RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-11-01 00:12:50 +01:00
Code Issues Projects Releases Wiki Activity
5395ec6ee4
llvm-mirror/test/CodeGen/X86/break-sse-dep.ll

64 lines
1.5 KiB
LLVM
Raw Normal View History

Add mcpu to tests to prevent them from using AVX instructions on Sandy Bridge after r155618. llvm-svn: 155696
2012-04-27 09:11:58 +02:00
; RUN: llc < %s -mtriple=x86_64-linux -mattr=+sse2 -mcpu=nehalem | FileCheck %s
; RUN: llc < %s -mtriple=x86_64-win32 -mattr=+sse2 -mcpu=nehalem | FileCheck %s
On recent Intel u-arch's, folding loads into some unary SSE instructions can be non-optimal. To be precise, we should avoid folding loads if the instructions only update part of the destination register, and the non-updated part is not needed. e.g. cvtss2sd, sqrtss. Unfolding the load from these instructions breaks the partial register dependency and it can improve performance. e.g. movss (%rdi), %xmm0 cvtss2sd %xmm0, %xmm0 instead of cvtss2sd (%rdi), %xmm0 An alternative method to break dependency is to clear the register first. e.g. xorps %xmm0, %xmm0 cvtss2sd (%rdi), %xmm0 llvm-svn: 91672
2009-12-18 08:40:29 +01:00
define double @t1(float* nocapture %x) nounwind readonly ssp {
entry:
Remove target attribute break-sse-dep. Instead, do not fold load into sse partial update instructions unless optimizing for size. llvm-svn: 91910
2009-12-22 18:47:23 +01:00
; CHECK: t1:
Relax expressions and add explicit triplets -linux and -win32. llvm-svn: 126197
2011-02-22 08:19:20 +01:00
; CHECK: movss ([[A0:%rdi|%rcx]]), %xmm0
Fix some tests that didn't test anything. llvm-svn: 106954
2010-06-26 22:05:06 +02:00
; CHECK: cvtss2sd %xmm0, %xmm0
On recent Intel u-arch's, folding loads into some unary SSE instructions can be non-optimal. To be precise, we should avoid folding loads if the instructions only update part of the destination register, and the non-updated part is not needed. e.g. cvtss2sd, sqrtss. Unfolding the load from these instructions breaks the partial register dependency and it can improve performance. e.g. movss (%rdi), %xmm0 cvtss2sd %xmm0, %xmm0 instead of cvtss2sd (%rdi), %xmm0 An alternative method to break dependency is to clear the register first. e.g. xorps %xmm0, %xmm0 cvtss2sd (%rdi), %xmm0 llvm-svn: 91672
2009-12-18 08:40:29 +01:00
%0 = load float* %x, align 4
%1 = fpext float %0 to double
ret double %1
}
Remove target attribute break-sse-dep. Instead, do not fold load into sse partial update instructions unless optimizing for size. llvm-svn: 91910
2009-12-22 18:47:23 +01:00
define float @t2(double* nocapture %x) nounwind readonly ssp optsize {
On recent Intel u-arch's, folding loads into some unary SSE instructions can be non-optimal. To be precise, we should avoid folding loads if the instructions only update part of the destination register, and the non-updated part is not needed. e.g. cvtss2sd, sqrtss. Unfolding the load from these instructions breaks the partial register dependency and it can improve performance. e.g. movss (%rdi), %xmm0 cvtss2sd %xmm0, %xmm0 instead of cvtss2sd (%rdi), %xmm0 An alternative method to break dependency is to clear the register first. e.g. xorps %xmm0, %xmm0 cvtss2sd (%rdi), %xmm0 llvm-svn: 91672
2009-12-18 08:40:29 +01:00
entry:
Remove target attribute break-sse-dep. Instead, do not fold load into sse partial update instructions unless optimizing for size. llvm-svn: 91910
2009-12-22 18:47:23 +01:00
; CHECK: t2:
Relax expressions and add explicit triplets -linux and -win32. llvm-svn: 126197
2011-02-22 08:19:20 +01:00
; CHECK: cvtsd2ss ([[A0]]), %xmm0
On recent Intel u-arch's, folding loads into some unary SSE instructions can be non-optimal. To be precise, we should avoid folding loads if the instructions only update part of the destination register, and the non-updated part is not needed. e.g. cvtss2sd, sqrtss. Unfolding the load from these instructions breaks the partial register dependency and it can improve performance. e.g. movss (%rdi), %xmm0 cvtss2sd %xmm0, %xmm0 instead of cvtss2sd (%rdi), %xmm0 An alternative method to break dependency is to clear the register first. e.g. xorps %xmm0, %xmm0 cvtss2sd (%rdi), %xmm0 llvm-svn: 91672
2009-12-18 08:40:29 +01:00
%0 = load double* %x, align 8
%1 = fptrunc double %0 to float
ret float %1
}
Apply the SSE dependence idiom for SSE unary operations to SD instructions too, in addition to SS instructions. And add a comment about it. llvm-svn: 108191
2010-07-12 22:46:04 +02:00
define float @squirtf(float* %x) nounwind {
entry:
; CHECK: squirtf:
Relax expressions and add explicit triplets -linux and -win32. llvm-svn: 126197
2011-02-22 08:19:20 +01:00
; CHECK: movss ([[A0]]), %xmm0
Apply the SSE dependence idiom for SSE unary operations to SD instructions too, in addition to SS instructions. And add a comment about it. llvm-svn: 108191
2010-07-12 22:46:04 +02:00
; CHECK: sqrtss %xmm0, %xmm0
%z = load float* %x
%t = call float @llvm.sqrt.f32(float %z)
ret float %t
}
define double @squirt(double* %x) nounwind {
entry:
; CHECK: squirt:
Relax expressions and add explicit triplets -linux and -win32. llvm-svn: 126197
2011-02-22 08:19:20 +01:00
; CHECK: movsd ([[A0]]), %xmm0
Apply the SSE dependence idiom for SSE unary operations to SD instructions too, in addition to SS instructions. And add a comment about it. llvm-svn: 108191
2010-07-12 22:46:04 +02:00
; CHECK: sqrtsd %xmm0, %xmm0
%z = load double* %x
%t = call double @llvm.sqrt.f64(double %z)
ret double %t
}
define float @squirtf_size(float* %x) nounwind optsize {
entry:
; CHECK: squirtf_size:
Relax expressions and add explicit triplets -linux and -win32. llvm-svn: 126197
2011-02-22 08:19:20 +01:00
; CHECK: sqrtss ([[A0]]), %xmm0
Apply the SSE dependence idiom for SSE unary operations to SD instructions too, in addition to SS instructions. And add a comment about it. llvm-svn: 108191
2010-07-12 22:46:04 +02:00
%z = load float* %x
%t = call float @llvm.sqrt.f32(float %z)
ret float %t
}
define double @squirt_size(double* %x) nounwind optsize {
entry:
; CHECK: squirt_size:
Relax expressions and add explicit triplets -linux and -win32. llvm-svn: 126197
2011-02-22 08:19:20 +01:00
; CHECK: sqrtsd ([[A0]]), %xmm0
Apply the SSE dependence idiom for SSE unary operations to SD instructions too, in addition to SS instructions. And add a comment about it. llvm-svn: 108191
2010-07-12 22:46:04 +02:00
%z = load double* %x
%t = call double @llvm.sqrt.f64(double %z)
ret double %t
}
declare float @llvm.sqrt.f32(float)
declare double @llvm.sqrt.f64(double)
Reference in New Issue Copy Permalink