-
Notifications
You must be signed in to change notification settings - Fork 14.3k
[AArch64][ARM] Make neon fp16 generic intrinsics always available. #87467
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
By generic intrinsics, this mean things like dup, ext, zip and bsl that can always be executed with integer s16 operations and do not require fullfp16. This makes them always available, and brings them inline with GCC. https://godbolt.org/z/azs8eMv54
You can test this locally with the following command:git-clang-format --diff 29c7d1a60c9d45e82f08cd7487178846ed5f9c6d c7c95cbaa6f57a4d9776d100482526d3245e3da9 -- clang/test/CodeGen/aarch64-v8.2a-neon-intrinsics-generic.c clang/test/CodeGen/arm-v8.2a-neon-intrinsics-generic.c clang/lib/CodeGen/CGBuiltin.cpp clang/test/CodeGen/aarch64-v8.2a-neon-intrinsics.c clang/test/CodeGen/arm-v8.2a-neon-intrinsics.c View the diff from clang-format here.diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index 2537e715b6..b0750893c9 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -7273,149 +7273,397 @@ static const ARMVectorIntrinsicInfo AArch64SISDIntrinsicMap[] = {
// Some intrinsics are equivalent for codegen.
static const std::pair<unsigned, unsigned> NEONEquivalentIntrinsicMap[] = {
- { NEON::BI__builtin_neon_splat_lane_bf16, NEON::BI__builtin_neon_splat_lane_v, },
- { NEON::BI__builtin_neon_splat_laneq_bf16, NEON::BI__builtin_neon_splat_laneq_v, },
- { NEON::BI__builtin_neon_splatq_lane_bf16, NEON::BI__builtin_neon_splatq_lane_v, },
- { NEON::BI__builtin_neon_splatq_laneq_bf16, NEON::BI__builtin_neon_splatq_laneq_v, },
- { NEON::BI__builtin_neon_vabd_f16, NEON::BI__builtin_neon_vabd_v, },
- { NEON::BI__builtin_neon_vabdq_f16, NEON::BI__builtin_neon_vabdq_v, },
- { NEON::BI__builtin_neon_vabs_f16, NEON::BI__builtin_neon_vabs_v, },
- { NEON::BI__builtin_neon_vabsq_f16, NEON::BI__builtin_neon_vabsq_v, },
- { NEON::BI__builtin_neon_vcage_f16, NEON::BI__builtin_neon_vcage_v, },
- { NEON::BI__builtin_neon_vcageq_f16, NEON::BI__builtin_neon_vcageq_v, },
- { NEON::BI__builtin_neon_vcagt_f16, NEON::BI__builtin_neon_vcagt_v, },
- { NEON::BI__builtin_neon_vcagtq_f16, NEON::BI__builtin_neon_vcagtq_v, },
- { NEON::BI__builtin_neon_vcale_f16, NEON::BI__builtin_neon_vcale_v, },
- { NEON::BI__builtin_neon_vcaleq_f16, NEON::BI__builtin_neon_vcaleq_v, },
- { NEON::BI__builtin_neon_vcalt_f16, NEON::BI__builtin_neon_vcalt_v, },
- { NEON::BI__builtin_neon_vcaltq_f16, NEON::BI__builtin_neon_vcaltq_v, },
- { NEON::BI__builtin_neon_vceqz_f16, NEON::BI__builtin_neon_vceqz_v, },
- { NEON::BI__builtin_neon_vceqzq_f16, NEON::BI__builtin_neon_vceqzq_v, },
- { NEON::BI__builtin_neon_vcgez_f16, NEON::BI__builtin_neon_vcgez_v, },
- { NEON::BI__builtin_neon_vcgezq_f16, NEON::BI__builtin_neon_vcgezq_v, },
- { NEON::BI__builtin_neon_vcgtz_f16, NEON::BI__builtin_neon_vcgtz_v, },
- { NEON::BI__builtin_neon_vcgtzq_f16, NEON::BI__builtin_neon_vcgtzq_v, },
- { NEON::BI__builtin_neon_vclez_f16, NEON::BI__builtin_neon_vclez_v, },
