From f3c6c76cfd0dbb29ec444a1e82e2ce034cad606d Mon Sep 17 00:00:00 2001 From: Roman Lebedev Date: Sun, 3 Oct 2021 16:31:23 +0300 Subject: [PATCH] [NFC][X86][LV] Add costmodel test coverage for interleaved i32/f32 load/store stride=3 --- .../CostModel/X86/interleaved-load-f32-stride-3.ll | 75 +++++++++++++++++++++ .../CostModel/X86/interleaved-load-i32-stride-3.ll | 75 +++++++++++++++++++++ .../X86/interleaved-store-f32-stride-3.ll | 76 ++++++++++++++++++++++ .../X86/interleaved-store-i32-stride-3.ll | 76 ++++++++++++++++++++++ 4 files changed, 302 insertions(+) create mode 100644 llvm/test/Analysis/CostModel/X86/interleaved-load-f32-stride-3.ll create mode 100644 llvm/test/Analysis/CostModel/X86/interleaved-load-i32-stride-3.ll create mode 100644 llvm/test/Analysis/CostModel/X86/interleaved-store-f32-stride-3.ll create mode 100644 llvm/test/Analysis/CostModel/X86/interleaved-store-i32-stride-3.ll diff --git a/llvm/test/Analysis/CostModel/X86/interleaved-load-f32-stride-3.ll b/llvm/test/Analysis/CostModel/X86/interleaved-load-f32-stride-3.ll new file mode 100644 index 0000000..946e024 --- /dev/null +++ b/llvm/test/Analysis/CostModel/X86/interleaved-load-f32-stride-3.ll @@ -0,0 +1,75 @@ +; RUN: opt -loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+sse2 --debug-only=loop-vectorize < %s 2>&1 | FileCheck %s --check-prefixes=CHECK,SSE2 +; RUN: opt -loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+avx --debug-only=loop-vectorize < %s 2>&1 | FileCheck %s --check-prefixes=CHECK,AVX1 +; RUN: opt -loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+avx2 --debug-only=loop-vectorize < %s 2>&1 | FileCheck %s --check-prefixes=CHECK,AVX2 +; RUN: opt -loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+avx512bw,+avx512vl --debug-only=loop-vectorize < %s 2>&1 | FileCheck %s --check-prefixes=CHECK,AVX512 +; REQUIRES: asserts + +target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128" +target triple = "x86_64-unknown-linux-gnu" + +@A = global [1024 x float] zeroinitializer, align 128 +@B = global [1024 x i8] zeroinitializer, align 128 + +; CHECK: LV: Checking a loop in "test" +; +; SSE2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v0 = load float, float* %in0, align 4 +; SSE2: LV: Found an estimated cost of 9 for VF 2 For instruction: %v0 = load float, float* %in0, align 4 +; SSE2: LV: Found an estimated cost of 21 for VF 4 For instruction: %v0 = load float, float* %in0, align 4 +; SSE2: LV: Found an estimated cost of 42 for VF 8 For instruction: %v0 = load float, float* %in0, align 4 +; +; AVX1: LV: Found an estimated cost of 1 for VF 1 For instruction: %v0 = load float, float* %in0, align 4 +; AVX1: LV: Found an estimated cost of 12 for VF 2 For instruction: %v0 = load float, float* %in0, align 4 +; AVX1: LV: Found an estimated cost of 24 for VF 4 For instruction: %v0 = load float, float* %in0, align 4 +; AVX1: LV: Found an estimated cost of 57 for VF 8 For instruction: %v0 = load float, float* %in0, align 4 +; AVX1: LV: Found an estimated cost of 114 for VF 16 For instruction: %v0 = load float, float* %in0, align 4 +; +; AVX2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v0 = load float, float* %in0, align 4 +; AVX2: LV: Found an estimated cost of 12 for VF 2 For instruction: %v0 = load float, float* %in0, align 4 +; AVX2: LV: Found an estimated cost of 24 for VF 4 For instruction: %v0 = load float, float* %in0, align 4 +; AVX2: LV: Found an estimated cost of 20 for VF 8 For instruction: %v0 = load float, float* %in0, align 4 +; AVX2: LV: Found an estimated cost of 114 for VF 16 For instruction: %v0 = load float, float* %in0, align 4 +; +; AVX512: LV: Found an estimated cost of 1 for VF 1 For instruction: %v0 = load float, float* %in0, align 4 +; AVX512: LV: Found an estimated cost of 4 for VF 2 For instruction: %v0 = load float, float* %in0, align 4 +; AVX512: LV: Found an estimated cost of 4 for VF 4 For instruction: %v0 = load float, float* %in0, align 4 +; AVX512: LV: Found an estimated cost of 6 for VF 8 For instruction: %v0 = load float, float* %in0, align 4 +; AVX512: LV: Found an estimated cost of 12 for VF 16 For instruction: %v0 = load float, float* %in0, align 4 +; AVX512: LV: Found an estimated cost of 51 for VF 32 For instruction: %v0 = load float, float* %in0, align 4 +; AVX512: LV: Found an estimated cost of 210 for VF 64 For instruction: %v0 = load float, float* %in0, align 4 +; +; CHECK-NOT: LV: Found an estimated cost of {{[0-9]+}} for VF {{[0-9]+}} For instruction: %v0 = load float, float* %in0, align 4 + +define void @test() { +entry: + br label %for.body + +for.body: + %iv = phi i64 [ 0, %entry ], [ %iv.next, %for.body ] + + %iv.0 = add nuw nsw i64 %iv, 0 + %iv.1 = add nuw nsw i64 %iv, 1 + %iv.2 = add nuw nsw i64 %iv, 2 + + %in0 = getelementptr inbounds [1024 x float], [1024 x float]* @A, i64 0, i64 %iv.0 + %in1 = getelementptr inbounds [1024 x float], [1024 x float]* @A, i64 0, i64 %iv.1 + %in2 = getelementptr inbounds [1024 x float], [1024 x float]* @A, i64 0, i64 %iv.2 + + %v0 = load float, float* %in0 + %v1 = load float, float* %in1 + %v2 = load float, float* %in2 + + %reduce.add.0 = fadd float %v0, %v1 + %reduce.add.1 = fadd float %reduce.add.0, %v2 + + %reduce.add.1.narrow = fptoui float %reduce.add.1 to i8 + + %out = getelementptr inbounds [1024 x i8], [1024 x i8]* @B, i64 0, i64 %iv.0 + store i8 %reduce.add.1.narrow, i8* %out + + %iv.next = add nuw nsw i64 %iv.0, 3 + %cmp = icmp ult i64 %iv.next, 1024 + br i1 %cmp, label %for.body, label %for.cond.cleanup + +for.cond.cleanup: + ret void +} diff --git a/llvm/test/Analysis/CostModel/X86/interleaved-load-i32-stride-3.ll b/llvm/test/Analysis/CostModel/X86/interleaved-load-i32-stride-3.ll new file mode 100644 index 0000000..15f81d4 --- /dev/null +++ b/llvm/test/Analysis/CostModel/X86/interleaved-load-i32-stride-3.ll @@ -0,0 +1,75 @@ +; RUN: opt -loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+sse2 --debug-only=loop-vectorize < %s 2>&1 | FileCheck %s --check-prefixes=CHECK,SSE2 +; RUN: opt -loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+avx --debug-only=loop-vectorize < %s 2>&1 | FileCheck %s --check-prefixes=CHECK,AVX1 +; RUN: opt -loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+avx2 --debug-only=loop-vectorize < %s 2>&1 | FileCheck %s --check-prefixes=CHECK,AVX2 +; RUN: opt -loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+avx512bw,+avx512vl --debug-only=loop-vectorize < %s 2>&1 | FileCheck %s --check-prefixes=CHECK,AVX512 +; REQUIRES: asserts + +target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128" +target triple = "x86_64-unknown-linux-gnu" + +@A = global [1024 x i32] zeroinitializer, align 128 +@B = global [1024 x i8] zeroinitializer, align 128 + +; CHECK: LV: Checking a loop in "test" +; +; SSE2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v0 = load i32, i32* %in0, align 4 +; SSE2: LV: Found an estimated cost of 21 for VF 2 For instruction: %v0 = load i32, i32* %in0, align 4 +; SSE2: LV: Found an estimated cost of 45 for VF 4 For instruction: %v0 = load i32, i32* %in0, align 4 +; SSE2: LV: Found an estimated cost of 90 for VF 8 For instruction: %v0 = load i32, i32* %in0, align 4 +; +; AVX1: LV: Found an estimated cost of 1 for VF 1 For instruction: %v0 = load i32, i32* %in0, align 4 +; AVX1: LV: Found an estimated cost of 17 for VF 2 For instruction: %v0 = load i32, i32* %in0, align 4 +; AVX1: LV: Found an estimated cost of 30 for VF 4 For instruction: %v0 = load i32, i32* %in0, align 4 +; AVX1: LV: Found an estimated cost of 69 for VF 8 For instruction: %v0 = load i32, i32* %in0, align 4 +; AVX1: LV: Found an estimated cost of 138 for VF 16 For instruction: %v0 = load i32, i32* %in0, align 4 +; +; AVX2: LV: Found an estimated cost of 1 for VF 1 For instruction: %v0 = load i32, i32* %in0, align 4 +; AVX2: LV: Found an estimated cost of 17 for VF 2 For instruction: %v0 = load i32, i32* %in0, align 4 +; AVX2: LV: Found an estimated cost of 30 for VF 4 For instruction: %v0 = load i32, i32* %in0, align 4 +; AVX2: LV: Found an estimated cost of 20 for VF 8 For instruction: %v0 = load i32, i32* %in0, align 4 +; AVX2: LV: Found an estimated cost of 138 for VF 16 For instruction: %v0 = load i32, i32* %in0, align 4 +; +; AVX512: LV: Found an estimated cost of 1 for VF 1 For instruction: %v0 = load i32, i32* %in0, align 4 +; AVX512: LV: Found an estimated cost of 4 for VF 2 For instruction: %v0 = load i32, i32* %in0, align 4 +; AVX512: LV: Found an estimated cost of 4 for VF 4 For instruction: %v0 = load i32, i32* %in0, align 4 +; AVX512: LV: Found an estimated cost of 6 for VF 8 For instruction: %v0 = load i32, i32* %in0, align 4 +; AVX512: LV: Found an estimated cost of 12 for VF 16 For instruction: %v0 = load i32, i32* %in0, align 4 +; AVX512: LV: Found an estimated cost of 51 for VF 32 For instruction: %v0 = load i32, i32* %in0, align 4 +; AVX512: LV: Found an estimated cost of 210 for VF 64 For instruction: %v0 = load i32, i32* %in0, align 4 +; +; CHECK-NOT: LV: Found an estimated cost of {{[0-9]+}} for VF {{[0-9]+}} For instruction: %v0 = load i32, i32* %in0, align 4 + +define void @test() { +entry: + br label %for.body + +for.body: + %iv = phi i64 [ 0, %entry ], [ %iv.next, %for.body ] + + %iv.0 = add nuw nsw i64 %iv, 0 + %iv.1 = add nuw nsw i64 %iv, 1 + %iv.2 = add nuw nsw i64 %iv, 2 + + %in0 = getelementptr inbounds [1024 x i32], [1024 x i32]* @A, i64 0, i64 %iv.0 + %in1 = getelementptr inbounds [1024 x i32], [1024 x i32]* @A, i64 0, i64 %iv.1 + %in2 = getelementptr inbounds [1024 x i32], [1024 x i32]* @A, i64 0, i64 %iv.2 + + %v0 = load i32, i32* %in0 + %v1 = load