maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity
llvm-project/llvm/test/CodeGen/X86/break-false-dep.ll

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

358 lines
15 KiB
LLVM
Raw Normal View History

[X86] Add missing memory variants to AVX false dependency breaking Adds missing memory instruction variants to AVX false dependency breaking handling. (SSE was handled in r224246) Differential Revision: http://reviews.llvm.org/D6780 llvm-svn: 224900
2014-12-28 21:15:05 +08:00
; RUN: llc < %s -mtriple=x86_64-linux -mattr=+sse2 -mcpu=nehalem | FileCheck %s --check-prefix=SSE
; RUN: llc < %s -mtriple=x86_64-win32 -mattr=+sse2 -mcpu=nehalem | FileCheck %s --check-prefix=SSE
; RUN: llc < %s -mtriple=x86_64-win32 -mattr=+avx -mcpu=corei7-avx | FileCheck %s --check-prefix=AVX
[X86] Add avx512vl command line to break-false-dep.ll llvm-svn: 336169
2018-07-03 12:43:49 +08:00
; RUN: llc < %s -mtriple=x86_64-win32 -mattr=+avx512vl -mcpu=skx | FileCheck %s --check-prefix=AVX
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 15:40:29 +08:00
define double @t1(float* nocapture %x) nounwind readonly ssp {
entry:
[X86] Add missing memory variants to AVX false dependency breaking Adds missing memory instruction variants to AVX false dependency breaking handling. (SSE was handled in r224246) Differential Revision: http://reviews.llvm.org/D6780 llvm-svn: 224900
2014-12-28 21:15:05 +08:00
; SSE-LABEL: t1:
; SSE: movss ([[A0:%rdi|%rcx]]), %xmm0
; SSE: 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 15:40:29 +08:00
[opaque pointer type] Add textual IR support for explicit type parameter to load instruction Essentially the same as the GEP change in r230786. A similar migration script can be used to update test cases, though a few more test case improvements/changes were required this time around: (r229269-r229278) import fileinput import sys import re pat = re.compile(r"((?:=|:|^)\s*load (?:atomic )?(?:volatile )?(.*?))(| addrspace\(\d+\) *)\*($| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$)") for line in sys.stdin: sys.stdout.write(re.sub(pat, r"\1, \2\3*\4", line)) Reviewers: rafael, dexonsmith, grosser Differential Revision: http://reviews.llvm.org/D7649 llvm-svn: 230794
2015-02-28 05:17:42 +08:00
%0 = load float, float* %x, align 4
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 15:40:29 +08:00
%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
2009-12-23 01:47:23 +08:00
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 15:40:29 +08:00
entry:
[X86] Add missing memory variants to AVX false dependency breaking Adds missing memory instruction variants to AVX false dependency breaking handling. (SSE was handled in r224246) Differential Revision: http://reviews.llvm.org/D6780 llvm-svn: 224900
2014-12-28 21:15:05 +08:00
; SSE-LABEL: t2:
; SSE: cvtsd2ss ([[A0]]), %xmm0
[opaque pointer type] Add textual IR support for explicit type parameter to load instruction Essentially the same as the GEP change in r230786. A similar migration script can be used to update test cases, though a few more test case improvements/changes were required this time around: (r229269-r229278) import fileinput import sys import re pat = re.compile(r"((?:=|:|^)\s*load (?:atomic )?(?:volatile )?(.*?))(| addrspace\(\d+\) *)\*($| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$)") for line in sys.stdin: sys.stdout.write(re.sub(pat, r"\1, \2\3*\4", line)) Reviewers: rafael, dexonsmith, grosser Differential Revision: http://reviews.llvm.org/D7649 llvm-svn: 230794
2015-02-28 05:17:42 +08:00
%0 = load double, double* %x, align 8
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 15:40:29 +08:00
%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
2010-07-13 04:46:04 +08:00
define float @squirtf(float* %x) nounwind {
entry:
[X86] Add missing memory variants to AVX false dependency breaking Adds missing memory instruction variants to AVX false dependency breaking handling. (SSE was handled in r224246) Differential Revision: http://reviews.llvm.org/D6780 llvm-svn: 224900
2014-12-28 21:15:05 +08:00
; SSE-LABEL: squirtf:
; SSE: movss ([[A0]]), %xmm0
; SSE: sqrtss %xmm0, %xmm0
[opaque pointer type] Add textual IR support for explicit type parameter to load instruction Essentially the same as the GEP change in r230786. A similar migration script can be used to update test cases, though a few more test case improvements/changes were required this time around: (r229269-r229278) import fileinput import sys import re pat = re.compile(r"((?:=|:|^)\s*load (?:atomic )?(?:volatile )?(.*?))(| addrspace\(\d+\) *)\*($| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$)") for line in sys.stdin: sys.stdout.write(re.sub(pat, r"\1, \2\3*\4", line)) Reviewers: rafael, dexonsmith, grosser Differential Revision: http://reviews.llvm.org/D7649 llvm-svn: 230794
2015-02-28 05:17:42 +08:00
%z = load float, float* %x
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
2010-07-13 04:46:04 +08:00
%t = call float @llvm.sqrt.f32(float %z)
ret float %t
}
define double @squirt(double* %x) nounwind {
entry:
[X86] Add missing memory variants to AVX false dependency breaking Adds missing memory instruction variants to AVX false dependency breaking handling. (SSE was handled in r224246) Differential Revision: http://reviews.llvm.org/D6780 llvm-svn: 224900
2014-12-28 21:15:05 +08:00
; SSE-LABEL: squirt:
; SSE: movsd ([[A0]]), %xmm0
; SSE: sqrtsd %xmm0, %xmm0
[opaque pointer type] Add textual IR support for explicit type parameter to load instruction Essentially the same as the GEP change in r230786. A similar migration script can be used to update test cases, though a few more test case improvements/changes were required this time around: (r229269-r229278) import fileinput import sys import re pat = re.compile(r"((?:=|:|^)\s*load (?:atomic )?(?:volatile )?(.*?))(| addrspace\(\d+\) *)\*($| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$)") for line in sys.stdin: sys.stdout.write(re.sub(pat, r"\1, \2\3*\4", line)) Reviewers: rafael, dexonsmith, grosser Differential Revision: http://reviews.llvm.org/D7649 llvm-svn: 230794
2015-02-28 05:17:42 +08:00
%z = load double, double* %x
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
2010-07-13 04:46:04 +08:00
%t = call double @llvm.sqrt.f64(double %z)
ret double %t
}
define float @squirtf_size(float* %x) nounwind optsize {
entry:
[X86] Add missing memory variants to AVX false dependency breaking Adds missing memory instruction variants to AVX false dependency breaking handling. (SSE was handled in r224246) Differential Revision: http://reviews.llvm.org/D6780 llvm-svn: 224900
2014-12-28 21:15:05 +08:00
; SSE-LABEL: squirtf_size:
; SSE: sqrtss ([[A0]]), %xmm0
[opaque pointer type] Add textual IR support for explicit type parameter to load instruction Essentially the same as the GEP change in r230786. A similar migration script can be used to update test cases, though a few more test case improvements/changes were required this time around: (r229269-r229278) import fileinput import sys import re pat = re.compile(r"((?:=|:|^)\s*load (?:atomic )?(?:volatile )?(.*?))(| addrspace\(\d+\) *)\*($| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$)") for line in sys.stdin: sys.stdout.write(re.sub(pat, r"\1, \2\3*\4", line)) Reviewers: rafael, dexonsmith, grosser Differential Revision: http://reviews.llvm.org/D7649 llvm-svn: 230794
2015-02-28 05:17:42 +08:00
%z = load float, float* %x
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
2010-07-13 04:46:04 +08:00
%t = call float @llvm.sqrt.f32(float %z)
ret float %t
}
define double @squirt_size(double* %x) nounwind optsize {
entry:
[X86] Add missing memory variants to AVX false dependency breaking Adds missing memory instruction variants to AVX false dependency breaking handling. (SSE was handled in r224246) Differential Revision: http://reviews.llvm.org/D6780 llvm-svn: 224900
2014-12-28 21:15:05 +08:00
; SSE-LABEL: squirt_size:
; SSE: sqrtsd ([[A0]]), %xmm0
[opaque pointer type] Add textual IR support for explicit type parameter to load instruction Essentially the same as the GEP change in r230786. A similar migration script can be used to update test cases, though a few more test case improvements/changes were required this time around: (r229269-r229278) import fileinput import sys import re pat = re.compile(r"((?:=|:|^)\s*load (?:atomic )?(?:volatile )?(.*?))(| addrspace\(\d+\) *)\*($| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$)") for line in sys.stdin: sys.stdout.write(re.sub(pat, r"\1, \2\3*\4", line)) Reviewers: rafael, dexonsmith, grosser Differential Revision: http://reviews.llvm.org/D7649 llvm-svn: 230794
2015-02-28 05:17:42 +08:00
%z = load double, double* %x
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
2010-07-13 04:46:04 +08:00
%t = call double @llvm.sqrt.f64(double %z)
ret double %t
}
declare float @llvm.sqrt.f32(float)
declare double @llvm.sqrt.f64(double)
[X86] Break false dependencies before partial register updates when the source operand is in memory Adds the various "rm" instruction variants into the list of instructions that have a partial register update. Also adds all variants of SQRTSD that were missing in the original list. Differential Revision: http://reviews.llvm.org/D6620 llvm-svn: 224246
2014-12-15 21:18:21 +08:00