- { NEON::BI__builtin_neon_vclezq_f16, NEON::BI__builtin_neon_vclezq_v, },
- { NEON::BI__builtin_neon_vcltz_f16, NEON::BI__builtin_neon_vcltz_v, },
- { NEON::BI__builtin_neon_vcltzq_f16, NEON::BI__builtin_neon_vcltzq_v, },
- { NEON::BI__builtin_neon_vfma_f16, NEON::BI__builtin_neon_vfma_v, },
- { NEON::BI__builtin_neon_vfma_lane_f16, NEON::BI__builtin_neon_vfma_lane_v, },
- { NEON::BI__builtin_neon_vfma_laneq_f16, NEON::BI__builtin_neon_vfma_laneq_v, },
- { NEON::BI__builtin_neon_vfmaq_f16, NEON::BI__builtin_neon_vfmaq_v, },
- { NEON::BI__builtin_neon_vfmaq_lane_f16, NEON::BI__builtin_neon_vfmaq_lane_v, },
- { NEON::BI__builtin_neon_vfmaq_laneq_f16, NEON::BI__builtin_neon_vfmaq_laneq_v, },
- { NEON::BI__builtin_neon_vld1_bf16_x2, NEON::BI__builtin_neon_vld1_x2_v },
- { NEON::BI__builtin_neon_vld1_bf16_x3, NEON::BI__builtin_neon_vld1_x3_v },
- { NEON::BI__builtin_neon_vld1_bf16_x4, NEON::BI__builtin_neon_vld1_x4_v },
- { NEON::BI__builtin_neon_vld1_bf16, NEON::BI__builtin_neon_vld1_v },
- { NEON::BI__builtin_neon_vld1_dup_bf16, NEON::BI__builtin_neon_vld1_dup_v },
- { NEON::BI__builtin_neon_vld1_lane_bf16, NEON::BI__builtin_neon_vld1_lane_v },
- { NEON::BI__builtin_neon_vld1q_bf16_x2, NEON::BI__builtin_neon_vld1q_x2_v },
- { NEON::BI__builtin_neon_vld1q_bf16_x3, NEON::BI__builtin_neon_vld1q_x3_v },
- { NEON::BI__builtin_neon_vld1q_bf16_x4, NEON::BI__builtin_neon_vld1q_x4_v },
- { NEON::BI__builtin_neon_vld1q_bf16, NEON::BI__builtin_neon_vld1q_v },
- { NEON::BI__builtin_neon_vld1q_dup_bf16, NEON::BI__builtin_neon_vld1q_dup_v },
- { NEON::BI__builtin_neon_vld1q_lane_bf16, NEON::BI__builtin_neon_vld1q_lane_v },
- { NEON::BI__builtin_neon_vld2_bf16, NEON::BI__builtin_neon_vld2_v },
- { NEON::BI__builtin_neon_vld2_dup_bf16, NEON::BI__builtin_neon_vld2_dup_v },
- { NEON::BI__builtin_neon_vld2_lane_bf16, NEON::BI__builtin_neon_vld2_lane_v },
- { NEON::BI__builtin_neon_vld2q_bf16, NEON::BI__builtin_neon_vld2q_v },
- { NEON::BI__builtin_neon_vld2q_dup_bf16, NEON::BI__builtin_neon_vld2q_dup_v },
- { NEON::BI__builtin_neon_vld2q_lane_bf16, NEON::BI__builtin_neon_vld2q_lane_v },
- { NEON::BI__builtin_neon_vld3_bf16, NEON::BI__builtin_neon_vld3_v },
- { NEON::BI__builtin_neon_vld3_dup_bf16, NEON::BI__builtin_neon_vld3_dup_v },
- { NEON::BI__builtin_neon_vld3_lane_bf16, NEON::BI__builtin_neon_vld3_lane_v },
- { NEON::BI__builtin_neon_vld3q_bf16, NEON::BI__builtin_neon_vld3q_v },
- { NEON::BI__builtin_neon_vld3q_dup_bf16, NEON::BI__builtin_neon_vld3q_dup_v },
- { NEON::BI__builtin_neon_vld3q_lane_bf16, NEON::BI__builtin_neon_vld3q_lane_v },
- { NEON::BI__builtin_neon_vld4_bf16, NEON::BI__builtin_neon_vld4_v },
- { NEON::BI__builtin_neon_vld4_dup_bf16, NEON::BI__builtin_neon_vld4_dup_v },
- { NEON::BI__builtin_neon_vld4_lane_bf16, NEON::BI__builtin_neon_vld4_lane_v },
- { NEON::BI__builtin_neon_vld4q_bf16, NEON::BI__builtin_neon_vld4q_v },
- { NEON::BI__builtin_neon_vld4q_dup_bf16, NEON::BI__builtin_neon_vld4q_dup_v },
- { NEON::BI__builtin_neon_vld4q_lane_bf16, NEON::BI__builtin_neon_vld4q_lane_v },
