Initial support for vectorization using svml (short vector math library).

Differential Revision: https://reviews.llvm.org/D19544

llvm-svn: 277166
This commit is contained in:
Matt Masten 2016-07-29 16:42:44 +00:00
parent 9e69450ec7
commit a6669a1e05
3 changed files with 259 additions and 3 deletions

View File

@ -85,8 +85,9 @@ public:
/// addVectorizableFunctionsFromVecLib for filling up the tables of
/// vectorizable functions.
enum VectorLibrary {
NoLibrary, // Don't use any vector library.
Accelerate // Use Accelerate framework.
NoLibrary, // Don't use any vector library.
Accelerate, // Use Accelerate framework.
SVML // Intel short vector math library.
};
TargetLibraryInfoImpl();

View File

@ -23,9 +23,10 @@ static cl::opt<TargetLibraryInfoImpl::VectorLibrary> ClVectorLibrary(
"No vector functions library"),
clEnumValN(TargetLibraryInfoImpl::Accelerate, "Accelerate",
"Accelerate framework"),
clEnumValN(TargetLibraryInfoImpl::SVML, "SVML",
"Intel SVML library"),
clEnumValEnd));
const char *const TargetLibraryInfoImpl::StandardNames[LibFunc::NumLibFuncs] = {
#define TLI_DEFINE_STRING
#include "llvm/Analysis/TargetLibraryInfo.def"
@ -1075,6 +1076,75 @@ void TargetLibraryInfoImpl::addVectorizableFunctionsFromVecLib(
addVectorizableFunctions(VecFuncs);
break;
}
case SVML: {
const VecDesc VecFuncs[] = {
{"sin", "__svml_sin2", 2},
{"sin", "__svml_sin4", 4},
{"sin", "__svml_sin8", 8},
{"sinf", "__svml_sinf4", 4},
{"sinf", "__svml_sinf8", 8},
{"sinf", "__svml_sinf16", 16},
{"cos", "__svml_cos2", 2},
{"cos", "__svml_cos4", 4},
{"cos", "__svml_cos8", 8},
{"cosf", "__svml_cosf4", 4},
{"cosf", "__svml_cosf8", 8},
{"cosf", "__svml_cosf16", 16},
{"pow", "__svml_pow2", 2},
{"pow", "__svml_pow4", 4},
{"pow", "__svml_pow8", 8},
{"powf", "__svml_powf4", 4},
{"powf", "__svml_powf8", 8},
{"powf", "__svml_powf16", 16},
{"llvm.pow.f64", "__svml_pow2", 2},
{"llvm.pow.f64", "__svml_pow4", 4},
{"llvm.pow.f64", "__svml_pow8", 8},
{"llvm.pow.f32", "__svml_powf4", 4},
{"llvm.pow.f32", "__svml_powf8", 8},
{"llvm.pow.f32", "__svml_powf16", 16},
{"exp", "__svml_exp2", 2},
{"exp", "__svml_exp4", 4},
{"exp", "__svml_exp8", 8},
{"expf", "__svml_expf4", 4},
{"expf", "__svml_expf8", 8},
{"expf", "__svml_expf16", 16},
{"llvm.exp.f64", "__svml_exp2", 2},
{"llvm.exp.f64", "__svml_exp4", 4},
{"llvm.exp.f64", "__svml_exp8", 8},
{"llvm.exp.f32", "__svml_expf4", 4},
{"llvm.exp.f32", "__svml_expf8", 8},
{"llvm.exp.f32", "__svml_expf16", 16},
{"log", "__svml_log2", 2},
{"log", "__svml_log4", 4},
{"log", "__svml_log8", 8},
{"logf", "__svml_logf4", 4},
{"logf", "__svml_logf8", 8},
{"logf", "__svml_logf16", 16},
{"llvm.log.f64", "__svml_log2", 2},
{"llvm.log.f64", "__svml_log4", 4},
{"llvm.log.f64", "__svml_log8", 8},
{"llvm.log.f32", "__svml_logf4", 4},
{"llvm.log.f32", "__svml_logf8", 8},
{"llvm.log.f32", "__svml_logf16", 16},
};
addVectorizableFunctions(VecFuncs);
break;
}
case NoLibrary:
break;
}