[X86] Add missing memory variants to AVX false dependency breaking Adds missing memory instruction variants to AVX false dependency breaking handling. (SSE was handled in r224246) Differential Revision: http://reviews.llvm.org/D6780 llvm-svn: 224900
2014-12-28 21:15:05 +08:00
; SSE-LABEL: loopdep1
Don't delete empty preheaders in CodeGenPrepare if it would create a critical edge Presently, CodeGenPrepare deletes all nearly empty (only phi and branch) basic blocks. This pass can delete loop preheaders which frequently creates critical edges. A preheader can be a convenient place to spill registers to the stack. If the entrance to a loop body is a critical edge, then spills may occur in the loop body rather than immediately before it. This patch protects loop preheaders from deletion in CodeGenPrepare even if they are nearly empty. Since the patch alters the CFG, it affects a large number of test cases. In most cases, the changes are merely cosmetic (basic blocks have different names or instruction orders change slightly). I am somewhat concerned about the test/CodeGen/Mips/brdelayslot.ll test case. If the loop preheader is not deleted, then the MIPS backend does not take advantage of a branch delay slot. Consequently, I would like some close review by a MIPS expert. The patch also partially subsumes D16893 from George Burgess IV. George correctly notes that CodeGenPrepare does not actually preserve the dominator tree. I think the dominator tree was usually not valid when CodeGenPrepare ran, but I am using LoopInfo to mark preheaders, so the dominator tree is now always valid before CodeGenPrepare. Author: Tom Jablin (tjablin) Reviewers: hfinkel george.burgess.iv vkalintiris dsanders kbarton cycheng http://reviews.llvm.org/D16984 llvm-svn: 265397
2016-04-05 22:06:20 +08:00
; SSE: for.body{{$}}
[X86] Break false dependencies before partial register updates when the source operand is in memory Adds the various "rm" instruction variants into the list of instructions that have a partial register update. Also adds all variants of SQRTSD that were missing in the original list. Differential Revision: http://reviews.llvm.org/D6620 llvm-svn: 224246
2014-12-15 21:18:21 +08:00
;
; This loop contains two cvtsi2ss instructions that update the same xmm
Separate ExecutionDepsFix into 4 parts: 1. ReachingDefsAnalysis - Allows to identify for each instruction what is the “closest” reaching def of a certain register. Used by BreakFalseDeps (for clearance calculation) and ExecutionDomainFix (for arbitrating conflicting domains). 2. ExecutionDomainFix - Changes the variant of the instructions in order to minimize domain crossings. 3. BreakFalseDeps - Breaks false dependencies. 4. LoopTraversal - Creatws a traversal order of the basic blocks that is optimal for loops (introduced in revision L293571). Both ExecutionDomainFix and ReachingDefsAnalysis use this to determine the order they will traverse the basic blocks. This also included the following changes to ExcecutionDepsFix original logic: 1. BreakFalseDeps and ReachingDefsAnalysis logic no longer restricted by a register class. 2. ReachingDefsAnalysis tracks liveness of reg units instead of reg indices into a given reg class. Additional changes in affected files: 1. X86 and ARM targets now inherit from ExecutionDomainFix instead of ExecutionDepsFix. BreakFalseDeps also was added to the passes they activate. 2. Comments and references to ExecutionDepsFix replaced with ExecutionDomainFix and BreakFalseDeps, as appropriate. Additional refactoring changes will follow. This commit is (almost) NFC. The only functional change is that now BreakFalseDeps will break dependency for all register classes. Since no additional instructions were added to the list of instructions that have false dependencies, there is no actual change yet. In a future commit several instructions (and tests) will be added. This is the first of multiple patches that fix bugzilla https://bugs.llvm.org/show_bug.cgi?id=33869 Most of the patches are intended at refactoring the existent code. Additional relevant reviews: https://reviews.llvm.org/D40331 https://reviews.llvm.org/D40332 https://reviews.llvm.org/D40333 https://reviews.llvm.org/D40334 Differential Revision: https://reviews.llvm.org/D40330 Change-Id: Icaeb75e014eff96a8f721377783f9a3e6c679275 llvm-svn: 323087
2018-01-22 18:05:23 +08:00
; register. Verify that the break false dependency fix pass breaks those
[X86] Break false dependencies before partial register updates when the source operand is in memory Adds the various "rm" instruction variants into the list of instructions that have a partial register update. Also adds all variants of SQRTSD that were missing in the original list. Differential Revision: http://reviews.llvm.org/D6620 llvm-svn: 224246
2014-12-15 21:18:21 +08:00
; dependencies by inserting xorps instructions.
;
; If the register allocator chooses different registers for the two cvtsi2ss
; instructions, they are still dependent on themselves.
[X86] Add missing memory variants to AVX false dependency breaking Adds missing memory instruction variants to AVX false dependency breaking handling. (SSE was handled in r224246) Differential Revision: http://reviews.llvm.org/D6780 llvm-svn: 224900
2014-12-28 21:15:05 +08:00
; SSE: xorps [[XMM1:%xmm[0-9]+]]
; SSE: , [[XMM1]]
[X86] Remove the suffix on vcvt[u]si2ss/sd register variants in assembly printing. We require d/q suffixes on the memory form of these instructions to disambiguate the memory size. We don't require it on the register forms, but need to support parsing both with and without it. Previously we always printed the d/q suffix on the register forms, but it's redundant and inconsistent with gcc and objdump. After this patch we should support the d/q for parsing, but not print it when its unneeded. llvm-svn: 360085
2019-05-07 05:39:51 +08:00
; SSE: cvtsi2ss %{{.*}}, [[XMM1]]
[X86] Add missing memory variants to AVX false dependency breaking Adds missing memory instruction variants to AVX false dependency breaking handling. (SSE was handled in r224246) Differential Revision: http://reviews.llvm.org/D6780 llvm-svn: 224900
2014-12-28 21:15:05 +08:00
; SSE: xorps [[XMM2:%xmm[0-9]+]]
; SSE: , [[XMM2]]
[X86] Remove the suffix on vcvt[u]si2ss/sd register variants in assembly printing. We require d/q suffixes on the memory form of these instructions to disambiguate the memory size. We don't require it on the register forms, but need to support parsing both with and without it. Previously we always printed the d/q suffix on the register forms, but it's redundant and inconsistent with gcc and objdump. After this patch we should support the d/q for parsing, but not print it when its unneeded. llvm-svn: 360085
2019-05-07 05:39:51 +08:00
; SSE: cvtsi2ss %{{.*}}, [[XMM2]]
[X86] Break false dependencies before partial register updates when the source operand is in memory Adds the various "rm" instruction variants into the list of instructions that have a partial register update. Also adds all variants of SQRTSD that were missing in the original list. Differential Revision: http://reviews.llvm.org/D6620 llvm-svn: 224246
2014-12-15 21:18:21 +08:00
define float @loopdep1(i32 %m) nounwind uwtable readnone ssp {
entry:
%tobool3 = icmp eq i32 %m, 0
br i1 %tobool3, label %for.end, label %for.body
for.body: ; preds = %entry, %for.body
%m.addr.07 = phi i32 [ %dec, %for.body ], [ %m, %entry ]
%s1.06 = phi float [ %add, %for.body ], [ 0.000000e+00, %entry ]
%s2.05 = phi float [ %add2, %for.body ], [ 0.000000e+00, %entry ]
%n.04 = phi i32 [ %inc, %for.body ], [ 1, %entry ]
%conv = sitofp i32 %n.04 to float
%add = fadd float %s1.06, %conv
%conv1 = sitofp i32 %m.addr.07 to float
%add2 = fadd float %s2.05, %conv1
%inc = add nsw i32 %n.04, 1
%dec = add nsw i32 %m.addr.07, -1
%tobool = icmp eq i32 %dec, 0
br i1 %tobool, label %for.end, label %for.body
for.end: ; preds = %for.body, %entry
%s1.0.lcssa = phi float [ 0.000000e+00, %entry ], [ %add, %for.body ]
%s2.0.lcssa = phi float [ 0.000000e+00, %entry ], [ %add2, %for.body ]
%sub = fsub float %s1.0.lcssa, %s2.0.lcssa
ret float %sub
}
[X86] Add missing memory variants to AVX false dependency breaking Adds missing memory instruction variants to AVX false dependency breaking handling. (SSE was handled in r224246) Differential Revision: http://reviews.llvm.org/D6780 llvm-svn: 224900
2014-12-28 21:15:05 +08:00
; rdar:15221834 False AVX register dependencies cause 5x slowdown on
[X86] Remove the suffix on vcvt[u]si2ss/sd register variants in assembly printing. We require d/q suffixes on the memory form of these instructions to disambiguate the memory size. We don't require it on the register forms, but need to support parsing both with and without it. Previously we always printed the d/q suffix on the register forms, but it's redundant and inconsistent with gcc and objdump. After this patch we should support the d/q for parsing, but not print it when its unneeded. llvm-svn: 360085
2019-05-07 05:39:51 +08:00
; flops-6. Make sure the unused register read by vcvtsi2sd is zeroed
[X86] Add missing memory variants to AVX false dependency breaking Adds missing memory instruction variants to AVX false dependency breaking handling. (SSE was handled in r224246) Differential Revision: http://reviews.llvm.org/D6780 llvm-svn: 224900
2014-12-28 21:15:05 +08:00
; to avoid cyclic dependence on a write to the same register in a
; previous iteration.