i32, i32* %in1 + %v2 = load i32, i32* %in2 + + %reduce.add.0 = add i32 %v0, %v1 + %reduce.add.1 = add i32 %reduce.add.0, %v2 + + %reduce.add.1.narrow = trunc i32 %reduce.add.1 to i8 + + %out = getelementptr inbounds [1024 x i8], [1024 x i8]* @B, i64 0, i64 %iv.0 + store i8 %reduce.add.1.narrow, i8* %out + + %iv.next = add nuw nsw i64 %iv.0, 3 + %cmp = icmp ult i64 %iv.next, 1024 + br i1 %cmp, label %for.body, label %for.cond.cleanup + +for.cond.cleanup: + ret void +} diff --git a/llvm/test/Analysis/CostModel/X86/interleaved-store-f32-stride-3.ll b/llvm/test/Analysis/CostModel/X86/interleaved-store-f32-stride-3.ll new file mode 100644 index 0000000..2b3e9f6 --- /dev/null +++ b/llvm/test/Analysis/CostModel/X86/interleaved-store-f32-stride-3.ll @@ -0,0 +1,76 @@ +; RUN: opt -loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+sse2 --debug-only=loop-vectorize < %s 2>&1 | FileCheck %s --check-prefixes=CHECK,SSE2 +; RUN: opt -loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+avx --debug-only=loop-vectorize < %s 2>&1 | FileCheck %s --check-prefixes=CHECK,AVX1 +; RUN: opt -loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+avx2 --debug-only=loop-vectorize < %s 2>&1 | FileCheck %s --check-prefixes=CHECK,AVX2 +; RUN: opt -loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+avx512bw,+avx512vl --debug-only=loop-vectorize < %s 2>&1 | FileCheck %s --check-prefixes=CHECK,AVX512 +; REQUIRES: asserts + +target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128" +target triple = "x86_64-unknown-linux-gnu" + +@A = global [1024 x i8] zeroinitializer, align 128 +@B = global [1024 x float] zeroinitializer, align 128 + +; CHECK: LV: Checking a loop in "test" +; +; SSE2: LV: Found an estimated cost of 1 for VF 1 For instruction: store float %v2, float* %out2, align 4 +; SSE2: LV: Found an estimated cost of 11 for VF 2 For instruction: store float %v2, float* %out2, align 4 +; SSE2: LV: Found an estimated cost of 24 for VF 4 For instruction: store float %v2, float* %out2, align 4 +; SSE2: LV: Found an estimated cost of 48 for VF 8 For instruction: store float %v2, float* %out2, align 4 +; +; AVX1: LV: Found an estimated cost of 1 for VF 1 For instruction: store float %v2, float* %out2, align 4 +; AVX1: LV: Found an estimated cost of 13 for VF 2 For instruction: store float %v2, float* %out2, align 4 +; AVX1: LV: Found an estimated cost of 23 for VF 4 For instruction: store float %v2, float* %out2, align 4 +; AVX1: LV: Found an estimated cost of 57 for VF 8 For instruction: store float %v2, float* %out2, align 4 +; AVX1: LV: Found an estimated cost of 114 for VF 16 For instruction: store float %v2, float* %out2, align 4 +; +; AVX2: LV: Found an estimated cost of 1 for VF 1 For instruction: store float %v2, float* %out2, align 4 +; AVX2: LV: Found an estimated cost of 13 for VF 2 For instruction: store float %v2, float* %out2, align 4 +; AVX2: LV: Found an estimated cost of 23 for VF 4 For instruction: store float %v2, float* %out2, align 4 +; AVX2: LV: Found an estimated cost of 57 for VF 8 For instruction: store float %v2, float* %out2, align 4 +; AVX2: LV: Found an estimated cost of 114 for VF 16 For instruction: store float %v2, float* %out2, align 