View File

@ -0,0 +1,185 @@
; RUN: opt -vector-library=SVML -loop-vectorize -S < %s | FileCheck %s
target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
target triple = "x86_64-unknown-linux-gnu"
; CHECK-LABEL: @sin_f32
; CHECK: <4 x float> @__svml_sinf4
; CHECK: ret
declare float @sinf(float) #0
define void @sin_f32(float* nocapture %varray) {
entry:
br label %for.body
for.body: ; preds = %for.body, %entry
%indvars.iv = phi i64 [ 0, %entry ], [ %indvars.iv.next, %for.body ]
%tmp = trunc i64 %indvars.iv to i32
%conv = sitofp i32 %tmp to float
%call = tail call fast float @sinf(float %conv)
%arrayidx = getelementptr inbounds float, float* %varray, i64 %indvars.iv
store float %call, float* %arrayidx, align 4
%indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
%exitcond = icmp eq i64 %indvars.iv.next, 1000
br i1 %exitcond, label %for.end, label %for.body
for.end: ; preds = %for.body
ret void
}
; CHECK-LABEL: @cos_f32
; CHECK: <4 x float> @__svml_cosf4
; CHECK: ret
declare float @cosf(float) #0
define void @cos_f32(float* nocapture %varray) {
entry:
br label %for.body
for.body: ; preds = %for.body, %entry
%indvars.iv = phi i64 [ 0, %entry ], [ %indvars.iv.next, %for.body ]
%tmp = trunc i64 %indvars.iv to i32
%conv = sitofp i32 %tmp to float
%call = tail call fast float @cosf(float %conv)
%arrayidx = getelementptr inbounds float, float* %varray, i64 %indvars.iv
store float %call, float* %arrayidx, align 4
%indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
%exitcond = icmp eq i64 %indvars.iv.next, 1000
br i1 %exitcond, label %for.end, label %for.body
for.end: ; preds = %for.body
ret void
}
; CHECK-LABEL: @exp_f32
; CHECK: <4 x float> @__svml_expf4
; CHECK: ret
declare float @expf(float) #0
define void @exp_f32(float* nocapture %varray) {
entry:
br label %for.body
for.body: ; preds = %for.body, %entry
%indvars.iv = phi i64 [ 0, %entry ], [ %indvars.iv.next, %for.body ]
%tmp = trunc i64 %indvars.iv to i32
%conv = sitofp i32 %tmp to float
%call = tail call fast float @expf(float %conv)
%arrayidx = getelementptr inbounds float, float* %varray, i64 %indvars.iv
store float %call, float* %arrayidx, align 4
%indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
%exitcond = icmp eq i64 %indvars.iv.next, 1000
br i1 %exitcond, label %for.end, label %for.body
for.end: ; preds = %for.body
ret void
}
; CHECK-LABEL: @exp_f32_intrin
; CHECK: <4 x float> @__svml_expf4
; CHECK: ret
declare float @llvm.exp.f32(float) #0
define void @exp_f32_intrin(float* nocapture %varray) {
entry:
br label %for.body
for.body: ; preds = %for.body, %entry
%indvars.iv = phi i64 [ 0, %entry ], [ %indvars.iv.next, %for.body ]
%tmp = trunc i64 %indvars.iv to i32
%conv = sitofp i32 %tmp to float
%call = tail call fast float @llvm.exp.f32(float %conv)
%arrayidx = getelementptr inbounds float, float* %varray, i64 %indvars.iv
store float %call, float* %arrayidx, align 4
%indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
%exitcond = icmp eq i64 %indvars.iv.next, 1000
br i1 %exitcond, label %for.end, label %for.body
for.end: ; preds = %for.body
ret void
}
; CHECK-LABEL: @log_f32
; CHECK: <4 x float> @__svml_logf4
; CHECK: ret
declare float @logf(float) #0
define void @log_f32(float* nocapture %varray) {
entry:
br label %for.body
for.body: ; preds = %for.body, %entry
%indvars.iv = phi i64 [ 0, %entry ], [ %indvars.iv.next, %for.body ]
%tmp = trunc i64 %indvars.iv to i32
%conv = sitofp i32 %tmp to float
%call = tail call fast float @logf(float %conv)
%arrayidx = getelementptr inbounds float, float* %varray, i64 %indvars.iv
store float %call, float* %arrayidx, align 4
%indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
%exitcond = icmp eq i64 %indvars.iv.next, 1000
br i1 %exitcond, label %for.end, label %for.body
for.end: ; preds = %for.body
ret void
}
; CHECK-LABEL: @pow_f32
; CHECK: <4 x float> @__svml_powf4
; CHECK: ret
declare float @powf(float, float) #0
define void @pow_f32(float* nocapture %varray, float* nocapture readonly %exp) {
entry:
br label %for.body
for.body: ; preds = %for.body, %entry
%indvars.iv = phi i64 [ 0, %entry ], [ %indvars.iv.next, %for.body ]
%tmp = trunc i64 %indvars.iv to i32
%conv = sitofp i32 %tmp to float
%arrayidx = getelementptr inbounds float, float* %exp, i64 %indvars.iv
%tmp1 = load float, float* %arrayidx, align 4
%tmp2 = tail call fast float @powf(float %conv, float %tmp1)
%arrayidx2 = getelementptr inbounds float, float* %varray, i64 %indvars.iv
store float %tmp2, float* %arrayidx2, align 4
%indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
%exitcond = icmp eq i64 %indvars.iv.next, 1000
br i1 %exitcond, label %for.end, label %for.body
for.end: ; preds = %for.body
ret void
}
; CHECK-LABEL: @pow_f32_intrin
; CHECK: <4 x float> @__svml_powf4
; CHECK: ret
declare float @llvm.pow.f32(float, float) #0
define void @pow_f32_intrin(float* nocapture %varray, float* nocapture readonly %exp) {
entry:
br label %for.body
for.body: ; preds = %for.body, %entry
%indvars.iv = phi i64 [ 0, %entry ], [ %indvars.iv.next, %for.body ]
%tmp = trunc i64 %indvars.iv to i32
%conv = sitofp i32 %tmp to float
%arrayidx = getelementptr inbounds float, float* %exp, i64 %indvars.iv
%tmp1 = load float, float* %arrayidx, align 4
%tmp2 = tail call fast float @llvm.pow.f32(float %conv, float %tmp1)
%arrayidx2 = getelementptr inbounds float, float* %varray, i64 %indvars.iv
store float %tmp2, float* %arrayidx2, align 4
%indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
%exitcond = icmp eq i64 %indvars.iv.next, 1000
br i1 %exitcond, label %for.end, label %for.body
for.end: ; preds = %for.body
ret void
}
attributes #0 = { nounwind readnone }