- { NEON::BI__builtin_neon_vmax_f16, NEON::BI__builtin_neon_vmax_v, },
- { NEON::BI__builtin_neon_vmaxnm_f16, NEON::BI__builtin_neon_vmaxnm_v, },
- { NEON::BI__builtin_neon_vmaxnmq_f16, NEON::BI__builtin_neon_vmaxnmq_v, },
- { NEON::BI__builtin_neon_vmaxq_f16, NEON::BI__builtin_neon_vmaxq_v, },
- { NEON::BI__builtin_neon_vmin_f16, NEON::BI__builtin_neon_vmin_v, },
- { NEON::BI__builtin_neon_vminnm_f16, NEON::BI__builtin_neon_vminnm_v, },
- { NEON::BI__builtin_neon_vminnmq_f16, NEON::BI__builtin_neon_vminnmq_v, },
- { NEON::BI__builtin_neon_vminq_f16, NEON::BI__builtin_neon_vminq_v, },
- { NEON::BI__builtin_neon_vmulx_f16, NEON::BI__builtin_neon_vmulx_v, },
- { NEON::BI__builtin_neon_vmulxq_f16, NEON::BI__builtin_neon_vmulxq_v, },
- { NEON::BI__builtin_neon_vpadd_f16, NEON::BI__builtin_neon_vpadd_v, },
- { NEON::BI__builtin_neon_vpaddq_f16, NEON::BI__builtin_neon_vpaddq_v, },
- { NEON::BI__builtin_neon_vpmax_f16, NEON::BI__builtin_neon_vpmax_v, },
- { NEON::BI__builtin_neon_vpmaxnm_f16, NEON::BI__builtin_neon_vpmaxnm_v, },
- { NEON::BI__builtin_neon_vpmaxnmq_f16, NEON::BI__builtin_neon_vpmaxnmq_v, },
- { NEON::BI__builtin_neon_vpmaxq_f16, NEON::BI__builtin_neon_vpmaxq_v, },
- { NEON::BI__builtin_neon_vpmin_f16, NEON::BI__builtin_neon_vpmin_v, },
- { NEON::BI__builtin_neon_vpminnm_f16, NEON::BI__builtin_neon_vpminnm_v, },
- { NEON::BI__builtin_neon_vpminnmq_f16, NEON::BI__builtin_neon_vpminnmq_v, },
- { NEON::BI__builtin_neon_vpminq_f16, NEON::BI__builtin_neon_vpminq_v, },
- { NEON::BI__builtin_neon_vrecpe_f16, NEON::BI__builtin_neon_vrecpe_v, },
- { NEON::BI__builtin_neon_vrecpeq_f16, NEON::BI__builtin_neon_vrecpeq_v, },
- { NEON::BI__builtin_neon_vrecps_f16, NEON::BI__builtin_neon_vrecps_v, },
- { NEON::BI__builtin_neon_vrecpsq_f16, NEON::BI__builtin_neon_vrecpsq_v, },
- { NEON::BI__builtin_neon_vrnd_f16, NEON::BI__builtin_neon_vrnd_v, },
- { NEON::BI__builtin_neon_vrnda_f16, NEON::BI__builtin_neon_vrnda_v, },
- { NEON::BI__builtin_neon_vrndaq_f16, NEON::BI__builtin_neon_vrndaq_v, },
- { NEON::BI__builtin_neon_vrndi_f16, NEON::BI__builtin_neon_vrndi_v, },
- { NEON::BI__builtin_neon_vrndiq_f16, NEON::BI__builtin_neon_vrndiq_v, },
- { NEON::BI__builtin_neon_vrndm_f16, NEON::BI__builtin_neon_vrndm_v, },
- { NEON::BI__builtin_neon_vrndmq_f16, NEON::BI__builtin_neon_vrndmq_v, },
- { NEON::BI__builtin_neon_vrndn_f16, NEON::BI__builtin_neon_vrndn_v, },
- { NEON::BI__builtin_neon_vrndnq_f16, NEON::BI__builtin_neon_vrndnq_v, },
- { NEON::BI__builtin_neon_vrndp_f16, NEON::BI__builtin_neon_vrndp_v, },
- { NEON::BI__builtin_neon_vrndpq_f16, NEON::BI__builtin_neon_vrndpq_v, },
- { NEON::BI__builtin_neon_vrndq_f16, NEON::BI__builtin_neon_vrndq_v, },
- { NEON::BI__builtin_neon_vrndx_f16, NEON::BI__builtin_neon_vrndx_v, },
- { NEON::BI__builtin_neon_vrndxq_f16, NEON::BI__builtin_neon_vrndxq_v, },
- { NEON::BI__builtin_neon_vrsqrte_f16, NEON::BI__builtin_neon_vrsqrte_v, },
- { NEON::BI__builtin_neon_vrsqrteq_f16, NEON::BI__builtin_neon_vrsqrteq_v, },