4 +; +; AVX512: LV: Found an estimated cost of 1 for VF 1 For instruction: store float %v2, float* %out2, align 4 +; AVX512: LV: Found an estimated cost of 4 for VF 2 For instruction: store float %v2, float* %out2, align 4 +; AVX512: LV: Found an estimated cost of 4 for VF 4 For instruction: store float %v2, float* %out2, align 4 +; AVX512: LV: Found an estimated cost of 8 for VF 8 For instruction: store float %v2, float* %out2, align 4 +; AVX512: LV: Found an estimated cost of 12 for VF 16 For instruction: store float %v2, float* %out2, align 4 +; AVX512: LV: Found an estimated cost of 24 for VF 32 For instruction: store float %v2, float* %out2, align 4 +; AVX512: LV: Found an estimated cost of 48 for VF 64 For instruction: store float %v2, float* %out2, align 4 +; +; CHECK-NOT: LV: Found an estimated cost of {{[0-9]+}} for VF {{[0-9]+}} For instruction: store float %v2, float* %out2, align 4 + +define void @test() { +entry: + br label %for.body + +for.body: + %iv = phi i64 [ 0, %entry ], [ %iv.next, %for.body ] + + %iv.0 = add nuw nsw i64 %iv, 0 + %iv.1 = add nuw nsw i64 %iv, 1 + %iv.2 = add nuw nsw i64 %iv, 2 + + %in = getelementptr inbounds [1024 x i8], [1024 x i8]* @A, i64 0, i64 %iv.0 + %v.narrow = load i8, i8* %in + + %v = uitofp i8 %v.narrow to float + + %v0 = fadd float %v, 0.0 + %v1 = fadd float %v, 1.0 + %v2 = fadd float %v, 2.0 + + %out0 = getelementptr inbounds [1024 x float], [1024 x float]* @B, i64 0, i64 %iv.0 + %out1 = getelementptr inbounds [1024 x float], [1024 x float]* @B, i64 0, i64 %iv.1 + %out2 = getelementptr inbounds [1024 x float], [1024 x float]* @B, i64 0, i64 %iv.2 + + store float %v0, float* %out0 + store float %v1, float* %out1 + store float %v2, float* %out2 + + %iv.next = add nuw nsw i64 %iv.0, 3 + %cmp = icmp ult i64 %iv.next, 1024 + br i1 %cmp, label %for.body, label %for.cond.cleanup + +for.cond.cleanup: + ret void +} diff --git a/llvm/test/Analysis/CostModel/X86/interleaved-store-i32-stride-3.ll b/llvm/test/Analysis/CostModel/X86/interleaved-store-i32-stride-3.ll new file mode 100644 index 0000000..d548e0d --- /dev/null +++ b/llvm/test/Analysis/CostModel/X86/interleaved-store-i32-stride-3.ll @@ -0,0 +1,76 @@ +; RUN: opt -loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+sse2 --debug-only=loop-vectorize < %s 2>&1 | FileCheck %s --check-prefixes=CHECK,SSE2 +; RUN: opt -loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+avx --debug-only=loop-vectorize < %s 2>&1 | FileCheck %s --check-prefixes=CHECK,AVX1 +; RUN: opt -loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+avx2 --debug-only=loop-vectorize < %s 2>&1 | FileCheck %s --check-prefixes=CHECK,AVX2 +; RUN: opt -loop-vectorize -vectorizer-maximize-bandwidth -S -mattr=+avx512bw,+avx512vl --debug-only=loop-vectorize < %s 2>&1 | FileCheck %s --check-prefixes=CHECK,AVX512 +; REQUIRES: asserts + +target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128" +target triple = "x86_64-unknown-linux-gnu" + +@A = global [1024 x i8] zeroinitializer, align 128 +@B = global [1024 x i32] zeroinitializer, align 128 + +; CHECK: LV: Checking a loop in "test" +; +; SSE2: LV: Found an estimated cost of 1 for VF 1 For instruction: store i32 %v2, i32* %out2, align 4 +; SSE2: LV: Found an estimated cost of 23 for VF 2 For instruction: store i32 %v2, i32* %out2, align 4 +; SSE2: LV: Found an estimated cost of 48 for VF 4 For instruction: store i32 %v2, i32* %out2, align 4 +; SSE2: LV: Found an estimated cost of 96 for VF 8 For instruction: store i32 %v2, i32* %out2, align 4 +; +; AVX1: LV: Found an estimated cost of 1 for VF 1 For instruction: store i32 %v2, i32* %out2, align 4 +; AVX1: LV: Found an estimated cost of 18 for VF 2 For instruction: store i32 %v2, i32* %out2, align 4 +; AVX1: LV: Found an estimated cost of 29 for VF 4 For instruction: store i32 %v2, i32* %out2, align 4 +; AVX1: LV: Found an estimated cost of 69 for VF 8 For instruction: store i32 %v2, i32* %out2, align 4 +; AVX1: LV: Found an estimated cost of 138 for VF 16 For instruction: store i32 %v2, i32* %out2, align 4 +; +; AVX2: LV: Found an estimated cost of 1 for VF 1 For instruction: store i32 %v2, i32* %out2, align 4 +; AVX2: LV: Found an estimated cost of 18 for VF 2 For instruction: store i32 %v2, i32* %out2, align 4 +; AVX2: LV: Found an estimated cost of 29 for VF 4 For instruction: store i32 %v2, i32* %out2, align 4 +; AVX2: LV: Found an estimated cost of 69 for VF 8 For instruction: store i32 %v2, i32* %out2, align 4 +; AVX2: LV: Found an estimated cost of 138 for VF 16 For instruction: store i32 %v2, i32* %out2, align 4 +; +; AVX512: LV: Found an estimated cost of 1 for VF 1 For instruction: store i32 %v2, i32* %out2, align 4 +; AVX512: LV: Found an estimated cost of 4 for VF 2 For instruction: store i32 %v2, i32* %out2, align 4 +; AVX512: LV: Found an estimated cost of 4 for VF 4 For instruction: store i32 %v2, i32* %out2, align 4 +; AVX512: LV: Found an estimated cost of 8 for VF 8 For instruction: store i32 %v2, i32* %out2, align 4 +; AVX512: LV: Found an estimated cost of 12 for VF 16 For instruction: store i32 %v2, i32* %out2, align 4 +; AVX512: LV: Found an estimated cost of 24 for VF 32 For instruction: store i32 %v2, i32* %out2, align 4 +; AVX512: LV: Found an estimated cost of 48 for VF 64 For instruction: store i32 %v2, i32* %out2, align 4 +; +; CHECK-NOT: LV: Found an estimated cost of {{[0-9]+}} for VF {{[0-9]+}} For instruction: store i32 %v2, i32* %out2, align 4 + +define void @test() { +entry: + br label %for.body + +for.body: + %iv = phi i64 [ 0, %entry ], [ %iv.next, %for.body ] + + %iv.0 = add nuw nsw i64 %iv, 0 + %iv.1 = add nuw nsw i64 %iv, 1 + %iv.2 = add nuw nsw i64 %iv, 2 + + %in = getelementptr inbounds [1024 x i8], [1024 x i8]* @A, i64 0, i64 %iv.0 + %v.narrow = load i8, i8* %in + + %v = zext i8 %v.narrow to i32 + + %v0 = add i32 %v, 0 + %v1 = add i32 %v, 1 + %v2 = add i32 %v, 2 + + %out0 = getelementptr inbounds [1024 x i32], [1024 x i32]* @B, i64 0, i64 %iv.0 + %out1 = getelementptr inbounds [1024 x i32], [1024 x i32]* @B, i64 0, i64 %iv.1 + %out2 = getelementptr inbounds [1024 x i32], [1024 x i32]* @B, i64 0, i64 %iv.2 + + store i32 %v0, i32* %out0 + store i32 %v1, i32* %out1 + store i32 %v2, i32* %out2 + + %iv.next = add nuw nsw i64 %iv.0, 3 + %cmp = icmp ult i64 %iv.next, 1024 + br i1 %cmp, label %for.body, label %for.cond.cleanup + +for.cond.cleanup: + ret void +} -- 2.7.4