; AVX-LABEL: loopdep2:
; AVX-LABEL: %loop
; AVX: vxorps %[[REG:xmm.]], %{{xmm.}}, %{{xmm.}}
[X86] Remove the suffix on vcvt[u]si2ss/sd register variants in assembly printing. We require d/q suffixes on the memory form of these instructions to disambiguate the memory size. We don't require it on the register forms, but need to support parsing both with and without it. Previously we always printed the d/q suffix on the register forms, but it's redundant and inconsistent with gcc and objdump. After this patch we should support the d/q for parsing, but not print it when its unneeded. llvm-svn: 360085
2019-05-07 05:39:51 +08:00
; AVX: vcvtsi2sd %{{r[0-9a-x]+}}, %[[REG]], %{{xmm.}}
[X86] Add missing memory variants to AVX false dependency breaking Adds missing memory instruction variants to AVX false dependency breaking handling. (SSE was handled in r224246) Differential Revision: http://reviews.llvm.org/D6780 llvm-svn: 224900
2014-12-28 21:15:05 +08:00
; SSE-LABEL: loopdep2:
; SSE-LABEL: %loop
; SSE: xorps %[[REG:xmm.]], %[[REG]]
[X86] Remove the suffix on vcvt[u]si2ss/sd register variants in assembly printing. We require d/q suffixes on the memory form of these instructions to disambiguate the memory size. We don't require it on the register forms, but need to support parsing both with and without it. Previously we always printed the d/q suffix on the register forms, but it's redundant and inconsistent with gcc and objdump. After this patch we should support the d/q for parsing, but not print it when its unneeded. llvm-svn: 360085
2019-05-07 05:39:51 +08:00
; SSE: cvtsi2sd %{{r[0-9a-x]+}}, %[[REG]]
[X86] Add missing memory variants to AVX false dependency breaking Adds missing memory instruction variants to AVX false dependency breaking handling. (SSE was handled in r224246) Differential Revision: http://reviews.llvm.org/D6780 llvm-svn: 224900
2014-12-28 21:15:05 +08:00
define i64 @loopdep2(i64* nocapture %x, double* nocapture %y) nounwind {
entry:
[opaque pointer type] Add textual IR support for explicit type parameter to load instruction Essentially the same as the GEP change in r230786. A similar migration script can be used to update test cases, though a few more test case improvements/changes were required this time around: (r229269-r229278) import fileinput import sys import re pat = re.compile(r"((?:=|:|^)\s*load (?:atomic )?(?:volatile )?(.*?))(| addrspace\(\d+\) *)\*($| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$)") for line in sys.stdin: sys.stdout.write(re.sub(pat, r"\1, \2\3*\4", line)) Reviewers: rafael, dexonsmith, grosser Differential Revision: http://reviews.llvm.org/D7649 llvm-svn: 230794
2015-02-28 05:17:42 +08:00
%vx = load i64, i64* %x
[X86] Add missing memory variants to AVX false dependency breaking Adds missing memory instruction variants to AVX false dependency breaking handling. (SSE was handled in r224246) Differential Revision: http://reviews.llvm.org/D6780 llvm-svn: 224900
2014-12-28 21:15:05 +08:00
br label %loop
loop:
%i = phi i64 [ 1, %entry ], [ %inc, %loop ]
%s1 = phi i64 [ %vx, %entry ], [ %s2, %loop ]
%fi = sitofp i64 %i to double
[X86] Add avx512vl command line to break-false-dep.ll llvm-svn: 336169
2018-07-03 12:43:49 +08:00
tail call void asm sideeffect "", "~{xmm0},~{xmm1},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{dirflag},~{fpsr},~{flags}"()
[opaque pointer type] Add textual IR support for explicit type parameter to load instruction Essentially the same as the GEP change in r230786. A similar migration script can be used to update test cases, though a few more test case improvements/changes were required this time around: (r229269-r229278) import fileinput import sys import re pat = re.compile(r"((?:=|:|^)\s*load (?:atomic )?(?:volatile )?(.*?))(| addrspace\(\d+\) *)\*($| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$)") for line in sys.stdin: sys.stdout.write(re.sub(pat, r"\1, \2\3*\4", line)) Reviewers: rafael, dexonsmith, grosser Differential Revision: http://reviews.llvm.org/D7649 llvm-svn: 230794
2015-02-28 05:17:42 +08:00
%vy = load double, double* %y
[X86] Add missing memory variants to AVX false dependency breaking Adds missing memory instruction variants to AVX false dependency breaking handling. (SSE was handled in r224246) Differential Revision: http://reviews.llvm.org/D6780 llvm-svn: 224900
2014-12-28 21:15:05 +08:00
%fipy = fadd double %fi, %vy
%iipy = fptosi double %fipy to i64
%s2 = add i64 %s1, %iipy
%inc = add nsw i64 %i, 1
%exitcond = icmp eq i64 %inc, 156250000
br i1 %exitcond, label %ret, label %loop
ret:
ret i64 %s2
}
[X86] Break false dependencies before partial register updates when the source operand is in memory Adds the various "rm" instruction variants into the list of instructions that have a partial register update. Also adds all variants of SQRTSD that were missing in the original list. Differential Revision: http://reviews.llvm.org/D6620 llvm-svn: 224246
2014-12-15 21:18:21 +08:00
; This loop contains a cvtsi2sd instruction that has a loop-carried
; false dependency on an xmm that is modified by other scalar instructions
Don't delete empty preheaders in CodeGenPrepare if it would create a critical edge Presently, CodeGenPrepare deletes all nearly empty (only phi and branch) basic blocks. This pass can delete loop preheaders which frequently creates critical edges. A preheader can be a convenient place to spill registers to the stack. If the entrance to a loop body is a critical edge, then spills may occur in the loop body rather than immediately before it. This patch protects loop preheaders from deletion in CodeGenPrepare even if they are nearly empty. Since the patch alters the CFG, it affects a large number of test cases. In most cases, the changes are merely cosmetic (basic blocks have different names or instruction orders change slightly). I am somewhat concerned about the test/CodeGen/Mips/brdelayslot.ll test case. If the loop preheader is not deleted, then the MIPS backend does not take advantage of a branch delay slot. Consequently, I would like some close review by a MIPS expert. The patch also partially subsumes D16893 from George Burgess IV. George correctly notes that CodeGenPrepare does not actually preserve the dominator tree. I think the dominator tree was usually not valid when CodeGenPrepare ran, but I am using LoopInfo to mark preheaders, so the dominator tree is now always valid before CodeGenPrepare. Author: Tom Jablin (tjablin) Reviewers: hfinkel george.burgess.iv vkalintiris dsanders kbarton cycheng http://reviews.llvm.org/D16984 llvm-svn: 265397
2016-04-05 22:06:20 +08:00
; that follow it in the loop. Additionally, the source of convert is a
Separate ExecutionDepsFix into 4 parts: 1. ReachingDefsAnalysis - Allows to identify for each instruction what is the “closest” reaching def of a certain register. Used by BreakFalseDeps (for clearance calculation) and ExecutionDomainFix (for arbitrating conflicting domains). 2. ExecutionDomainFix - Changes the variant of the instructions in order to minimize domain crossings. 3. BreakFalseDeps - Breaks false dependencies. 4. LoopTraversal - Creatws a traversal order of the basic blocks that is optimal for loops (introduced in revision L293571). Both ExecutionDomainFix and ReachingDefsAnalysis use this to determine the order they will traverse the basic blocks. This also included the following changes to ExcecutionDepsFix original logic: 1. BreakFalseDeps and ReachingDefsAnalysis logic no longer restricted by a register class. 2. ReachingDefsAnalysis tracks liveness of reg units instead of reg indices into a given reg class. Additional changes in affected files: 1. X86 and ARM targets now inherit from ExecutionDomainFix instead of ExecutionDepsFix. BreakFalseDeps also was added to the passes they activate. 2. Comments and references to ExecutionDepsFix replaced with ExecutionDomainFix and BreakFalseDeps, as appropriate. Additional refactoring changes will follow. This commit is (almost) NFC. The only functional change is that now BreakFalseDeps will break dependency for all register classes. Since no additional instructions were added to the list of instructions that have false dependencies, there is no actual change yet. In a future commit several instructions (and tests) will be added. This is the first of multiple patches that fix bugzilla https://bugs.llvm.org/show_bug.cgi?id=33869 Most of the patches are intended at refactoring the existent code. Additional relevant reviews: https://reviews.llvm.org/D40331 https://reviews.llvm.org/D40332 https://reviews.llvm.org/D40333 https://reviews.llvm.org/D40334 Differential Revision: https://reviews.llvm.org/D40330 Change-Id: Icaeb75e014eff96a8f721377783f9a3e6c679275 llvm-svn: 323087
2018-01-22 18:05:23 +08:00
; memory operand. Verify the break false dependency fix pass breaks this
[X86] Break false dependencies before partial register updates when the source operand is in memory Adds the various "rm" instruction variants into the list of instructions that have a partial register update. Also adds all variants of SQRTSD that were missing in the original list. Differential Revision: http://reviews.llvm.org/D6620 llvm-svn: 224246
2014-12-15 21:18:21 +08:00
; dependency by inserting a xor before the convert.
@x = common global [1024 x double] zeroinitializer, align 16
@y = common global [1024 x double] zeroinitializer, align 16
@z = common global [1024 x double] zeroinitializer, align 16
@w = common global [1024 x double] zeroinitializer, align 16
@v = common global [1024 x i32] zeroinitializer, align 16
[X86] Add missing memory variants to AVX false dependency breaking Adds missing memory instruction variants to AVX false dependency breaking handling. (SSE was handled in r224246) Differential Revision: http://reviews.llvm.org/D6780 llvm-svn: 224900
2014-12-28 21:15:05 +08:00
define void @loopdep3() {
[X86] Break false dependencies before partial register updates when the source operand is in memory Adds the various "rm" instruction variants into the list of instructions that have a partial register update. Also adds all variants of SQRTSD that were missing in the original list. Differential Revision: http://reviews.llvm.org/D6620 llvm-svn: 224246
2014-12-15 21:18:21 +08:00
entry:
br label %for.cond1.preheader
for.cond1.preheader: ; preds = %for.inc14, %entry
%i.025 = phi i32 [ 0, %entry ], [ %inc15, %for.inc14 ]
br label %for.body3
for.body3:
%indvars.iv = phi i64 [ 0, %for.cond1.preheader ], [ %indvars.iv.next, %for.body3 ]