- { NEON::BI__builtin_neon_vrsqrts_f16, NEON::BI__builtin_neon_vrsqrts_v, },
- { NEON::BI__builtin_neon_vrsqrtsq_f16, NEON::BI__builtin_neon_vrsqrtsq_v, },
- { NEON::BI__builtin_neon_vsqrt_f16, NEON::BI__builtin_neon_vsqrt_v, },
- { NEON::BI__builtin_neon_vsqrtq_f16, NEON::BI__builtin_neon_vsqrtq_v, },
- { NEON::BI__builtin_neon_vst1_bf16_x2, NEON::BI__builtin_neon_vst1_x2_v },
- { NEON::BI__builtin_neon_vst1_bf16_x3, NEON::BI__builtin_neon_vst1_x3_v },
- { NEON::BI__builtin_neon_vst1_bf16_x4, NEON::BI__builtin_neon_vst1_x4_v },
- { NEON::BI__builtin_neon_vst1_bf16, NEON::BI__builtin_neon_vst1_v },
- { NEON::BI__builtin_neon_vst1_lane_bf16, NEON::BI__builtin_neon_vst1_lane_v },
- { NEON::BI__builtin_neon_vst1q_bf16_x2, NEON::BI__builtin_neon_vst1q_x2_v },
- { NEON::BI__builtin_neon_vst1q_bf16_x3, NEON::BI__builtin_neon_vst1q_x3_v },
- { NEON::BI__builtin_neon_vst1q_bf16_x4, NEON::BI__builtin_neon_vst1q_x4_v },
- { NEON::BI__builtin_neon_vst1q_bf16, NEON::BI__builtin_neon_vst1q_v },
- { NEON::BI__builtin_neon_vst1q_lane_bf16, NEON::BI__builtin_neon_vst1q_lane_v },
- { NEON::BI__builtin_neon_vst2_bf16, NEON::BI__builtin_neon_vst2_v },
- { NEON::BI__builtin_neon_vst2_lane_bf16, NEON::BI__builtin_neon_vst2_lane_v },
- { NEON::BI__builtin_neon_vst2q_bf16, NEON::BI__builtin_neon_vst2q_v },
- { NEON::BI__builtin_neon_vst2q_lane_bf16, NEON::BI__builtin_neon_vst2q_lane_v },
- { NEON::BI__builtin_neon_vst3_bf16, NEON::BI__builtin_neon_vst3_v },
- { NEON::BI__builtin_neon_vst3_lane_bf16, NEON::BI__builtin_neon_vst3_lane_v },
- { NEON::BI__builtin_neon_vst3q_bf16, NEON::BI__builtin_neon_vst3q_v },
- { NEON::BI__builtin_neon_vst3q_lane_bf16, NEON::BI__builtin_neon_vst3q_lane_v },
- { NEON::BI__builtin_neon_vst4_bf16, NEON::BI__builtin_neon_vst4_v },
- { NEON::BI__builtin_neon_vst4_lane_bf16, NEON::BI__builtin_neon_vst4_lane_v },
- { NEON::BI__builtin_neon_vst4q_bf16, NEON::BI__builtin_neon_vst4q_v },
- { NEON::BI__builtin_neon_vst4q_lane_bf16, NEON::BI__builtin_neon_vst4q_lane_v },
- // The mangling rules cause us to have one ID for each type for vldap1(q)_lane
- // and vstl1(q)_lane, but codegen is equivalent for all of them. Choose an
- // arbitrary one to be handled as tha canonical variation.
- { NEON::BI__builtin_neon_vldap1_lane_u64, NEON::BI__builtin_neon_vldap1_lane_s64 },
- { NEON::BI__builtin_neon_vldap1_lane_f64, NEON::BI__builtin_neon_vldap1_lane_s64 },
- { NEON::BI__builtin_neon_vldap1_lane_p64, NEON::BI__builtin_neon_vldap1_lane_s64 },
- { NEON::BI__builtin_neon_vldap1q_lane_u64, NEON::BI__builtin_neon_vldap1q_lane_s64 },
- { NEON::BI__builtin_neon_vldap1q_lane_f64, NEON::BI__builtin_neon_vldap1q_lane_s64 },
- { NEON::BI__builtin_neon_vldap1q_lane_p64, NEON::BI__builtin_neon_vldap1q_lane_s64 },
- { NEON::BI__builtin_neon_vstl1_lane_u64, NEON::BI__builtin_neon_vstl1_lane_s64 },
- { NEON::BI__builtin_neon_vstl1_lane_f64, NEON::BI__builtin_neon_vstl1_lane_s64 },
- { NEON::BI__builtin_neon_vstl1_lane_p64, NEON::BI__builtin_neon_vstl1_lane_s64 },
- { NEON::BI__builtin_neon_vstl1q_lane_u64, NEON::BI__builtin_neon_vstl1q_lane_s64 },
