBD987629, 0xBD987629, 0xBD987629, 0xBD987629 // sA05 + .align 32 + .long 0x3DD96474, 0x3DD96474, 0x3DD96474, 0x3DD96474, 0x3DD96474, 0x3DD96474, 0x3DD96474, 0x3DD96474 // sA04 + .align 32 + .long 0xBE1161F8, 0xBE1161F8, 0xBE1161F8, 0xBE1161F8, 0xBE1161F8, 0xBE1161F8, 0xBE1161F8, 0xBE1161F8 // sA03 + .align 32 + .long 0x3E4CB79F, 0x3E4CB79F, 0x3E4CB79F, 0x3E4CB79F, 0x3E4CB79F, 0x3E4CB79F, 0x3E4CB79F, 0x3E4CB79F // sA02 + .align 32 + .long 0xBEAAAA49, 0xBEAAAA49, 0xBEAAAA49, 0xBEAAAA49, 0xBEAAAA49, 0xBEAAAA49, 0xBEAAAA49, 0xBEAAAA49 // sA01 + .align 32 + .long 0x3F800000, 0x3F800000, 0x3F800000, 0x3F800000, 0x3F800000, 0x3F800000, 0x3F800000, 0x3F800000 // sA00 + .align 32 + .long 0x81000000, 0x81000000, 0x81000000, 0x81000000, 0x81000000, 0x81000000, 0x81000000, 0x81000000 //iCHK_WORK_SUB + .align 32 + .long 0xFC000000, 0xFC000000, 0xFC000000, 0xFC000000, 0xFC000000, 0xFC000000, 0xFC000000, 0xFC000000 //iCHK_WORK_CMP + .align 32 + .type __svml_satan2_data_internal,@object + .size __svml_satan2_data_internal,.-__svml_satan2_data_internal diff --git a/sysdeps/x86_64/fpu/svml_d_atan22_core.S b/sysdeps/x86_64/fpu/svml_d_atan22_core.S new file mode 100644 index 0000000000..f3089e70f9 --- /dev/null +++ b/sysdeps/x86_64/fpu/svml_d_atan22_core.S @@ -0,0 +1,29 @@ +/* Function atan2 vectorized with SSE2. + Copyright (C) 2021 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, see + . */ + +#include +#include "svml_d_wrapper_impl.h" + + .text +ENTRY (_ZGVbN2vv_atan2) +WRAPPER_IMPL_SSE2_ff atan2 +END (_ZGVbN2vv_atan2) + +#ifndef USE_MULTIARCH + libmvec_hidden_def (_ZGVbN2vv_atan2) +#endif diff --git a/sysdeps/x86_64/fpu/svml_d_atan24_core.S b/sysdeps/x86_64/fpu/svml_d_atan24_core.S new file mode 100644 index 0000000000..8a163d12d2 --- /dev/null +++ b/sysdeps/x86_64/fpu/svml_d_atan24_core.S @@ -0,0 +1,29 @@ +/* Function atan2 vectorized with AVX2, wrapper version. + Copyright (C) 2021 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, see + . */ + +#include +#include "svml_d_wrapper_impl.h" + + .text +ENTRY (_ZGVdN4vv_atan2) +WRAPPER_IMPL_AVX_ff _ZGVbN2vv_atan2 +END (_ZGVdN4vv_atan2) + +#ifndef USE_MULTIARCH + libmvec_hidden_def (_ZGVdN4vv_atan2) +#endif diff --git a/sysdeps/x86_64/fpu/svml_d_atan24_core_avx.S b/sysdeps/x86_64/fpu/svml_d_atan24_core_avx.S new file mode 100644 index 0000000000..0ee5ae8faf --- /dev/null +++ b/sysdeps/x86_64/fpu/svml_d_atan24_core_avx.S @@ -0,0 +1,25 @@ +/* Function atan2 vectorized in AVX ISA as wrapper to SSE4 ISA version. + Copyright (C) 2021 