[opaque pointer type] Add textual IR support for explicit type parameter to getelementptr instruction One of several parallel first steps to remove the target type of pointers, replacing them with a single opaque pointer type. This adds an explicit type parameter to the gep instruction so that when the first parameter becomes an opaque pointer type, the type to gep through is still available to the instructions. * This doesn't modify gep operators, only instructions (operators will be handled separately) * Textual IR changes only. Bitcode (including upgrade) and changing the in-memory representation will be in separate changes. * geps of vectors are transformed as: getelementptr <4 x float*> %x, ... ->getelementptr float, <4 x float*> %x, ... Then, once the opaque pointer type is introduced, this will ultimately look like: getelementptr float, <4 x ptr> %x with the unambiguous interpretation that it is a vector of pointers to float. * address spaces remain on the pointer, not the type: getelementptr float addrspace(1)* %x ->getelementptr float, float addrspace(1)* %x Then, eventually: getelementptr float, ptr addrspace(1) %x Importantly, the massive amount of test case churn has been automated by same crappy python code. I had to manually update a few test cases that wouldn't fit the script's model (r228970,r229196,r229197,r229198). The python script just massages stdin and writes the result to stdout, I then wrapped that in a shell script to handle replacing files, then using the usual find+xargs to migrate all the files. update.py: import fileinput import sys import re ibrep = re.compile(r"(^.*?[^%\w]getelementptr inbounds )(((?:<\d* x )?)(.*?)(| addrspace\(\d\)) *\*(|>)(?:$| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$))") normrep = re.compile( r"(^.*?[^%\w]getelementptr )(((?:<\d* x )?)(.*?)(| addrspace\(\d\)) *\*(|>)(?:$| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$))") def conv(match, line): if not match: return line line = match.groups()[0] if len(match.groups()[5]) == 0: line += match.groups()[2] line += match.groups()[3] line += ", " line += match.groups()[1] line += "\n" return line for line in sys.stdin: if line.find("getelementptr ") == line.find("getelementptr inbounds"): if line.find("getelementptr inbounds") != line.find("getelementptr inbounds ("): line = conv(re.match(ibrep, line), line) elif line.find("getelementptr ") != line.find("getelementptr ("): line = conv(re.match(normrep, line), line) sys.stdout.write(line) apply.sh: for name in "$@" do python3 `dirname "$0"`/update.py < "$name" > "$name.tmp" && mv "$name.tmp" "$name" rm -f "$name.tmp" done The actual commands: From llvm/src: find test/ -name *.ll | xargs ./apply.sh From llvm/src/tools/clang: find test/ -name *.mm -o -name *.m -o -name *.cpp -o -name *.c | xargs -I '{}' ../../apply.sh "{}" From llvm/src/tools/polly: find test/ -name *.ll | xargs ./apply.sh After that, check-all (with llvm, clang, clang-tools-extra, lld, compiler-rt, and polly all checked out). The extra 'rm' in the apply.sh script is due to a few files in clang's test suite using interesting unicode stuff that my python script was throwing exceptions on. None of those files needed to be migrated, so it seemed sufficient to ignore those cases. Reviewers: rafael, dexonsmith, grosser Differential Revision: http://reviews.llvm.org/D7636 llvm-svn: 230786
2015-02-28 03:29:02 +08:00
%arrayidx = getelementptr inbounds [1024 x i32], [1024 x i32]* @v, i64 0, i64 %indvars.iv
[opaque pointer type] Add textual IR support for explicit type parameter to load instruction Essentially the same as the GEP change in r230786. A similar migration script can be used to update test cases, though a few more test case improvements/changes were required this time around: (r229269-r229278) import fileinput import sys import re pat = re.compile(r"((?:=|:|^)\s*load (?:atomic )?(?:volatile )?(.*?))(| addrspace\(\d+\) *)\*($| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$)") for line in sys.stdin: sys.stdout.write(re.sub(pat, r"\1, \2\3*\4", line)) Reviewers: rafael, dexonsmith, grosser Differential Revision: http://reviews.llvm.org/D7649 llvm-svn: 230794
2015-02-28 05:17:42 +08:00
%0 = load i32, i32* %arrayidx, align 4
[X86] Break false dependencies before partial register updates when the source operand is in memory Adds the various "rm" instruction variants into the list of instructions that have a partial register update. Also adds all variants of SQRTSD that were missing in the original list. Differential Revision: http://reviews.llvm.org/D6620 llvm-svn: 224246
2014-12-15 21:18:21 +08:00
%conv = sitofp i32 %0 to double
[opaque pointer type] Add textual IR support for explicit type parameter to getelementptr instruction One of several parallel first steps to remove the target type of pointers, replacing them with a single opaque pointer type. This adds an explicit type parameter to the gep instruction so that when the first parameter becomes an opaque pointer type, the type to gep through is still available to the instructions. * This doesn't modify gep operators, only instructions (operators will be handled separately) * Textual IR changes only. Bitcode (including upgrade) and changing the in-memory representation will be in separate changes. * geps of vectors are transformed as: getelementptr <4 x float*> %x, ... ->getelementptr float, <4 x float*> %x, ... Then, once the opaque pointer type is introduced, this will ultimately look like: getelementptr float, <4 x ptr> %x with the unambiguous interpretation that it is a vector of pointers to float. * address spaces remain on the pointer, not the type: getelementptr float addrspace(1)* %x ->getelementptr float, float addrspace(1)* %x Then, eventually: getelementptr float, ptr addrspace(1) %x Importantly, the massive amount of test case churn has been automated by same crappy python code. I had to manually update a few test cases that wouldn't fit the script's model (r228970,r229196,r229197,r229198). The python script just massages stdin and writes the result to stdout, I then wrapped that in a shell script to handle replacing files, then using the usual find+xargs to migrate all the files. update.py: import fileinput import sys import re ibrep = re.compile(r"(^.*?[^%\w]getelementptr inbounds )(((?:<\d* x )?)(.*?)(| addrspace\(\d\)) *\*(|>)(?:$| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$))") normrep = re.compile( r"(^.*?[^%\w]getelementptr )(((?:<\d* x )?)(.*?)(| addrspace\(\d\)) *\*(|>)(?:$| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$))") def conv(match, line): if not match: return line line = match.groups()[0] if len(match.groups()[5]) == 0: line += match.groups()[2] line += match.groups()[3] line += ", " line += match.groups()[1] line += "\n" return line for line in sys.stdin: if line.find("getelementptr ") == line.find("getelementptr inbounds"): if line.find("getelementptr inbounds") != line.find("getelementptr inbounds ("): line = conv(re.match(ibrep, line), line) elif line.find("getelementptr ") != line.find("getelementptr ("): line = conv(re.match(normrep, line), line) sys.stdout.write(line) apply.sh: for name in "$@" do python3 `dirname "$0"`/update.py < "$name" > "$name.tmp" && mv "$name.tmp" "$name" rm -f "$name.tmp" done The actual commands: From llvm/src: find test/ -name *.ll | xargs ./apply.sh From llvm/src/tools/clang: find test/ -name *.mm -o -name *.m -o -name *.cpp -o -name *.c | xargs -I '{}' ../../apply.sh "{}" From llvm/src/tools/polly: find test/ -name *.ll | xargs ./apply.sh After that, check-all (with llvm, clang, clang-tools-extra, lld, compiler-rt, and polly all checked out). The extra 'rm' in the apply.sh script is due to a few files in clang's test suite using interesting unicode stuff that my python script was throwing exceptions on. None of those files needed to be migrated, so it seemed sufficient to ignore those cases. Reviewers: rafael, dexonsmith, grosser Differential Revision: http://reviews.llvm.org/D7636 llvm-svn: 230786
2015-02-28 03:29:02 +08:00
%arrayidx5 = getelementptr inbounds [1024 x double], [1024 x double]* @x, i64 0, i64 %indvars.iv
[opaque pointer type] Add textual IR support for explicit type parameter to load instruction Essentially the same as the GEP change in r230786. A similar migration script can be used to update test cases, though a few more test case improvements/changes were required this time around: (r229269-r229278) import fileinput import sys import re pat = re.compile(r"((?:=|:|^)\s*load (?:atomic )?(?:volatile )?(.*?))(| addrspace\(\d+\) *)\*($| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$)") for line in sys.stdin: sys.stdout.write(re.sub(pat, r"\1, \2\3*\4", line)) Reviewers: rafael, dexonsmith, grosser Differential Revision: http://reviews.llvm.org/D7649 llvm-svn: 230794
2015-02-28 05:17:42 +08:00
%1 = load double, double* %arrayidx5, align 8
[X86] Break false dependencies before partial register updates when the source operand is in memory Adds the various "rm" instruction variants into the list of instructions that have a partial register update. Also adds all variants of SQRTSD that were missing in the original list. Differential Revision: http://reviews.llvm.org/D6620 llvm-svn: 224246
2014-12-15 21:18:21 +08:00
%mul = fmul double %conv, %1
[opaque pointer type] Add textual IR support for explicit type parameter to getelementptr instruction One of several parallel first steps to remove the target type of pointers, replacing them with a single opaque pointer type. This adds an explicit type parameter to the gep instruction so that when the first parameter becomes an opaque pointer type, the type to gep through is still available to the instructions. * This doesn't modify gep operators, only instructions (operators will be handled separately) * Textual IR changes only. Bitcode (including upgrade) and changing the in-memory representation will be in separate changes. * geps of vectors are transformed as: getelementptr <4 x float*> %x, ... ->getelementptr float, <4 x float*> %x, ... Then, once the opaque pointer type is introduced, this will ultimately look like: getelementptr float, <4 x ptr> %x with the unambiguous interpretation that it is a vector of pointers to float. * address spaces remain on the pointer, not the type: getelementptr float addrspace(1)* %x ->getelementptr float, float addrspace(1)* %x Then, eventually: getelementptr float, ptr addrspace(1) %x Importantly, the massive amount of test case churn has been automated by same crappy python code. I had to manually update a few test cases that wouldn't fit the script's model (r228970,r229196,r229197,r229198). The python script just massages stdin and writes the result to stdout, I then wrapped that in a shell script to handle replacing files, then using the usual find+xargs to migrate all the files. update.py: import fileinput import sys import re ibrep = re.compile(r"(^.*?[^%\w]getelementptr inbounds )(((?:<\d* x )?)(.*?)(| addrspace\(\d\)) *\*(|>)(?:$| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$))") normrep = re.compile( r"(^.*?[^%\w]getelementptr )(((?:<\d* x )?)(.*?)(| addrspace\(\d\)) *\*(|>)(?:$| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$))") def conv(match, line): if not match: return line line = match.groups()[0] if len(match.groups()[5]) == 0: line += match.groups()[2] line += match.groups()[3] line += ", " line += match.groups()[1] line += "\n" return line for line in sys.stdin: if line.find("getelementptr ") == line.find("getelementptr inbounds"): if line.find("getelementptr inbounds") != line.find("getelementptr inbounds ("): line = conv(re.match(ibrep, line), line) elif line.find("getelementptr ") != line.find("getelementptr ("): line = conv(re.match(normrep, line), line) sys.stdout.write(line) apply.sh: for name in "$@" do python3 `dirname "$0"`/update.py < "$name" > "$name.tmp" && mv "$name.tmp" "$name" rm -f "$name.tmp" done The actual commands: From llvm/src: find test/ -name *.ll | xargs ./apply.sh From llvm/src/tools/clang: find test/ -name *.mm -o -name *.m -o -name *.cpp -o -name *.c | xargs -I '{}' ../../apply.sh "{}" From llvm/src/tools/polly: find test/ -name *.ll | xargs ./apply.sh After that, check-all (with llvm, clang, clang-tools-extra, lld, compiler-rt, and polly all checked out). The extra 'rm' in the apply.sh script is due to a few files in clang's test suite using interesting unicode stuff that my python script was throwing exceptions on. None of those files needed to be migrated, so it seemed sufficient to ignore those cases. Reviewers: rafael, dexonsmith, grosser Differential Revision: http://reviews.llvm.org/D7636 llvm-svn: 230786