- { NEON::BI__builtin_neon_vstl1q_lane_f64, NEON::BI__builtin_neon_vstl1q_lane_s64 },
- { NEON::BI__builtin_neon_vstl1q_lane_p64, NEON::BI__builtin_neon_vstl1q_lane_s64 },
+ {
+ NEON::BI__builtin_neon_splat_lane_bf16,
+ NEON::BI__builtin_neon_splat_lane_v,
+ },
+ {
+ NEON::BI__builtin_neon_splat_laneq_bf16,
+ NEON::BI__builtin_neon_splat_laneq_v,
+ },
+ {
+ NEON::BI__builtin_neon_splatq_lane_bf16,
+ NEON::BI__builtin_neon_splatq_lane_v,
+ },
+ {
+ NEON::BI__builtin_neon_splatq_laneq_bf16,
+ NEON::BI__builtin_neon_splatq_laneq_v,
+ },
+ {
+ NEON::BI__builtin_neon_vabd_f16,
+ NEON::BI__builtin_neon_vabd_v,
+ },
+ {
+ NEON::BI__builtin_neon_vabdq_f16,
+ NEON::BI__builtin_neon_vabdq_v,
+ },
+ {
+ NEON::BI__builtin_neon_vabs_f16,
+ NEON::BI__builtin_neon_vabs_v,
+ },
+ {
+ NEON::BI__builtin_neon_vabsq_f16,
+ NEON::BI__builtin_neon_vabsq_v,
+ },
+ {
+ NEON::BI__builtin_neon_vcage_f16,
+ NEON::BI__builtin_neon_vcage_v,
+ },
+ {
+ NEON::BI__builtin_neon_vcageq_f16,
+ NEON::BI__builtin_neon_vcageq_v,
+ },
+ {
+ NEON::BI__builtin_neon_vcagt_f16,
+ NEON::BI__builtin_neon_vcagt_v,
+ },
+ {
+ NEON::BI__builtin_neon_vcagtq_f16,
+ NEON::BI__builtin_neon_vcagtq_v,
+ },
+ {
+ NEON::BI__builtin_neon_vcale_f16,
+ NEON::BI__builtin_neon_vcale_v,
+ },
+ {
+ NEON::BI__builtin_neon_vcaleq_f16,
+ NEON::BI__builtin_neon_vcaleq_v,
+ },
+ {
+ NEON::BI__builtin_neon_vcalt_f16,
+ NEON::BI__builtin_neon_vcalt_v,
+ },
+ {
+ NEON::BI__builtin_neon_vcaltq_f16,
+ NEON::BI__builtin_neon_vcaltq_v,
+ },
+ {
+ NEON::BI__builtin_neon_vceqz_f16,
+ NEON::BI__builtin_neon_vceqz_v,
+ },
+ {
+ NEON::BI__builtin_neon_vceqzq_f16,
+ NEON::BI__builtin_neon_vceqzq_v,
+ },
+ {
+ NEON::BI__builtin_neon_vcgez_f16,
+ NEON::BI__builtin_neon_vcgez_v,
+ },
+ {
+ NEON::BI__builtin_neon_vcgezq_f16,
+ NEON::BI__builtin_neon_vcgezq_v,
+ },
+ {
+ NEON::BI__builtin_neon_vcgtz_f16,
+ NEON::BI__builtin_neon_vcgtz_v,
+ },
+ {
+ NEON::BI__builtin_neon_vcgtzq_f16,
+ NEON::BI__builtin_neon_vcgtzq_v,
+ },
+ {
+ NEON::BI__builtin_neon_vclez_f16,
+ NEON::BI__builtin_neon_vclez_v,
+ },
+ {
+ NEON::BI__builtin_neon_vclezq_f16,
+ NEON::BI__builtin_neon_vclezq_v,
+ },
+ {
+ NEON::BI__builtin_neon_vcltz_f16,
+ NEON::BI__builtin_neon_vcltz_v,
+ },
+ {
+ NEON::BI__builtin_neon_vcltzq_f16,
+ NEON::BI__builtin_neon_vcltzq_v,
+ },
+ {
+ NEON::BI__builtin_neon_vfma_f16,
+ NEON::BI__builtin_neon_vfma_v,
+ },
+ {
+ NEON::BI__builtin_neon_vfma_lane_f16,
+ NEON::BI__builtin_neon_vfma_lane_v,
+ },
+ {
+ NEON::BI__builtin_neon_vfma_laneq_f16,
+ NEON::BI__builtin_neon_vfma_laneq_v,
+ },
+ {
+ NEON::BI__builtin_neon_vfmaq_f16,
+ NEON::BI__builtin_neon_vfmaq_v,
+ },
+ {
+ NEON::BI__builtin_neon_vfmaq_lane_f16,
+ NEON::BI__builtin_neon_vfmaq_lane_v,
+ },
+ {
+ NEON::BI__builtin_neon_vfmaq_laneq_f16,
+ NEON::BI__builtin_neon_vfmaq_laneq_v,
+ },
+ {NEON::BI__builtin_neon_vld1_bf16_x2, NEON::BI__builtin_neon_vld1_x2_v},
+ {NEON::BI__builtin_neon_vld1_bf16_x3, NEON::BI__builtin_neon_vld1_x3_v},
+ {NEON::BI__builtin_neon_vld1_bf16_x4, NEON::BI__builtin_neon_vld1_x4_v},