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, see + . */ + +#include +#include "svml_d_wrapper_impl.h" + + .text +ENTRY (_ZGVcN4vv_atan2) +WRAPPER_IMPL_AVX_ff _ZGVbN2vv_atan2 +END (_ZGVcN4vv_atan2) diff --git a/sysdeps/x86_64/fpu/svml_d_atan28_core.S b/sysdeps/x86_64/fpu/svml_d_atan28_core.S new file mode 100644 index 0000000000..b85f696686 --- /dev/null +++ b/sysdeps/x86_64/fpu/svml_d_atan28_core.S @@ -0,0 +1,25 @@ +/* Function atan2 vectorized with AVX-512. Wrapper to AVX2 version. + Copyright (C) 2021 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, see + . */ + +#include +#include "svml_d_wrapper_impl.h" + + .text +ENTRY (_ZGVeN8vv_atan2) +WRAPPER_IMPL_AVX512_ff _ZGVdN4vv_atan2 +END (_ZGVeN8vv_atan2) diff --git a/sysdeps/x86_64/fpu/svml_s_atan2f16_core.S b/sysdeps/x86_64/fpu/svml_s_atan2f16_core.S new file mode 100644 index 0000000000..25acb31dfb --- /dev/null +++ b/sysdeps/x86_64/fpu/svml_s_atan2f16_core.S @@ -0,0 +1,25 @@ +/* Function atan2f vectorized with AVX-512. Wrapper to AVX2 version. + Copyright (C) 2021 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, see + . */ + +#include +#include "svml_s_wrapper_impl.h" + + .text +ENTRY (_ZGVeN16vv_atan2f) +WRAPPER_IMPL_AVX512_ff _ZGVdN8vv_atan2f +END (_ZGVeN16vv_atan2f) diff --git a/sysdeps/x86_64/fpu/svml_s_atan2f4_core.S b/sysdeps/x86_64/fpu/svml_s_atan2f4_core.S new file mode 100644 index 0000000000..bc99f0ba10 --- /dev/null +++ b/sysdeps/x86_64/fpu/svml_s_atan2f4_core.S @@ -0,0 +1,29 @@ +/* Function atan2f vectorized with SSE2. + Copyright (C) 2021 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, see + . */ + +#include +#include "svml_s_wrapper_impl.h" + + .text +ENTRY (_ZGVbN4vv_atan2f) +WRAPPER_IMPL_SSE2_ff atan2f +END (_ZGVbN4vv_atan2f) + +#ifndef USE_MULTIARCH + libmvec_hidden_def (_ZGVbN4vv_atan2f) +#endif diff --git a/sysdeps/x86_64/fpu/svml_s_atan2f8_core.S b/sysdeps/x86_64/fpu/svml_s_atan2f8_core.S new file mode 100644 index 0000000000..bfcdb3c372 --- /dev/null +++ b/sysdeps/x86_64/fpu/svml_s_atan2f8_core.S @@ -0,0 +1,29 @@ +/* Function atan2f vectorized with AVX2, wrapper version. + Copyright (C) 2021 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, see + . */ + +#include +#include "svml_s_wrapper_impl.h" + + .text +ENTRY (_ZGVdN8vv_atan2f) +WRAPPER_IMPL_AVX_ff _ZGVbN4vv_atan2f +END (_ZGVdN8vv_atan2f) + +#ifndef USE_MULTIARCH + libmvec_hidden_def (_ZGVdN8vv_atan2f) +#endif diff --git a/sysdeps/x86_64/fpu/svml_s_atan2f8_core_avx.S b/sysdeps/x86_64/fpu/svml_s_atan2f8_core_avx.S new file mode 100644 index 0000000000..1aa8d05822 --- /dev/null +++ b/sysdeps/x86_64/fpu/svml_s_atan2f8_core_avx.S @@ -0,0 +1,25 @@ +/* Function atan2f vectorized in AVX ISA as wrapper to SSE4 ISA version. + Copyright (C) 2021 Free Software Foundation, Inc. + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, see + . */ + +#include +#include "svml_s_wrapper_impl.h" + + .text +ENTRY(_ZGVcN8vv_atan2f) +WRAPPER_IMPL_AVX_ff _ZGVbN4vv_atan2f +END(_ZGVcN8vv_atan2f) diff --git a/sysdeps/x86_64/fpu/test-double-libmvec-atan2-avx.c b/sysdeps/x86_64/fpu/test-double-libmvec-atan2-avx.c new file mode 100644 index 0000000000..e423bce25b --- /dev/null +++ b/sysdeps/x86_64/fpu/test-double-libmvec-atan2-avx.c @@ -0,0 +1 @@ +#include "test-double-libmvec-atan2.c" diff --git a/sysdeps/x86_64/fpu/test-double-libmvec-atan2-avx2.c b/sysdeps/x86_64/fpu/test-double-libmvec-atan2-avx2.c new file mode 100644 index 0000000000..e423bce25b --- /dev/null +++ b/sysdeps/x86_64/fpu/test-double-libmvec-atan2-avx2.c @@ -0,0 +1 @@ +#include "test-double-libmvec-atan2.c" diff --git a/sysdeps/x86_64/fpu/test-double-libmvec-atan2-avx512f.c b/sysdeps/x86_64/fpu/test-double-libmvec-atan2-avx512f.c new file mode 100644 index 0000000000..e423bce25b --- /dev/null +++ b/sysdeps/x86_64/fpu/test-double-libmvec-atan2-avx512f.c @@ -0,0 +1 @@ +#include "test-double-libmvec-atan2.c" diff --git a/sysdeps/x86_64/fpu/test-double-libmvec-atan2.c b/sysdeps/x86_64/fpu/test-double-libmvec-atan2.c new file mode 100644 index 0000000000..d0aa626d95 --- /dev/null +++ b/sysdeps/x86_64/fpu/test-double-libmvec-atan2.c @@ -0,0 +1,3 @@ +#define LIBMVEC_TYPE double +#define LIBMVEC_FUNC atan2 +#include "test-vector-abi-arg2.h" diff --git a/sysdeps/x86_64/fpu/test-double-vlen2-wrappers.c b/sysdeps/x86_64/fpu/test-double-vlen2-wrappers.c index b1981ac7e4..37a7a1c777 100644 --- a/sysdeps/x86_64/fpu/test-double-vlen2-wrappers.c +++ b/sysdeps/x86_64/fpu/test-double-vlen2-wrappers.c @@ -37,6 +37,7 @@ VECTOR_WRAPPER (WRAPPER_NAME (cosh), _ZGVbN2v_cosh) VECTOR_WRAPPER (WRAPPER_NAME (expm1), _ZGVbN2v_expm1) VECTOR_WRAPPER (WRAPPER_NAME (sinh), _ZGVbN2v_sinh) VECTOR_WRAPPER (WRAPPER_NAME (cbrt), _ZGVbN2v_cbrt) +VECTOR_WRAPPER_ff (WRAPPER_NAME (atan2), _ZGVbN2vv_atan2) #define VEC_INT_TYPE __m128i diff --git a/sysdeps/x86_64/fpu/test-double-vlen4-avx2-wrappers.c b/sysdeps/x86_64/fpu/test-double-vlen4-avx2-wrappers.c index 47915a7e59..4313f67e06 100644 --- a/sysdeps/x86_64/fpu/test-double-vlen4-avx2-wrappers.c +++ b/sysdeps/x86_64/fpu/test-double-vlen4-avx2-wrappers.c @@ -40,6 +40,7 @@ VECTOR_WRAPPER (WRAPPER_NAME (cosh), _ZGVdN4v_cosh) VECTOR_WRAPPER (WRAPPER_NAME (expm1), _ZGVdN4v_expm1) VECTOR_WRAPPER (WRAPPER_NAME (sinh), _ZGVdN4v_sinh) VECTOR_WRAPPER (WRAPPER_NAME (cbrt), _ZGVdN4v_cbrt) +VECTOR_WRAPPER_ff (WRAPPER_NAME (atan2), _ZGVdN4vv_atan2) #ifndef __ILP32__ # define VEC_INT_TYPE __m256i diff --git a/sysdeps/x86_64/fpu/test-double-vlen4-wrappers.c b/sysdeps/x86_64/fpu/test-double-vlen4-wrappers.c index 5cd5049807..4b8b00f16d 100644 --- a/sysdeps/x86_64/fpu/test-double-vlen4-wrappers.c +++ b/sysdeps/x86_64/fpu/test-double-vlen4-wrappers.c @@ -37,6 +37,7 @@ VECTOR_WRAPPER (WRAPPER_NAME (cosh), _ZGVcN4v_cosh) VECTOR_WRAPPER (WRAPPER_NAME (expm1), _ZGVcN4v_expm1) VECTOR_WRAPPER (WRAPPER_NAME (sinh), _ZGVcN4v_sinh) VECTOR_WRAPPER (WRAPPER_NAME (cbrt), _ZGVcN4v_cbrt) +VECTOR_WRAPPER_ff (WRAPPER_NAME (atan2), _ZGVcN4vv_atan2) #define VEC_INT_TYPE __m128i diff --git a/sysdeps/x86_64/fpu/test-double-vlen8-wrappers.c b/sysdeps/x86_64/fpu/test-double-vlen8-wrappers.c index 83970739ab..d06522a407 100644 --- a/sysdeps/x86_64/fpu/test-double-vlen8-wrappers.c +++ b/sysdeps/x86_64/fpu/test-double-vlen8-wrappers.c @@ -37,6 +37,7 @@ VECTOR_WRAPPER (WRAPPER_NAME (cosh), _ZGVeN8v_cosh) VECTOR_WRAPPER (WRAPPER_NAME (expm1), _ZGVeN8v_expm1) VECTOR_WRAPPER (WRAPPER_NAME (sinh), _ZGVeN8v_sinh) VECTOR_WRAPPER (WRAPPER_NAME (cbrt), _ZGVeN8v_cbrt) +VECTOR_WRAPPER_ff (WRAPPER_NAME (atan2), _ZGVeN8vv_atan2) #ifndef __ILP32__ # define VEC_INT_TYPE __m512i diff --git a/sysdeps/x86_64/fpu/test-float-libmvec-atan2f-avx.c b/sysdeps/x86_64/fpu/test-float-libmvec-atan2f-avx.c new file mode 100644 index 0000000000..5c7e2c9ad5 --- /dev/null +++ b/sysdeps/x86_64/fpu/test-float-libmvec-atan2f-avx.c @@ -0,0 +1 @@ +#include "test-float-libmvec-atan2f.c" diff --git a/sysdeps/x86_64/fpu/test-float-libmvec-atan2f-avx2.c b/sysdeps/x86_64/fpu/test-float-libmvec-atan2f-avx2.c new file mode 100644 index 0000000000..5c7e2c9ad5 --- /dev/null +++ b/sysdeps/x86_64/fpu/test-float-libmvec-atan2f-avx2.c @@ -0,0 +1 @@ +#include "test-float-libmvec-atan2f.c" diff --git a/sysdeps/x86_64/fpu/test-float-libmvec-atan2f-avx512f.c b/sysdeps/x86_64/fpu/test-float-libmvec-atan2f-avx512f.c new file mode 100644 index 0000000000..5c7e2c9ad5 --- /dev/null +++ b/sysdeps/x86_64/fpu/test-float-libmvec-atan2f-avx512f.c @@ -0,0 +1 @@ +#include "test-float-libmvec-atan2f.c" diff --git a/sysdeps/x86_64/fpu/test-float-libmvec-atan2f.c b/sysdeps/x86_64/fpu/test-float-libmvec-atan2f.c new file mode 100644 index 0000000000..beb5c745cb --- /dev/null +++ b/sysdeps/x86_64/fpu/test-float-libmvec-atan2f.c @@ -0,0 +1,3 @@ +#define LIBMVEC_TYPE float +#define LIBMVEC_FUNC atan2f +#include "test-vector-abi-arg2.h" diff --git a/sysdeps/x86_64/fpu/test-float-vlen16-wrappers.c b/sysdeps/x86_64/fpu/test-float-vlen16-wrappers.c index 0420f11c28..0bd631bf9a 100644 --- a/sysdeps/x86_64/fpu/test-float-vlen16-wrappers.c +++ b/sysdeps/x86_64/fpu/test-float-vlen16-wrappers.c @@ -37,6 +37,7 @@ VECTOR_WRAPPER (WRAPPER_NAME (coshf), _ZGVeN16v_coshf) VECTOR_WRAPPER (WRAPPER_NAME (expm1f), _ZGVeN16v_expm1f) VECTOR_WRAPPER (WRAPPER_NAME (sinhf), _ZGVeN16v_sinhf) VECTOR_WRAPPER (WRAPPER_NAME (cbrtf), _ZGVeN16v_cbrtf) +VECTOR_WRAPPER_ff (WRAPPER_NAME (atan2f), _ZGVeN16vv_atan2f) #define VEC_INT_TYPE __m512i diff --git a/sysdeps/x86_64/fpu/test-float-vlen4-wrappers.c b/sysdeps/x86_64/fpu/test-float-vlen4-wrappers.c index c8f7580265..1018398bd3 100644 --- a/sysdeps/x86_64/fpu/test-float-vlen4-wrappers.c +++ b/sysdeps/x86_64/fpu/test-float-vlen4-wrappers.c @@ -37,6 +37,7 @@ VECTOR_WRAPPER (WRAPPER_NAME (coshf), _ZGVbN4v_coshf) VECTOR_WRAPPER (WRAPPER_NAME (expm1f), _ZGVbN4v_expm1f) VECTOR_WRAPPER (WRAPPER_NAME (sinhf), _ZGVbN4v_sinhf) VECTOR_WRAPPER (WRAPPER_NAME (cbrtf), _ZGVbN4v_cbrtf) +VECTOR_WRAPPER_ff (WRAPPER_NAME (atan2f), _ZGVbN4vv_atan2f) #define VEC_INT_TYPE __m128i diff --git a/sysdeps/x86_64/fpu/test-float-vlen8-avx2-wrappers.c b/sysdeps/x86_64/fpu/test-float-vlen8-avx2-wrappers.c index b581796b88..42ea28f30f 100644 --- a/sysdeps/x86_64/fpu/test-float-vlen8-avx2-wrappers.c +++ b/sysdeps/x86_64/fpu/test-float-vlen8-avx2-wrappers.c @@ -40,6 +40,7 @@ VECTOR_WRAPPER (WRAPPER_NAME (coshf), _ZGVdN8v_coshf) VECTOR_WRAPPER (WRAPPER_NAME (expm1f), _ZGVdN8v_expm1f) VECTOR_WRAPPER (WRAPPER_NAME (sinhf), _ZGVdN8v_sinhf) VECTOR_WRAPPER (WRAPPER_NAME (cbrtf), _ZGVdN8v_cbrtf) +VECTOR_WRAPPER_ff (WRAPPER_NAME (atan2f), _ZGVdN8vv_atan2f) /* Redefinition of wrapper to be compatible with _ZGVdN8vvv_sincosf. */ #undef VECTOR_WRAPPER_fFF diff --git a/sysdeps/x86_64/fpu/test-float-vlen8-wrappers.c b/sysdeps/x86_64/fpu/test-float-vlen8-wrappers.c index f16789e5ff..70a0216a07 100644 --- a/sysdeps/x86_64/fpu/test-float-vlen8-wrappers.c +++ b/sysdeps/x86_64/fpu/test-float-vlen8-wrappers.c @@ -37,6 +37,7 @@ VECTOR_WRAPPER (WRAPPER_NAME (coshf), _ZGVcN8v_coshf) VECTOR_WRAPPER (WRAPPER_NAME (expm1f), _ZGVcN8v_expm1f) VECTOR_WRAPPER (WRAPPER_NAME (sinhf), _ZGVcN8v_sinhf) VECTOR_WRAPPER (WRAPPER_NAME (cbrtf), _ZGVcN8v_cbrtf) +VECTOR_WRAPPER_ff (WRAPPER_NAME (atan2f), _ZGVcN8vv_atan2f) #define VEC_INT_TYPE __m128i