2015-02-28 03:29:02 +08:00
%arrayidx7 = getelementptr inbounds [1024 x double], [1024 x double]* @y, i64 0, i64 %indvars.iv
[opaque pointer type] Add textual IR support for explicit type parameter to load instruction Essentially the same as the GEP change in r230786. A similar migration script can be used to update test cases, though a few more test case improvements/changes were required this time around: (r229269-r229278) import fileinput import sys import re pat = re.compile(r"((?:=|:|^)\s*load (?:atomic )?(?:volatile )?(.*?))(| addrspace\(\d+\) *)\*($| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$)") for line in sys.stdin: sys.stdout.write(re.sub(pat, r"\1, \2\3*\4", line)) Reviewers: rafael, dexonsmith, grosser Differential Revision: http://reviews.llvm.org/D7649 llvm-svn: 230794
2015-02-28 05:17:42 +08:00
%2 = load double, double* %arrayidx7, align 8
[X86] Break false dependencies before partial register updates when the source operand is in memory Adds the various "rm" instruction variants into the list of instructions that have a partial register update. Also adds all variants of SQRTSD that were missing in the original list. Differential Revision: http://reviews.llvm.org/D6620 llvm-svn: 224246
2014-12-15 21:18:21 +08:00
%mul8 = fmul double %mul, %2
[opaque pointer type] Add textual IR support for explicit type parameter to getelementptr instruction One of several parallel first steps to remove the target type of pointers, replacing them with a single opaque pointer type. This adds an explicit type parameter to the gep instruction so that when the first parameter becomes an opaque pointer type, the type to gep through is still available to the instructions. * This doesn't modify gep operators, only instructions (operators will be handled separately) * Textual IR changes only. Bitcode (including upgrade) and changing the in-memory representation will be in separate changes. * geps of vectors are transformed as: getelementptr <4 x float*> %x, ... ->getelementptr float, <4 x float*> %x, ... Then, once the opaque pointer type is introduced, this will ultimately look like: getelementptr float, <4 x ptr> %x with the unambiguous interpretation that it is a vector of pointers to float. * address spaces remain on the pointer, not the type: getelementptr float addrspace(1)* %x ->getelementptr float, float addrspace(1)* %x Then, eventually: getelementptr float, ptr addrspace(1) %x Importantly, the massive amount of test case churn has been automated by same crappy python code. I had to manually update a few test cases that wouldn't fit the script's model (r228970,r229196,r229197,r229198). The python script just massages stdin and writes the result to stdout, I then wrapped that in a shell script to handle replacing files, then using the usual find+xargs to migrate all the files. update.py: import fileinput import sys import re ibrep = re.compile(r"(^.*?[^%\w]getelementptr inbounds )(((?:<\d* x )?)(.*?)(| addrspace\(\d\)) *\*(|>)(?:$| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$))") normrep = re.compile( r"(^.*?[^%\w]getelementptr )(((?:<\d* x )?)(.*?)(| addrspace\(\d\)) *\*(|>)(?:$| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$))") def conv(match, line): if not match: return line line = match.groups()[0] if len(match.groups()[5]) == 0: line += match.groups()[2] line += match.groups()[3] line += ", " line += match.groups()[1] line += "\n" return line for line in sys.stdin: if line.find("getelementptr ") == line.find("getelementptr inbounds"): if line.find("getelementptr inbounds") != line.find("getelementptr inbounds ("): line = conv(re.match(ibrep, line), line) elif line.find("getelementptr ") != line.find("getelementptr ("): line = conv(re.match(normrep, line), line) sys.stdout.write(line) apply.sh: for name in "$@" do python3 `dirname "$0"`/update.py < "$name" > "$name.tmp" && mv "$name.tmp" "$name" rm -f "$name.tmp" done The actual commands: From llvm/src: find test/ -name *.ll | xargs ./apply.sh From llvm/src/tools/clang: find test/ -name *.mm -o -name *.m -o -name *.cpp -o -name *.c | xargs -I '{}' ../../apply.sh "{}" From llvm/src/tools/polly: find test/ -name *.ll | xargs ./apply.sh After that, check-all (with llvm, clang, clang-tools-extra, lld, compiler-rt, and polly all checked out). The extra 'rm' in the apply.sh script is due to a few files in clang's test suite using interesting unicode stuff that my python script was throwing exceptions on. None of those files needed to be migrated, so it seemed sufficient to ignore those cases. Reviewers: rafael, dexonsmith, grosser Differential Revision: http://reviews.llvm.org/D7636 llvm-svn: 230786
2015-02-28 03:29:02 +08:00
%arrayidx10 = getelementptr inbounds [1024 x double], [1024 x double]* @z, i64 0, i64 %indvars.iv
[opaque pointer type] Add textual IR support for explicit type parameter to load instruction Essentially the same as the GEP change in r230786. A similar migration script can be used to update test cases, though a few more test case improvements/changes were required this time around: (r229269-r229278) import fileinput import sys import re pat = re.compile(r"((?:=|:|^)\s*load (?:atomic )?(?:volatile )?(.*?))(| addrspace\(\d+\) *)\*($| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$)") for line in sys.stdin: sys.stdout.write(re.sub(pat, r"\1, \2\3*\4", line)) Reviewers: rafael, dexonsmith, grosser Differential Revision: http://reviews.llvm.org/D7649 llvm-svn: 230794
2015-02-28 05:17:42 +08:00
%3 = load double, double* %arrayidx10, align 8
[X86] Break false dependencies before partial register updates when the source operand is in memory Adds the various "rm" instruction variants into the list of instructions that have a partial register update. Also adds all variants of SQRTSD that were missing in the original list. Differential Revision: http://reviews.llvm.org/D6620 llvm-svn: 224246
2014-12-15 21:18:21 +08:00
%mul11 = fmul double %mul8, %3
[opaque pointer type] Add textual IR support for explicit type parameter to getelementptr instruction One of several parallel first steps to remove the target type of pointers, replacing them with a single opaque pointer type. This adds an explicit type parameter to the gep instruction so that when the first parameter becomes an opaque pointer type, the type to gep through is still available to the instructions. * This doesn't modify gep operators, only instructions (operators will be handled separately) * Textual IR changes only. Bitcode (including upgrade) and changing the in-memory representation will be in separate changes. * geps of vectors are transformed as: getelementptr <4 x float*> %x, ... ->getelementptr float, <4 x float*> %x, ... Then, once the opaque pointer type is introduced, this will ultimately look like: getelementptr float, <4 x ptr> %x with the unambiguous interpretation that it is a vector of pointers to float. * address spaces remain on the pointer, not the type: getelementptr float addrspace(1)* %x ->getelementptr float, float addrspace(1)* %x Then, eventually: getelementptr float, ptr addrspace(1) %x Importantly, the massive amount of test case churn has been automated by same crappy python code. I had to manually update a few test cases that wouldn't fit the script's model (r228970,r229196,r229197,r229198). The python script just massages stdin and writes the result to stdout, I then wrapped that in a shell script to handle replacing files, then using the usual find+xargs to migrate all the files. update.py: import fileinput import sys import re ibrep = re.compile(r"(^.*?[^%\w]getelementptr inbounds )(((?:<\d* x )?)(.*?)(| addrspace\(\d\)) *\*(|>)(?:$| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$))") normrep = re.compile( r"(^.*?[^%\w]getelementptr )(((?:<\d* x )?)(.*?)(| addrspace\(\d\)) *\*(|>)(?:$| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$))") def conv(match, line): if not match: return line line = match.groups()[0] if len(match.groups()[5]) == 0: line += match.groups()[2] line += match.groups()[3] line += ", " line += match.groups()[1] line += "\n" return line for line in sys.stdin: if line.find("getelementptr ") == line.find("getelementptr inbounds"): if line.find("getelementptr inbounds") != line.find("getelementptr inbounds ("): line = conv(re.match(ibrep, line), line) elif line.find("getelementptr ") != line.find("getelementptr ("): line = conv(re.match(normrep, line), line) sys.stdout.write(line) apply.sh: for name in "$@" do python3 `dirname "$0"`/update.py < "$name" > "$name.tmp" && mv "$name.tmp" "$name" rm -f "$name.tmp" done The actual commands: From llvm/src: find test/ -name *.ll | xargs ./apply.sh From llvm/src/tools/clang: find test/ -name *.mm -o -name *.m -o -name *.cpp -o -name *.c | xargs -I '{}' ../../apply.sh "{}" From llvm/src/tools/polly: find test/ -name *.ll | xargs ./apply.sh After that, check-all (with llvm, clang, clang-tools-extra, lld, compiler-rt, and polly all checked out). The extra 'rm' in the apply.sh script is due to a few files in clang's test suite using interesting unicode stuff that my python script was throwing exceptions on. None of those files needed to be migrated, so it seemed sufficient to ignore those cases. Reviewers: rafael, dexonsmith, grosser Differential Revision: http://reviews.llvm.org/D7636 llvm-svn: 230786
2015-02-28 03:29:02 +08:00
%arrayidx13 = getelementptr inbounds [1024 x double], [1024 x double]* @w, i64 0, i64 %indvars.iv
[X86] Break false dependencies before partial register updates when the source operand is in memory Adds the various "rm" instruction variants into the list of instructions that have a partial register update. Also adds all variants of SQRTSD that were missing in the original list. Differential Revision: http://reviews.llvm.org/D6620 llvm-svn: 224246
2014-12-15 21:18:21 +08:00
store double %mul11, double* %arrayidx13, align 8
%indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
%exitcond = icmp eq i64 %indvars.iv.next, 1024
[X86] Add avx512vl command line to break-false-dep.ll llvm-svn: 336169
2018-07-03 12:43:49 +08:00
tail call void asm sideeffect "", "~{xmm0},~{xmm1},~{xmm2},~{xmm3},~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{dirflag},~{fpsr},~{flags}"()