+ {NEON::BI__builtin_neon_vld1_bf16, NEON::BI__builtin_neon_vld1_v},
+ {NEON::BI__builtin_neon_vld1_dup_bf16, NEON::BI__builtin_neon_vld1_dup_v},
+ {NEON::BI__builtin_neon_vld1_lane_bf16, NEON::BI__builtin_neon_vld1_lane_v},
+ {NEON::BI__builtin_neon_vld1q_bf16_x2, NEON::BI__builtin_neon_vld1q_x2_v},
+ {NEON::BI__builtin_neon_vld1q_bf16_x3, NEON::BI__builtin_neon_vld1q_x3_v},
+ {NEON::BI__builtin_neon_vld1q_bf16_x4, NEON::BI__builtin_neon_vld1q_x4_v},
+ {NEON::BI__builtin_neon_vld1q_bf16, NEON::BI__builtin_neon_vld1q_v},
+ {NEON::BI__builtin_neon_vld1q_dup_bf16, NEON::BI__builtin_neon_vld1q_dup_v},
+ {NEON::BI__builtin_neon_vld1q_lane_bf16,
+ NEON::BI__builtin_neon_vld1q_lane_v},
+ {NEON::BI__builtin_neon_vld2_bf16, NEON::BI__builtin_neon_vld2_v},
+ {NEON::BI__builtin_neon_vld2_dup_bf16, NEON::BI__builtin_neon_vld2_dup_v},
+ {NEON::BI__builtin_neon_vld2_lane_bf16, NEON::BI__builtin_neon_vld2_lane_v},
+ {NEON::BI__builtin_neon_vld2q_bf16, NEON::BI__builtin_neon_vld2q_v},
+ {NEON::BI__builtin_neon_vld2q_dup_bf16, NEON::BI__builtin_neon_vld2q_dup_v},
+ {NEON::BI__builtin_neon_vld2q_lane_bf16,
+ NEON::BI__builtin_neon_vld2q_lane_v},
+ {NEON::BI__builtin_neon_vld3_bf16, NEON::BI__builtin_neon_vld3_v},
+ {NEON::BI__builtin_neon_vld3_dup_bf16, NEON::BI__builtin_neon_vld3_dup_v},
+ {NEON::BI__builtin_neon_vld3_lane_bf16, NEON::BI__builtin_neon_vld3_lane_v},
+ {NEON::BI__builtin_neon_vld3q_bf16, NEON::BI__builtin_neon_vld3q_v},
+ {NEON::BI__builtin_neon_vld3q_dup_bf16, NEON::BI__builtin_neon_vld3q_dup_v},
+ {NEON::BI__builtin_neon_vld3q_lane_bf16,
+ NEON::BI__builtin_neon_vld3q_lane_v},
+ {NEON::BI__builtin_neon_vld4_bf16, NEON::BI__builtin_neon_vld4_v},
+ {NEON::BI__builtin_neon_vld4_dup_bf16, NEON::BI__builtin_neon_vld4_dup_v},
+ {NEON::BI__builtin_neon_vld4_lane_bf16, NEON::BI__builtin_neon_vld4_lane_v},
+ {NEON::BI__builtin_neon_vld4q_bf16, NEON::BI__builtin_neon_vld4q_v},
+ {NEON::BI__builtin_neon_vld4q_dup_bf16, NEON::BI__builtin_neon_vld4q_dup_v},
+ {NEON::BI__builtin_neon_vld4q_lane_bf16,
+ NEON::BI__builtin_neon_vld4q_lane_v},
+ {
+ NEON::BI__builtin_neon_vmax_f16,
+ NEON::BI__builtin_neon_vmax_v,
+ },
+ {
+ NEON::BI__builtin_neon_vmaxnm_f16,
+ NEON::BI__builtin_neon_vmaxnm_v,
+ },
+ {
+ NEON::BI__builtin_neon_vmaxnmq_f16,
+ NEON::BI__builtin_neon_vmaxnmq_v,
+ },
+ {
+ NEON::BI__builtin_neon_vmaxq_f16,
+ NEON::BI__builtin_neon_vmaxq_v,
+ },
+ {
+ NEON::BI__builtin_neon_vmin_f16,
+ NEON::BI__builtin_neon_vmin_v,
+ },
+ {
+ NEON::BI__builtin_neon_vminnm_f16,
+ NEON::BI__builtin_neon_vminnm_v,
+ },
+ {
+ NEON::BI__builtin_neon_vminnmq_f16,
+ NEON::BI__builtin_neon_vminnmq_v,
+ },
+ {
+ NEON::BI__builtin_neon_vminq_f16,
+ NEON::BI__builtin_neon_vminq_v,
+ },
+ {
+ NEON::BI__builtin_neon_vmulx_f16,
+ NEON::BI__builtin_neon_vmulx_v,
+ },
+ {
+ NEON::BI__builtin_neon_vmulxq_f16,
+ NEON::BI__builtin_neon_vmulxq_v,
+ },
+ {
+ NEON::BI__builtin_neon_vpadd_f16,
+ NEON::BI__builtin_neon_vpadd_v,
+ },
+ {
+ NEON::BI__builtin_neon_vpaddq_f16,
