target/arm: Implement SVE2 fp multiply-add long
RFC PATCH 1/1 target/arm: Implement SVE2 fp multiply-add long
Implements FMLALB, FMLALT, FMLSLB, FMLSLT
Signed-off-by Stephen Long <[email protected]>
--- target/arm/helper-sve.h | 5 ++++ target/arm/sve.decode | 6 ++++ target/arm/sve_helper.c | 27 ++++++++++++++++++ target/arm/translate-sve.c | 58 ++++++++++++++++++++++++++++++++++++++ 4 files changed, 96 insertions(+)
target/arm/helper-sve.h +5 -0
@@ -2731,3 +2731,8 @@ DEF_HELPER_FLAGS_5(sve2_sqrdcmlah_idx_h, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, i32) DEF_HELPER_FLAGS_5(sve2_sqrdcmlah_idx_s, TCG_CALL_NO_RWG, void, ptr, ptr, ptr, ptr, i32) + +DEF_HELPER_FLAGS_6(sve2_fmlal_zzzw_s, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, ptr, i32) +DEF_HELPER_FLAGS_6(sve2_fmlsl_zzzw_s, TCG_CALL_NO_RWG, + void, ptr, ptr, ptr, ptr, ptr, i32)
target/arm/sve.decode +6 -0
@@ -1568,3 +1568,9 @@ SM4E 01000101 00 10001 1 11100 0 ..... ..... @rdn_rm_e0 # SVE2 crypto constructive binary operations SM4EKEY 01000101 00 1 ..... 11110 0 ..... ..... @rd_rn_rm_e0 RAX1 01000101 00 1 ..... 11110 1 ..... ..... @rd_rn_rm_e0 + +### SVE2 floating-point multiply-add long +FMLALB_zzzw 01100100 .. 1 ..... 10 0 00 0 ..... ..... @rda_rn_rm +FMLALT_zzzw 01100100 .. 1 ..... 10 0 00 1 ..... ..... @rda_rn_rm +FMLSLB_zzzw 01100100 .. 1 ..... 10 1 00 0 ..... ..... @rda_rn_rm +FMLSLT_zzzw 01100100 .. 1 ..... 10 1 00 1 ..... ..... @rda_rn_rm
target/arm/sve_helper.c +27 -0
@@ -7622,3 +7622,30 @@ void HELPER(fmmla_d)(void *vd, void *va, void *vn, void *vm, d[3 ] =float64_add (a[3 ],float64_add (p0, p1, status), status); } } + +void HELPER (sve2_fmlal_zzzw_s )(void *vd,void *vn,void *vm,void *va, +void *status,uint32_t desc) +{ +intptr_t i, opr_sz =simd_oprsz (desc); +intptr_t sel1 =simd_data (desc) *sizeof (float16); +for (i =0 ; i < opr_sz; i +=sizeof (float32)) { +float32 nn = *(float16 *)(vn +H1_2 (i + sel1)); +float32 mm = *(float16 *)(vm +H1_2 (i + sel1)); +float32 aa = *(float32 *)(va +H1_4 (i)); + *(float32 *)(vd +H1_4 (i)) =float32_muladd (nn, mm, aa,0 , status); + } +} + +void HELPER (sve2_fmlsl_zzzw_s )(void *vd,void *vn,void *vm,void *va, +void *status,uint32_t desc) +{ +intptr_t i, opr_sz =simd_oprsz (desc); +intptr_t sel1 =simd_data (desc) *sizeof (float16); +for (i =0 ; i < opr_sz; i +=sizeof (float32)) { +float32 nn = *(float16 *)(vn +H1_2 (i + sel1)); +float32 mm = *(float16 *)(vm +H1_2 (i + sel1)); +float32 aa = *(float32 *)(va +H1_4 (i)); + nn =float32_set_sign (nn,float32_is_neg (nn) ^1 ); + *(float32 *)(vd +H1_4 (i)) =float32_muladd (nn, mm, aa,0 , status); + } +}
target/arm/translate-sve.c +58 -0
61 lines not on this page · raw message
Signature · 1 line
2.17.1
target/arm/sve_helper.c void HELPER(fmmla_d)(void *vd, void *va, void *vn, void *vm, · line 7633
Quoted 3 lines · Stephen Long
target/arm/sve_helper.c +3
-0+for (i =0 ; i < opr_sz; i +=sizeof (float32)) { +float32 nn = *(float16 *)(vn +H1_2 (i + sel1)); +float32 mm = *(float16 *)(vm +H1_2 (i + sel1));
For integers we could rely on C types for the extension. For floating point, you need to call float16_to_float32 here.
r~