[X86] Break false dependencies before partial register updates when the source operand is in memory Adds the various "rm" instruction variants into the list of instructions that have a partial register update. Also adds all variants of SQRTSD that were missing in the original list. Differential Revision: http://reviews.llvm.org/D6620 llvm-svn: 224246
2014-12-15 21:18:21 +08:00
br i1 %exitcond, label %for.inc14, label %for.body3
for.inc14: ; preds = %for.body3
%inc15 = add nsw i32 %i.025, 1
%exitcond26 = icmp eq i32 %inc15, 100000
br i1 %exitcond26, label %for.end16, label %for.cond1.preheader
for.end16: ; preds = %for.inc14
ret void
[X86] Add missing memory variants to AVX false dependency breaking Adds missing memory instruction variants to AVX false dependency breaking handling. (SSE was handled in r224246) Differential Revision: http://reviews.llvm.org/D6780 llvm-svn: 224900
2014-12-28 21:15:05 +08:00
;SSE-LABEL:@loopdep3
;SSE: xorps [[XMM0:%xmm[0-9]+]], [[XMM0]]
;SSE-NEXT: cvtsi2sdl {{.*}}, [[XMM0]]
;SSE-NEXT: mulsd {{.*}}, [[XMM0]]
;SSE-NEXT: mulsd {{.*}}, [[XMM0]]
;SSE-NEXT: mulsd {{.*}}, [[XMM0]]
;SSE-NEXT: movsd [[XMM0]],
;AVX-LABEL:@loopdep3
;AVX: vxorps [[XMM0:%xmm[0-9]+]], [[XMM0]]
Avoid false dependencies of undef machine operands This patch helps avoid false dependencies on undef registers by updating the machine instructions' undef operand to use a register that the instruction is truly dependent on, or use a register with clearance higher than Pref. Pseudo example: loop: xmm0 = ... xmm1 = vcvtsi2sdl eax, xmm0<undef> ... = inst xmm0 jmp loop In this example, selecting xmm0 as the undef register creates false dependency between loop iterations. This false dependency cannot be solved by inserting an xor before vcvtsi2sdl because xmm0 is alive at the point of the vcvtsi2sdl instruction. Selecting a different register instead of xmm0, especially a register that is not used in the loop, will eliminate this problem. Differential Revision: https://reviews.llvm.org/D22466 llvm-svn: 278321
2016-08-11 15:32:08 +08:00
;AVX-NEXT: vcvtsi2sdl {{.*}}, [[XMM0]], {{%xmm[0-9]+}}
[X86] Add missing memory variants to AVX false dependency breaking Adds missing memory instruction variants to AVX false dependency breaking handling. (SSE was handled in r224246) Differential Revision: http://reviews.llvm.org/D6780 llvm-svn: 224900
2014-12-28 21:15:05 +08:00
;AVX-NEXT: vmulsd {{.*}}, [[XMM0]], [[XMM0]]
;AVX-NEXT: vmulsd {{.*}}, [[XMM0]], [[XMM0]]
;AVX-NEXT: vmulsd {{.*}}, [[XMM0]], [[XMM0]]
;AVX-NEXT: vmovsd [[XMM0]],
[X86] Break false dependencies before partial register updates when the source operand is in memory Adds the various "rm" instruction variants into the list of instructions that have a partial register update. Also adds all variants of SQRTSD that were missing in the original list. Differential Revision: http://reviews.llvm.org/D6620 llvm-svn: 224246
2014-12-15 21:18:21 +08:00
}
ExecutionDepsFix - Fix bug in clearance calculation The clearance calculation did not take into account registers defined as outputs or clobbers in inline assembly machine instructions because these register defs are implicit. Differential Revision: http://reviews.llvm.org/D22580 llvm-svn: 276266
2016-07-21 20:37:07 +08:00
define double @inlineasmdep(i64 %arg) {
top:
Avoid false dependencies of undef machine operands This patch helps avoid false dependencies on undef registers by updating the machine instructions' undef operand to use a register that the instruction is truly dependent on, or use a register with clearance higher than Pref. Pseudo example: loop: xmm0 = ... xmm1 = vcvtsi2sdl eax, xmm0<undef> ... = inst xmm0 jmp loop In this example, selecting xmm0 as the undef register creates false dependency between loop iterations. This false dependency cannot be solved by inserting an xor before vcvtsi2sdl because xmm0 is alive at the point of the vcvtsi2sdl instruction. Selecting a different register instead of xmm0, especially a register that is not used in the loop, will eliminate this problem. Differential Revision: https://reviews.llvm.org/D22466 llvm-svn: 278321
2016-08-11 15:32:08 +08:00
tail call void asm sideeffect "", "~{xmm0},~{xmm1},~{xmm2},~{xmm3},~{dirflag},~{fpsr},~{flags}"()
tail call void asm sideeffect "", "~{xmm4},~{xmm5},~{xmm6},~{xmm7},~{dirflag},~{fpsr},~{flags}"()
tail call void asm sideeffect "", "~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{dirflag},~{fpsr},~{flags}"()
tail call void asm sideeffect "", "~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{dirflag},~{fpsr},~{flags}"()
[X86] Add avx512vl command line to break-false-dep.ll llvm-svn: 336169
2018-07-03 12:43:49 +08:00
tail call void asm sideeffect "", "~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{dirflag},~{fpsr},~{flags}"()
tail call void asm sideeffect "", "~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{dirflag},~{fpsr},~{flags}"()
tail call void asm sideeffect "", "~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{dirflag},~{fpsr},~{flags}"()
tail call void asm sideeffect "", "~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{dirflag},~{fpsr},~{flags}"()
ExecutionDepsFix - Fix bug in clearance calculation The clearance calculation did not take into account registers defined as outputs or clobbers in inline assembly machine instructions because these register defs are implicit. Differential Revision: http://reviews.llvm.org/D22580 llvm-svn: 276266
2016-07-21 20:37:07 +08:00
%tmp1 = sitofp i64 %arg to double
ret double %tmp1
;AVX-LABEL:@inlineasmdep
;AVX: vxorps [[XMM0:%xmm[0-9]+]], [[XMM0]], [[XMM0]]
[X86] Remove the suffix on vcvt[u]si2ss/sd register variants in assembly printing. We require d/q suffixes on the memory form of these instructions to disambiguate the memory size. We don't require it on the register forms, but need to support parsing both with and without it. Previously we always printed the d/q suffix on the register forms, but it's redundant and inconsistent with gcc and objdump. After this patch we should support the d/q for parsing, but not print it when its unneeded. llvm-svn: 360085
2019-05-07 05:39:51 +08:00
;AVX-NEXT: vcvtsi2sd {{.*}}, [[XMM0]], {{%xmm[0-9]+}}
ExecutionDepsFix - Fix bug in clearance calculation The clearance calculation did not take into account registers defined as outputs or clobbers in inline assembly machine instructions because these register defs are implicit. Differential Revision: http://reviews.llvm.org/D22580 llvm-svn: 276266
2016-07-21 20:37:07 +08:00
}
Avoid false dependencies of undef machine operands This patch helps avoid false dependencies on undef registers by updating the machine instructions' undef operand to use a register that the instruction is truly dependent on, or use a register with clearance higher than Pref. Pseudo example: loop: xmm0 = ... xmm1 = vcvtsi2sdl eax, xmm0<undef> ... = inst xmm0 jmp loop In this example, selecting xmm0 as the undef register creates false dependency between loop iterations. This false dependency cannot be solved by inserting an xor before vcvtsi2sdl because xmm0 is alive at the point of the vcvtsi2sdl instruction. Selecting a different register instead of xmm0, especially a register that is not used in the loop, will eliminate this problem. Differential Revision: https://reviews.llvm.org/D22466 llvm-svn: 278321
2016-08-11 15:32:08 +08:00
; Make sure we are making a smart choice regarding undef registers and
; hiding the false dependency behind a true dependency
define double @truedeps(float %arg) {
top:
tail call void asm sideeffect "", "~{xmm6},~{dirflag},~{fpsr},~{flags}"()
tail call void asm sideeffect "", "~{xmm0},~{xmm1},~{xmm2},~{xmm3},~{dirflag},~{fpsr},~{flags}"()
tail call void asm sideeffect "", "~{xmm4},~{xmm5},~{xmm7},~{dirflag},~{fpsr},~{flags}"()
tail call void asm sideeffect "", "~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{dirflag},~{fpsr},~{flags}"()
tail call void asm sideeffect "", "~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{dirflag},~{fpsr},~{flags}"()
[X86] Add avx512vl command line to break-false-dep.ll llvm-svn: 336169
2018-07-03 12:43:49 +08:00
tail call void asm sideeffect "", "~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{dirflag},~{fpsr},~{flags}"()
tail call void asm sideeffect "", "~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{dirflag},~{fpsr},~{flags}"()
tail call void asm sideeffect "", "~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{dirflag},~{fpsr},~{flags}"()
tail call void asm sideeffect "", "~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{dirflag},~{fpsr},~{flags}"()
Avoid false dependencies of undef machine operands This patch helps avoid false dependencies on undef registers by updating the machine instructions' undef operand to use a register that the instruction is truly dependent on, or use a register with clearance higher than Pref. Pseudo example: loop: xmm0 = ... xmm1 = vcvtsi2sdl eax, xmm0<undef> ... = inst xmm0 jmp loop In this example, selecting xmm0 as the undef register creates false dependency between loop iterations. This false dependency cannot be solved by inserting an xor before vcvtsi2sdl because xmm0 is alive at the point of the vcvtsi2sdl instruction. Selecting a different register instead of xmm0, especially a register that is not used in the loop, will eliminate this problem. Differential Revision: https://reviews.llvm.org/D22466 llvm-svn: 278321
2016-08-11 15:32:08 +08:00
%tmp1 = fpext float %arg to double
ret double %tmp1
;AVX-LABEL:@truedeps
[X86] Prevent folding stack reloads into instructions in hasUndefRegUpdate. An earlier commit prevented folds from the peephole pass by checking for IMPLICIT_DEF. But later in the pipeline IMPLICIT_DEF just becomes and Undef flag on the input register so we need to check for that case too. llvm-svn: 334848
2018-06-16 01:56:17 +08:00
;AVX-NOT: vxorps
;AVX: vcvtss2sd [[XMM0:%xmm[0-9]+]], [[XMM0]], {{%xmm[0-9]+}}
Avoid false dependencies of undef machine operands This patch helps avoid false dependencies on undef registers by updating the machine instructions' undef operand to use a register that the instruction is truly dependent on, or use a register with clearance higher than Pref. Pseudo example: loop: xmm0 = ... xmm1 = vcvtsi2sdl eax, xmm0<undef> ... = inst xmm0 jmp loop In this example, selecting xmm0 as the undef register creates false dependency between loop iterations. This false dependency cannot be solved by inserting an xor before vcvtsi2sdl because xmm0 is alive at the point of the vcvtsi2sdl instruction. Selecting a different register instead of xmm0, especially a register that is not used in the loop, will eliminate this problem. Differential Revision: https://reviews.llvm.org/D22466 llvm-svn: 278321
2016-08-11 15:32:08 +08:00
}
; Make sure we are making a smart choice regarding undef registers and
; choosing the register with the highest clearence
define double @clearence(i64 %arg) {
top:
tail call void asm sideeffect "", "~{xmm6},~{dirflag},~{fpsr},~{flags}"()
tail call void asm sideeffect "", "~{xmm0},~{xmm1},~{xmm2},~{xmm3},~{dirflag},~{fpsr},~{flags}"()