+ NEON::BI__builtin_neon_vpaddq_v,
+ },
+ {
+ NEON::BI__builtin_neon_vpmax_f16,
+ NEON::BI__builtin_neon_vpmax_v,
+ },
+ {
+ NEON::BI__builtin_neon_vpmaxnm_f16,
+ NEON::BI__builtin_neon_vpmaxnm_v,
+ },
+ {
+ NEON::BI__builtin_neon_vpmaxnmq_f16,
+ NEON::BI__builtin_neon_vpmaxnmq_v,
+ },
+ {
+ NEON::BI__builtin_neon_vpmaxq_f16,
+ NEON::BI__builtin_neon_vpmaxq_v,
+ },
+ {
+ NEON::BI__builtin_neon_vpmin_f16,
+ NEON::BI__builtin_neon_vpmin_v,
+ },
+ {
+ NEON::BI__builtin_neon_vpminnm_f16,
+ NEON::BI__builtin_neon_vpminnm_v,
+ },
+ {
+ NEON::BI__builtin_neon_vpminnmq_f16,
+ NEON::BI__builtin_neon_vpminnmq_v,
+ },
+ {
+ NEON::BI__builtin_neon_vpminq_f16,
+ NEON::BI__builtin_neon_vpminq_v,
+ },
+ {
+ NEON::BI__builtin_neon_vrecpe_f16,
+ NEON::BI__builtin_neon_vrecpe_v,
+ },
+ {
+ NEON::BI__builtin_neon_vrecpeq_f16,
+ NEON::BI__builtin_neon_vrecpeq_v,
+ },
+ {
+ NEON::BI__builtin_neon_vrecps_f16,
+ NEON::BI__builtin_neon_vrecps_v,
+ },
+ {
+ NEON::BI__builtin_neon_vrecpsq_f16,
+ NEON::BI__builtin_neon_vrecpsq_v,
+ },
+ {
+ NEON::BI__builtin_neon_vrnd_f16,
+ NEON::BI__builtin_neon_vrnd_v,
+ },
+ {
+ NEON::BI__builtin_neon_vrnda_f16,
+ NEON::BI__builtin_neon_vrnda_v,
+ },
+ {
+ NEON::BI__builtin_neon_vrndaq_f16,
+ NEON::BI__builtin_neon_vrndaq_v,
+ },
+ {
+ NEON::BI__builtin_neon_vrndi_f16,
+ NEON::BI__builtin_neon_vrndi_v,
+ },
+ {
+ NEON::BI__builtin_neon_vrndiq_f16,
+ NEON::BI__builtin_neon_vrndiq_v,
+ },
+ {
+ NEON::BI__builtin_neon_vrndm_f16,
+ NEON::BI__builtin_neon_vrndm_v,
+ },
+ {
+ NEON::BI__builtin_neon_vrndmq_f16,
+ NEON::BI__builtin_neon_vrndmq_v,
+ },
+ {
+ NEON::BI__builtin_neon_vrndn_f16,
+ NEON::BI__builtin_neon_vrndn_v,
+ },
+ {
+ NEON::BI__builtin_neon_vrndnq_f16,
+ NEON::BI__builtin_neon_vrndnq_v,
+ },
+ {
+ NEON::BI__builtin_neon_vrndp_f16,
+ NEON::BI__builtin_neon_vrndp_v,
+ },
+ {
+ NEON::BI__builtin_neon_vrndpq_f16,
+ NEON::BI__builtin_neon_vrndpq_v,
+ },
+ {
+ NEON::BI__builtin_neon_vrndq_f16,
+ NEON::BI__builtin_neon_vrndq_v,
+ },
+ {
+ NEON::BI__builtin_neon_vrndx_f16,
+ NEON::BI__builtin_neon_vrndx_v,
+ },
+ {
+ NEON::BI__builtin_neon_vrndxq_f16,
+ NEON::BI__builtin_neon_vrndxq_v,
+ },
+ {
+ NEON::BI__builtin_neon_vrsqrte_f16,
+ NEON::BI__builtin_neon_vrsqrte_v,
+ },
+ {
+ NEON::BI__builtin_neon_vrsqrteq_f16,
+ NEON::BI__builtin_neon_vrsqrteq_v,
+ },
+ {
+ NEON::BI__builtin_neon_vrsqrts_f16,
+ NEON::BI__builtin_neon_vrsqrts_v,
+ },
+ {
+ NEON::BI__builtin_neon_vrsqrtsq_f16,
+ NEON::BI__builtin_neon_vrsqrtsq_v,
+ },
+ {
+ NEON::BI__builtin_neon_vsqrt_f16,
+ NEON::BI__builtin_neon_vsqrt_v,
+ },
+ {
+ NEON::BI__builtin_neon_vsqrtq_f16,
+ NEON::BI__builtin_neon_vsqrtq_v,
+ },
+ {NEON::BI__builtin_neon_vst1_bf16_x2, NEON::BI__builtin_neon_vst1_x2_v},
+ {NEON::BI__builtin_neon_vst1_bf16_x3, NEON::BI__builtin_neon_vst1_x3_v},
+ {NEON::BI__builtin_neon_vst1_bf16_x4, NEON::BI__builtin_neon_vst1_x4_v},
+ {NEON::BI__builtin_neon_vst1_bf16, NEON::BI__builtin_neon_vst1_v},
+ {NEON::BI__builtin_neon_vst1_lane_bf16, NEON::BI__builtin_neon_vst1_lane_v},
