@@ -1157,3 +1157,6 @@ DEF_HELPER_FLAGS_2(cm_jalt, TCG_CALL_NO_WG, tl, env, i32)
/* BF16 functions */
DEF_HELPER_FLAGS_2(fcvt_bf16_s, TCG_CALL_NO_RWG, i64, env, i64)
DEF_HELPER_FLAGS_2(fcvt_s_bf16, TCG_CALL_NO_RWG, i64, env, i64)
+
+DEF_HELPER_5(vfncvtbf16_f_f_w, void, ptr, ptr, ptr, env, i32)
+DEF_HELPER_5(vfwcvtbf16_f_f_v, void, ptr, ptr, ptr, env, i32)
@@ -912,3 +912,7 @@ czero_nez 0000111 ..... ..... 111 ..... 0110011 @r
# *** Zfbfmin Standard Extension ***
fcvt_bf16_s 0100010 01000 ..... ... ..... 1010011 @r2_rm
fcvt_s_bf16 0100000 00110 ..... ... ..... 1010011 @r2_rm
+
+# *** Zvfbfmin Standard Extension ***
+vfncvtbf16_f_f_w 010010 . ..... 11101 001 ..... 1010111 @r2_vm
+vfwcvtbf16_f_f_v 010010 . ..... 01101 001 ..... 1010111 @r2_vm
@@ -4535,6 +4535,9 @@ RVVCALL(OPFVV1, vfwcvt_f_f_v_w, WOP_UU_W, H8, H4, float32_to_float64)
GEN_VEXT_V_ENV(vfwcvt_f_f_v_h, 4)
GEN_VEXT_V_ENV(vfwcvt_f_f_v_w, 8)
+RVVCALL(OPFVV1, vfwcvtbf16_f_f_v, WOP_UU_H, H4, H2, bfloat16_to_float32)
+GEN_VEXT_V_ENV(vfwcvtbf16_f_f_v, 4)
+
/* Narrowing Floating-Point/Integer Type-Convert Instructions */
/* (TD, T2, TX2) */
#define NOP_UU_B uint8_t, uint16_t, uint32_t
@@ -4581,6 +4584,9 @@ RVVCALL(OPFVV1, vfncvt_f_f_w_w, NOP_UU_W, H4, H8, float64_to_float32)
GEN_VEXT_V_ENV(vfncvt_f_f_w_h, 2)
GEN_VEXT_V_ENV(vfncvt_f_f_w_w, 4)
+RVVCALL(OPFVV1, vfncvtbf16_f_f_w, NOP_UU_H, H2, H4, float32_to_bfloat16)
+GEN_VEXT_V_ENV(vfncvtbf16_f_f_w, 2)
+
/*
* Vector Reduction Operations
*/
@@ -22,6 +22,12 @@
} \
} while (0)
+#define REQUIRE_ZVFBFMIN(ctx) do { \
+ if (!ctx->cfg_ptr->ext_zvfbfmin) { \
+ return false; \
+ } \
+} while (0)
+
static bool trans_fcvt_bf16_s(DisasContext *ctx, arg_fcvt_bf16_s *a)
{
REQUIRE_FPU;
@@ -51,3 +57,61 @@ static bool trans_fcvt_s_bf16(DisasContext *ctx, arg_fcvt_s_bf16 *a)
mark_fs_dirty(ctx);
return true;
}
+
+static bool trans_vfncvtbf16_f_f_w(DisasContext *ctx, arg_vfncvtbf16_f_f_w *a)
+{
+ REQUIRE_FPU;
+ REQUIRE_ZVFBFMIN(ctx);
+
+ if (opfv_narrow_check(ctx, a) && (ctx->sew == MO_16)) {
+ uint32_t data = 0;
+ TCGLabel *over = gen_new_label();
+
+ gen_set_rm_chkfrm(ctx, RISCV_FRM_DYN);
+ tcg_gen_brcondi_tl(TCG_COND_EQ, cpu_vl, 0, over);
+ tcg_gen_brcond_tl(TCG_COND_GEU, cpu_vstart, cpu_vl, over);
+
+ data = FIELD_DP32(data, VDATA, VM, a->vm);
+ data = FIELD_DP32(data, VDATA, LMUL, ctx->lmul);
+ data = FIELD_DP32(data, VDATA, VTA, ctx->vta);
+ data = FIELD_DP32(data, VDATA, VMA, ctx->vma);
+ tcg_gen_gvec_3_ptr(vreg_ofs(ctx, a->rd), vreg_ofs(ctx, 0),
+ vreg_ofs(ctx, a->rs2), cpu_env,
+ ctx->cfg_ptr->vlen / 8,
+ ctx->cfg_ptr->vlen / 8, data,
+ gen_helper_vfncvtbf16_f_f_w);
+ mark_vs_dirty(ctx);
+ gen_set_label(over);
+ return true;
+ }
+ return false;
+}
+
+static bool trans_vfwcvtbf16_f_f_v(DisasContext *ctx, arg_vfwcvtbf16_f_f_v *a)
+{
+ REQUIRE_FPU;
+ REQUIRE_ZVFBFMIN(ctx);
+
+ if (opfv_widen_check(ctx, a) && (ctx->sew == MO_16)) {
+ uint32_t data = 0;
+ TCGLabel *over = gen_new_label();
+
+ gen_set_rm_chkfrm(ctx, RISCV_FRM_DYN);
+ tcg_gen_brcondi_tl(TCG_COND_EQ, cpu_vl, 0, over);
+ tcg_gen_brcond_tl(TCG_COND_GEU, cpu_vstart, cpu_vl, over);
+
+ data = FIELD_DP32(data, VDATA, VM, a->vm);
+ data = FIELD_DP32(data, VDATA, LMUL, ctx->lmul);
+ data = FIELD_DP32(data, VDATA, VTA, ctx->vta);
+ data = FIELD_DP32(data, VDATA, VMA, ctx->vma);
+ tcg_gen_gvec_3_ptr(vreg_ofs(ctx, a->rd), vreg_ofs(ctx, 0),
+ vreg_ofs(ctx, a->rs2), cpu_env,
+ ctx->cfg_ptr->vlen / 8,
+ ctx->cfg_ptr->vlen / 8, data,
+ gen_helper_vfwcvtbf16_f_f_v);
+ mark_vs_dirty(ctx);
+ gen_set_label(over);
+ return true;
+ }
+ return false;
+}