tail call void asm sideeffect "", "~{xmm4},~{xmm5},~{xmm7},~{dirflag},~{fpsr},~{flags}"()
tail call void asm sideeffect "", "~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{dirflag},~{fpsr},~{flags}"()
tail call void asm sideeffect "", "~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{dirflag},~{fpsr},~{flags}"()
[X86] Add avx512vl command line to break-false-dep.ll llvm-svn: 336169
2018-07-03 12:43:49 +08:00
tail call void asm sideeffect "", "~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{dirflag},~{fpsr},~{flags}"()
tail call void asm sideeffect "", "~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{dirflag},~{fpsr},~{flags}"()
tail call void asm sideeffect "", "~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{dirflag},~{fpsr},~{flags}"()
tail call void asm sideeffect "", "~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{dirflag},~{fpsr},~{flags}"()
Avoid false dependencies of undef machine operands This patch helps avoid false dependencies on undef registers by updating the machine instructions' undef operand to use a register that the instruction is truly dependent on, or use a register with clearance higher than Pref. Pseudo example: loop: xmm0 = ... xmm1 = vcvtsi2sdl eax, xmm0<undef> ... = inst xmm0 jmp loop In this example, selecting xmm0 as the undef register creates false dependency between loop iterations. This false dependency cannot be solved by inserting an xor before vcvtsi2sdl because xmm0 is alive at the point of the vcvtsi2sdl instruction. Selecting a different register instead of xmm0, especially a register that is not used in the loop, will eliminate this problem. Differential Revision: https://reviews.llvm.org/D22466 llvm-svn: 278321
2016-08-11 15:32:08 +08:00
%tmp1 = sitofp i64 %arg to double
ret double %tmp1
;AVX-LABEL:@clearence
;AVX: vxorps [[XMM6:%xmm6]], [[XMM6]], [[XMM6]]
[X86] Remove the suffix on vcvt[u]si2ss/sd register variants in assembly printing. We require d/q suffixes on the memory form of these instructions to disambiguate the memory size. We don't require it on the register forms, but need to support parsing both with and without it. Previously we always printed the d/q suffix on the register forms, but it's redundant and inconsistent with gcc and objdump. After this patch we should support the d/q for parsing, but not print it when its unneeded. llvm-svn: 360085
2019-05-07 05:39:51 +08:00
;AVX-NEXT: vcvtsi2sd {{.*}}, [[XMM6]], {{%xmm[0-9]+}}
Avoid false dependencies of undef machine operands This patch helps avoid false dependencies on undef registers by updating the machine instructions' undef operand to use a register that the instruction is truly dependent on, or use a register with clearance higher than Pref. Pseudo example: loop: xmm0 = ... xmm1 = vcvtsi2sdl eax, xmm0<undef> ... = inst xmm0 jmp loop In this example, selecting xmm0 as the undef register creates false dependency between loop iterations. This false dependency cannot be solved by inserting an xor before vcvtsi2sdl because xmm0 is alive at the point of the vcvtsi2sdl instruction. Selecting a different register instead of xmm0, especially a register that is not used in the loop, will eliminate this problem. Differential Revision: https://reviews.llvm.org/D22466 llvm-svn: 278321
2016-08-11 15:32:08 +08:00
}
; Make sure we are making a smart choice regarding undef registers in order to
; avoid a cyclic dependence on a write to the same register in a previous
; iteration, especially when we cannot zero out the undef register because it
; is alive.
define i64 @loopclearence(i64* nocapture %x, double* nocapture %y) nounwind {
entry:
%vx = load i64, i64* %x
br label %loop
loop:
%i = phi i64 [ 1, %entry ], [ %inc, %loop ]
%s1 = phi i64 [ %vx, %entry ], [ %s2, %loop ]
%fi = sitofp i64 %i to double
tail call void asm sideeffect "", "~{xmm0},~{xmm1},~{xmm2},~{xmm3},~{dirflag},~{fpsr},~{flags}"()
tail call void asm sideeffect "", "~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{dirflag},~{fpsr},~{flags}"()
tail call void asm sideeffect "", "~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{dirflag},~{fpsr},~{flags}"()
[X86] Add avx512vl command line to break-false-dep.ll llvm-svn: 336169
2018-07-03 12:43:49 +08:00
tail call void asm sideeffect "", "~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{dirflag},~{fpsr},~{flags}"()
tail call void asm sideeffect "", "~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{dirflag},~{fpsr},~{flags}"()
tail call void asm sideeffect "", "~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{dirflag},~{fpsr},~{flags}"()
tail call void asm sideeffect "", "~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{dirflag},~{fpsr},~{flags}"()
Avoid false dependencies of undef machine operands This patch helps avoid false dependencies on undef registers by updating the machine instructions' undef operand to use a register that the instruction is truly dependent on, or use a register with clearance higher than Pref. Pseudo example: loop: xmm0 = ... xmm1 = vcvtsi2sdl eax, xmm0<undef> ... = inst xmm0 jmp loop In this example, selecting xmm0 as the undef register creates false dependency between loop iterations. This false dependency cannot be solved by inserting an xor before vcvtsi2sdl because xmm0 is alive at the point of the vcvtsi2sdl instruction. Selecting a different register instead of xmm0, especially a register that is not used in the loop, will eliminate this problem. Differential Revision: https://reviews.llvm.org/D22466 llvm-svn: 278321
2016-08-11 15:32:08 +08:00
%vy = load double, double* %y
%fipy = fadd double %fi, %vy
%iipy = fptosi double %fipy to i64
%s2 = add i64 %s1, %iipy
%inc = add nsw i64 %i, 1
%exitcond = icmp eq i64 %inc, 156250000
br i1 %exitcond, label %ret, label %loop
ret:
ret i64 %s2
;AVX-LABEL:@loopclearence
;Registers 4-7 are not used and therefore one of them should be chosen
;AVX-NOT: {{%xmm[4-7]}}
[X86] Remove the suffix on vcvt[u]si2ss/sd register variants in assembly printing. We require d/q suffixes on the memory form of these instructions to disambiguate the memory size. We don't require it on the register forms, but need to support parsing both with and without it. Previously we always printed the d/q suffix on the register forms, but it's redundant and inconsistent with gcc and objdump. After this patch we should support the d/q for parsing, but not print it when its unneeded. llvm-svn: 360085
2019-05-07 05:39:51 +08:00
;AVX: vcvtsi2sd {{.*}}, [[XMM4_7:%xmm[4-7]]], {{%xmm[0-9]+}}
Avoid false dependencies of undef machine operands This patch helps avoid false dependencies on undef registers by updating the machine instructions' undef operand to use a register that the instruction is truly dependent on, or use a register with clearance higher than Pref. Pseudo example: loop: xmm0 = ... xmm1 = vcvtsi2sdl eax, xmm0<undef> ... = inst xmm0 jmp loop In this example, selecting xmm0 as the undef register creates false dependency between loop iterations. This false dependency cannot be solved by inserting an xor before vcvtsi2sdl because xmm0 is alive at the point of the vcvtsi2sdl instruction. Selecting a different register instead of xmm0, especially a register that is not used in the loop, will eliminate this problem. Differential Revision: https://reviews.llvm.org/D22466 llvm-svn: 278321
2016-08-11 15:32:08 +08:00
;AVX-NOT: [[XMM4_7]]
}
[ExecutionDepsFix] Improve clearance calculation for loops Summary: In revision rL278321, ExecutionDepsFix learned how to pick a better register for undef register reads, e.g. for instructions such as `vcvtsi2sdq`. While this revision improved performance on a good number of our benchmarks, it unfortunately also caused significant regressions (up to 3x) on others. This regression turned out to be caused by loops such as: PH -> A -> B (xmm<Undef> -> xmm<Def>) -> C -> D -> EXIT ^ | +----------------------------------+ In the previous version of the clearance calculation, we would visit the blocks in order, remembering for each whether there were any incoming backedges from blocks that we hadn't processed yet and if so queuing up the block to be re-processed. However, for loop structures such as the above, this is clearly insufficient, since the block B does not have any unknown backedges, so we do not see the false dependency from the previous interation's Def of xmm registers in B. To fix this, we need to consider all blocks that are part of the loop and reprocess them one the correct clearance values are known. As an optimization, we also want to avoid reprocessing any later blocks that are not part of the loop. In summary, the iteration order is as follows: Before: PH A B C D A' Corrected (Naive): PH A B C D A' B' C' D' Corrected (w/ optimization): PH A B C A' B' C' D To facilitate this optimization we introduce two new counters for each basic block. The first counts how many of it's predecssors have completed primary processing. The second counts how many of its predecessors have completed all processing (we will call such a block *done*. Now, the criteria to reprocess a block is as follows: - All Predecessors have completed primary processing - For x the number of predecessors that have completed primary processing *at the time of primary processing of this block*, the number of predecessors that are done has reached x. The intuition behind this criterion is as follows: We need to perform primary processing on all predecessors in order to find out any direct defs in those predecessors. When predecessors are done, we also know that we have information about indirect defs (e.g. in block B though that were inherited through B->C->A->B). However, we can't wait for all predecessors to be done, since that would cause cyclic dependencies. However, it is guaranteed that all those predecessors that are prior to us in reverse postorder will be done before us. Since we iterate of the basic blocks in reverse postorder, the number x above, is precisely the count of the number of predecessors prior to us in reverse postorder. Reviewers: myatsina Differential Revision: https://reviews.llvm.org/D28759 llvm-svn: 293571
2017-01-31 07:37:03 +08:00
; Make sure we are making a smart choice regarding undef registers even for more
; complicated loop structures. This example is the inner loop from
; julia> a = falses(10000); a[1:4:end] = true
; julia> linspace(1.0,2.0,10000)[a]
define void @loopclearance2(double* nocapture %y, i64* %x, double %c1, double %c2, double %c3, double %c4, i64 %size) {
entry:
tail call void asm sideeffect "", "~{xmm7},~{dirflag},~{fpsr},~{flags}"()
tail call void asm sideeffect "", "~{xmm8},~{xmm9},~{xmm10},~{xmm11},~{dirflag},~{fpsr},~{flags}"()
tail call void asm sideeffect "", "~{xmm12},~{xmm13},~{xmm14},~{xmm15},~{dirflag},~{fpsr},~{flags}"()
[X86] Add avx512vl command line to break-false-dep.ll llvm-svn: 336169
2018-07-03 12:43:49 +08:00
tail call void asm sideeffect "", "~{xmm16},~{xmm17},~{xmm18},~{xmm19},~{dirflag},~{fpsr},~{flags}"()
tail call void asm sideeffect "", "~{xmm20},~{xmm21},~{xmm22},~{xmm23},~{dirflag},~{fpsr},~{flags}"()