+ {NEON::BI__builtin_neon_vst1q_bf16_x2, NEON::BI__builtin_neon_vst1q_x2_v},
+ {NEON::BI__builtin_neon_vst1q_bf16_x3, NEON::BI__builtin_neon_vst1q_x3_v},
+ {NEON::BI__builtin_neon_vst1q_bf16_x4, NEON::BI__builtin_neon_vst1q_x4_v},
+ {NEON::BI__builtin_neon_vst1q_bf16, NEON::BI__builtin_neon_vst1q_v},
+ {NEON::BI__builtin_neon_vst1q_lane_bf16,
+ NEON::BI__builtin_neon_vst1q_lane_v},
+ {NEON::BI__builtin_neon_vst2_bf16, NEON::BI__builtin_neon_vst2_v},
+ {NEON::BI__builtin_neon_vst2_lane_bf16, NEON::BI__builtin_neon_vst2_lane_v},
+ {NEON::BI__builtin_neon_vst2q_bf16, NEON::BI__builtin_neon_vst2q_v},
+ {NEON::BI__builtin_neon_vst2q_lane_bf16,
+ NEON::BI__builtin_neon_vst2q_lane_v},
+ {NEON::BI__builtin_neon_vst3_bf16, NEON::BI__builtin_neon_vst3_v},
+ {NEON::BI__builtin_neon_vst3_lane_bf16, NEON::BI__builtin_neon_vst3_lane_v},
+ {NEON::BI__builtin_neon_vst3q_bf16, NEON::BI__builtin_neon_vst3q_v},
+ {NEON::BI__builtin_neon_vst3q_lane_bf16,
+ NEON::BI__builtin_neon_vst3q_lane_v},
+ {NEON::BI__builtin_neon_vst4_bf16, NEON::BI__builtin_neon_vst4_v},
+ {NEON::BI__builtin_neon_vst4_lane_bf16, NEON::BI__builtin_neon_vst4_lane_v},
+ {NEON::BI__builtin_neon_vst4q_bf16, NEON::BI__builtin_neon_vst4q_v},
+ {NEON::BI__builtin_neon_vst4q_lane_bf16,
+ NEON::BI__builtin_neon_vst4q_lane_v},
+ // The mangling rules cause us to have one ID for each type for
+ // vldap1(q)_lane and vstl1(q)_lane, but codegen is equivalent for all of
+ // them. Choose an arbitrary one to be handled as tha canonical variation.
+ {NEON::BI__builtin_neon_vldap1_lane_u64,
+ NEON::BI__builtin_neon_vldap1_lane_s64},
+ {NEON::BI__builtin_neon_vldap1_lane_f64,
+ NEON::BI__builtin_neon_vldap1_lane_s64},
+ {NEON::BI__builtin_neon_vldap1_lane_p64,
+ NEON::BI__builtin_neon_vldap1_lane_s64},
+ {NEON::BI__builtin_neon_vldap1q_lane_u64,
+ NEON::BI__builtin_neon_vldap1q_lane_s64},
+ {NEON::BI__builtin_neon_vldap1q_lane_f64,
+ NEON::BI__builtin_neon_vldap1q_lane_s64},
+ {NEON::BI__builtin_neon_vldap1q_lane_p64,
+ NEON::BI__builtin_neon_vldap1q_lane_s64},
+ {NEON::BI__builtin_neon_vstl1_lane_u64,
+ NEON::BI__builtin_neon_vstl1_lane_s64},
+ {NEON::BI__builtin_neon_vstl1_lane_f64,
+ NEON::BI__builtin_neon_vstl1_lane_s64},
+ {NEON::BI__builtin_neon_vstl1_lane_p64,
+ NEON::BI__builtin_neon_vstl1_lane_s64},
+ {NEON::BI__builtin_neon_vstl1q_lane_u64,
+ NEON::BI__builtin_neon_vstl1q_lane_s64},
+ {NEON::BI__builtin_neon_vstl1q_lane_f64,
+ NEON::BI__builtin_neon_vstl1q_lane_s64},
+ {NEON::BI__builtin_neon_vstl1q_lane_p64,
+ NEON::BI__builtin_neon_vstl1q_lane_s64},
};
#undef NEONMAP0
|
ostannard
approved these changes
Apr 3, 2024
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
By generic intrinsics this mean things like dup, ext, zip and bsl that can always be executed with integer s16 operations and do not require fullfp16. This makes them always available, and brings them inline with GCC.
https://godbolt.org/z/azs8eMv54
The relevant test cases have been moved into their own files, to allow them to be tested with armv8-a and armv8.2-a+fp16.