diff options
| author | Simon Pilgrim <llvm-dev@redking.me.uk> | 2016-06-13 09:57:52 +0000 |
|---|---|---|
| committer | Simon Pilgrim <llvm-dev@redking.me.uk> | 2016-06-13 09:57:52 +0000 |
| commit | beca5f295cf6720add28ae40aa87378a86bf52a1 (patch) | |
| tree | 5aae895ce9ee76a41683d444558fa2a67f58c496 /clang/test/CodeGen/builtins-x86.c | |
| parent | 7900ca1e7e80d85fc22fefdf8d723df16570017e (diff) | |
| download | bcm5719-llvm-beca5f295cf6720add28ae40aa87378a86bf52a1.tar.gz bcm5719-llvm-beca5f295cf6720add28ae40aa87378a86bf52a1.zip | |
[Clang][X86] Convert non-temporal store builtins to generic __builtin_nontemporal_store in headers
We can now use __builtin_nontemporal_store instead of target specific builtins for naturally aligned nontemporal stores which avoids the need for handling in CGBuiltin.cpp
The scalar integer nontemporal (unaligned) store builtins will have to wait as __builtin_nontemporal_store currently assumes natural alignment and doesn't accept the 'packed struct' trick that we use for normal unaligned load/stores.
The nontemporal loads require further backend support before we can safely convert them to __builtin_nontemporal_load
Differential Revision: http://reviews.llvm.org/D21272
llvm-svn: 272540
Diffstat (limited to 'clang/test/CodeGen/builtins-x86.c')
| -rw-r--r-- | clang/test/CodeGen/builtins-x86.c | 6 |
1 files changed, 0 insertions, 6 deletions
diff --git a/clang/test/CodeGen/builtins-x86.c b/clang/test/CodeGen/builtins-x86.c index 24e491c0262..2dedd7913a7 100644 --- a/clang/test/CodeGen/builtins-x86.c +++ b/clang/test/CodeGen/builtins-x86.c @@ -300,7 +300,6 @@ void f0() { (void) __builtin_ia32_storelps(tmp_V2ip, tmp_V4f); tmp_i = __builtin_ia32_movmskps(tmp_V4f); tmp_i = __builtin_ia32_pmovmskb(tmp_V8c); - (void) __builtin_ia32_movntps(tmp_fp, tmp_V4f); (void) __builtin_ia32_movntq(tmp_V1LLip, tmp_V1LLi); (void) __builtin_ia32_sfence(); @@ -318,8 +317,6 @@ void f0() { #ifdef USE_64 (void) __builtin_ia32_movnti64(tmp_LLip, tmp_LLi); #endif - (void) __builtin_ia32_movntpd(tmp_dp, tmp_V2d); - (void) __builtin_ia32_movntdq(tmp_V2LLip, tmp_V2LLi); tmp_V2LLi = __builtin_ia32_psadbw128(tmp_V16c, tmp_V16c); tmp_V2d = __builtin_ia32_sqrtpd(tmp_V2d); tmp_V2d = __builtin_ia32_sqrtsd(tmp_V2d); @@ -446,9 +443,6 @@ void f0() { tmp_V4d = __builtin_ia32_vbroadcastf128_pd256(tmp_V2dCp); tmp_V8f = __builtin_ia32_vbroadcastf128_ps256(tmp_V4fCp); tmp_V32c = __builtin_ia32_lddqu256(tmp_cCp); - __builtin_ia32_movntdq256(tmp_V4LLip, tmp_V4LLi); - __builtin_ia32_movntpd256(tmp_dp, tmp_V4d); - __builtin_ia32_movntps256(tmp_fp, tmp_V8f); tmp_V2d = __builtin_ia32_maskloadpd(tmp_V2dCp, tmp_V2LLi); tmp_V4f = __builtin_ia32_maskloadps(tmp_V4fCp, tmp_V4i); tmp_V4d = __builtin_ia32_maskloadpd256(tmp_V4dCp, tmp_V4LLi); |