tail call void asm sideeffect "", "~{xmm24},~{xmm25},~{xmm26},~{xmm27},~{dirflag},~{fpsr},~{flags}"()
tail call void asm sideeffect "", "~{xmm28},~{xmm29},~{xmm30},~{xmm31},~{dirflag},~{fpsr},~{flags}"()
[ExecutionDepsFix] Improve clearance calculation for loops Summary: In revision rL278321, ExecutionDepsFix learned how to pick a better register for undef register reads, e.g. for instructions such as `vcvtsi2sdq`. While this revision improved performance on a good number of our benchmarks, it unfortunately also caused significant regressions (up to 3x) on others. This regression turned out to be caused by loops such as: PH -> A -> B (xmm<Undef> -> xmm<Def>) -> C -> D -> EXIT ^ | +----------------------------------+ In the previous version of the clearance calculation, we would visit the blocks in order, remembering for each whether there were any incoming backedges from blocks that we hadn't processed yet and if so queuing up the block to be re-processed. However, for loop structures such as the above, this is clearly insufficient, since the block B does not have any unknown backedges, so we do not see the false dependency from the previous interation's Def of xmm registers in B. To fix this, we need to consider all blocks that are part of the loop and reprocess them one the correct clearance values are known. As an optimization, we also want to avoid reprocessing any later blocks that are not part of the loop. In summary, the iteration order is as follows: Before: PH A B C D A' Corrected (Naive): PH A B C D A' B' C' D' Corrected (w/ optimization): PH A B C A' B' C' D To facilitate this optimization we introduce two new counters for each basic block. The first counts how many of it's predecssors have completed primary processing. The second counts how many of its predecessors have completed all processing (we will call such a block *done*. Now, the criteria to reprocess a block is as follows: - All Predecessors have completed primary processing - For x the number of predecessors that have completed primary processing *at the time of primary processing of this block*, the number of predecessors that are done has reached x. The intuition behind this criterion is as follows: We need to perform primary processing on all predecessors in order to find out any direct defs in those predecessors. When predecessors are done, we also know that we have information about indirect defs (e.g. in block B though that were inherited through B->C->A->B). However, we can't wait for all predecessors to be done, since that would cause cyclic dependencies. However, it is guaranteed that all those predecessors that are prior to us in reverse postorder will be done before us. Since we iterate of the basic blocks in reverse postorder, the number x above, is precisely the count of the number of predecessors prior to us in reverse postorder. Reviewers: myatsina Differential Revision: https://reviews.llvm.org/D28759 llvm-svn: 293571
2017-01-31 07:37:03 +08:00
br label %loop
loop:
%phi_i = phi i64 [ 1, %entry ], [ %nexti, %loop_end ]
%phi_j = phi i64 [ 1, %entry ], [ %nextj, %loop_end ]
%phi_k = phi i64 [ 0, %entry ], [ %nextk, %loop_end ]
br label %inner_loop
inner_loop:
%phi = phi i64 [ %phi_k, %loop ], [ %nextk, %inner_loop ]
%idx = lshr i64 %phi, 6
%inputptr = getelementptr i64, i64* %x, i64 %idx
%input = load i64, i64* %inputptr, align 8
%masked = and i64 %phi, 63
%shiftedmasked = shl i64 1, %masked
%maskedinput = and i64 %input, %shiftedmasked
%cmp = icmp eq i64 %maskedinput, 0
%nextk = add i64 %phi, 1
br i1 %cmp, label %inner_loop, label %loop_end
loop_end:
%nexti = add i64 %phi_i, 1
%nextj = add i64 %phi_j, 1
; Register use, plus us clobbering 7-15 above, basically forces xmm6 here as
; the only reasonable choice. The primary thing we care about is that it's
; not one of the registers used in the loop (e.g. not the output reg here)
;AVX-NOT: %xmm6
[X86] Remove the suffix on vcvt[u]si2ss/sd register variants in assembly printing. We require d/q suffixes on the memory form of these instructions to disambiguate the memory size. We don't require it on the register forms, but need to support parsing both with and without it. Previously we always printed the d/q suffix on the register forms, but it's redundant and inconsistent with gcc and objdump. After this patch we should support the d/q for parsing, but not print it when its unneeded. llvm-svn: 360085
2019-05-07 05:39:51 +08:00
;AVX: vcvtsi2sd {{.*}}, %xmm6, {{%xmm[0-9]+}}
[ExecutionDepsFix] Improve clearance calculation for loops Summary: In revision rL278321, ExecutionDepsFix learned how to pick a better register for undef register reads, e.g. for instructions such as `vcvtsi2sdq`. While this revision improved performance on a good number of our benchmarks, it unfortunately also caused significant regressions (up to 3x) on others. This regression turned out to be caused by loops such as: PH -> A -> B (xmm<Undef> -> xmm<Def>) -> C -> D -> EXIT ^ | +----------------------------------+ In the previous version of the clearance calculation, we would visit the blocks in order, remembering for each whether there were any incoming backedges from blocks that we hadn't processed yet and if so queuing up the block to be re-processed. However, for loop structures such as the above, this is clearly insufficient, since the block B does not have any unknown backedges, so we do not see the false dependency from the previous interation's Def of xmm registers in B. To fix this, we need to consider all blocks that are part of the loop and reprocess them one the correct clearance values are known. As an optimization, we also want to avoid reprocessing any later blocks that are not part of the loop. In summary, the iteration order is as follows: Before: PH A B C D A' Corrected (Naive): PH A B C D A' B' C' D' Corrected (w/ optimization): PH A B C A' B' C' D To facilitate this optimization we introduce two new counters for each basic block. The first counts how many of it's predecssors have completed primary processing. The second counts how many of its predecessors have completed all processing (we will call such a block *done*. Now, the criteria to reprocess a block is as follows: - All Predecessors have completed primary processing - For x the number of predecessors that have completed primary processing *at the time of primary processing of this block*, the number of predecessors that are done has reached x. The intuition behind this criterion is as follows: We need to perform primary processing on all predecessors in order to find out any direct defs in those predecessors. When predecessors are done, we also know that we have information about indirect defs (e.g. in block B though that were inherited through B->C->A->B). However, we can't wait for all predecessors to be done, since that would cause cyclic dependencies. However, it is guaranteed that all those predecessors that are prior to us in reverse postorder will be done before us. Since we iterate of the basic blocks in reverse postorder, the number x above, is precisely the count of the number of predecessors prior to us in reverse postorder. Reviewers: myatsina Differential Revision: https://reviews.llvm.org/D28759 llvm-svn: 293571
2017-01-31 07:37:03 +08:00
;AVX-NOT: %xmm6
%nexti_f = sitofp i64 %nexti to double
%sub = fsub double %c1, %nexti_f
%mul = fmul double %sub, %c2
[X86] Remove the suffix on vcvt[u]si2ss/sd register variants in assembly printing. We require d/q suffixes on the memory form of these instructions to disambiguate the memory size. We don't require it on the register forms, but need to support parsing both with and without it. Previously we always printed the d/q suffix on the register forms, but it's redundant and inconsistent with gcc and objdump. After this patch we should support the d/q for parsing, but not print it when its unneeded. llvm-svn: 360085
2019-05-07 05:39:51 +08:00
;AVX: vcvtsi2sd {{.*}}, %xmm6, {{%xmm[0-9]+}}
[ExecutionDepsFix] Improve clearance calculation for loops Summary: In revision rL278321, ExecutionDepsFix learned how to pick a better register for undef register reads, e.g. for instructions such as `vcvtsi2sdq`. While this revision improved performance on a good number of our benchmarks, it unfortunately also caused significant regressions (up to 3x) on others. This regression turned out to be caused by loops such as: PH -> A -> B (xmm<Undef> -> xmm<Def>) -> C -> D -> EXIT ^ | +----------------------------------+ In the previous version of the clearance calculation, we would visit the blocks in order, remembering for each whether there were any incoming backedges from blocks that we hadn't processed yet and if so queuing up the block to be re-processed. However, for loop structures such as the above, this is clearly insufficient, since the block B does not have any unknown backedges, so we do not see the false dependency from the previous interation's Def of xmm registers in B. To fix this, we need to consider all blocks that are part of the loop and reprocess them one the correct clearance values are known. As an optimization, we also want to avoid reprocessing any later blocks that are not part of the loop. In summary, the iteration order is as follows: Before: PH A B C D A' Corrected (Naive): PH A B C D A' B' C' D' Corrected (w/ optimization): PH A B C A' B' C' D To facilitate this optimization we introduce two new counters for each basic block. The first counts how many of it's predecssors have completed primary processing. The second counts how many of its predecessors have completed all processing (we will call such a block *done*. Now, the criteria to reprocess a block is as follows: - All Predecessors have completed primary processing - For x the number of predecessors that have completed primary processing *at the time of primary processing of this block*, the number of predecessors that are done has reached x. The intuition behind this criterion is as follows: We need to perform primary processing on all predecessors in order to find out any direct defs in those predecessors. When predecessors are done, we also know that we have information about indirect defs (e.g. in block B though that were inherited through B->C->A->B). However, we can't wait for all predecessors to be done, since that would cause cyclic dependencies. However, it is guaranteed that all those predecessors that are prior to us in reverse postorder will be done before us. Since we iterate of the basic blocks in reverse postorder, the number x above, is precisely the count of the number of predecessors prior to us in reverse postorder. Reviewers: myatsina Differential Revision: https://reviews.llvm.org/D28759 llvm-svn: 293571
2017-01-31 07:37:03 +08:00
;AVX-NOT: %xmm6
%phi_f = sitofp i64 %phi to double
%mul2 = fmul double %phi_f, %c3
%add2 = fadd double %mul, %mul2
%div = fdiv double %add2, %c4
%prev_j = add i64 %phi_j, -1
%outptr = getelementptr double, double* %y, i64 %prev_j
store double %div, double* %outptr, align 8
%done = icmp slt i64 %size, %nexti
br i1 %done, label %loopdone, label %loop
loopdone:
ret void
}