derek.holloway
7502bf64a4
CodeQL / Analyze (actions) (push) Canceled after 0s
CodeQL / Analyze (python) (push) Canceled after 0s
pre-commit / pre-commit (push) Canceled after 0s
Multi-Arch CI / setup (push) Canceled after 0s
Publish build_manylinux_x86_64 images / publish_build_manylinux_x86_64 (push) Canceled after 0s
Setup (Multi-Arch) / setup (push) Canceled after 0s
Multi-Arch CI / Linux::${{ fromJSON(needs.setup.outputs.linux_build_config || '{}').build_variant_label || 'skip' }} (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish no_rocm_image_ubi10 image / publish_no_rocm_image_ubi10 (push) Canceled after 0s
Multi-Arch CI / Windows::${{ fromJSON(needs.setup.outputs.windows_build_config || '{}').build_variant_label || 'skip' }} (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Multi-Arch CI / Manifest Diff (push) Canceled after 0s
TheRock Manifest Diff Report / Generate Manifest Diff Report (push) Canceled after 0s
Multi-Arch CI / CI Summary (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04 (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04_ocl_rt images / publish_no_rocm_image_ubuntu24_04_ocl_rt (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_rocgdb (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04_rocgdb images / publish_no_rocm_image_ubuntu24_04_rocgdb (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_ocl_rt (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Unit Tests / Unit Tests :: ubuntu-24.04 (push) Canceled after 0s
Unit Tests / Unit Tests :: windows-2022 (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_media (push) Canceled after 0s
Unit Tests / Unit Tests Summary (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_openmpi (push) Canceled after 0s
145 lines
5.8 KiB
LLVM
145 lines
5.8 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 5
|
|
; RUN: opt -passes=loop-interchange -cache-line-size=64 -S < %s | FileCheck %s
|
|
|
|
; Check that nsw/nuw flags are dropped when interchanging loops.
|
|
;
|
|
; int sum = 0;
|
|
; for (int i = 0; i < 2; i++)
|
|
; for (int j = 0; j < 2; j++)
|
|
; sum += A[j][i];
|
|
;
|
|
define void @reduction_add(ptr %A) {
|
|
; CHECK-LABEL: define void @reduction_add(
|
|
; CHECK-SAME: ptr [[A:%.*]]) {
|
|
; CHECK-NEXT: [[ENTRY:.*:]]
|
|
; CHECK-NEXT: br label %[[FOR_J_PREHEADER:.*]]
|
|
; CHECK: [[FOR_I_HEADER_PREHEADER:.*]]:
|
|
; CHECK-NEXT: br label %[[FOR_I_HEADER:.*]]
|
|
; CHECK: [[FOR_I_HEADER]]:
|
|
; CHECK-NEXT: [[I:%.*]] = phi i32 [ [[I_INC:%.*]], %[[FOR_I_LATCH:.*]] ], [ 0, %[[FOR_I_HEADER_PREHEADER]] ]
|
|
; CHECK-NEXT: [[SUM_J:%.*]] = phi i32 [ [[SUM_J_NEXT:%.*]], %[[FOR_I_LATCH]] ], [ [[SUM_I:%.*]], %[[FOR_I_HEADER_PREHEADER]] ]
|
|
; CHECK-NEXT: br label %[[FOR_J_SPLIT1:.*]]
|
|
; CHECK: [[FOR_J_PREHEADER]]:
|
|
; CHECK-NEXT: br label %[[FOR_J:.*]]
|
|
; CHECK: [[FOR_J]]:
|
|
; CHECK-NEXT: [[J:%.*]] = phi i32 [ [[TMP0:%.*]], %[[FOR_J_SPLIT:.*]] ], [ 0, %[[FOR_J_PREHEADER]] ]
|
|
; CHECK-NEXT: [[SUM_I]] = phi i32 [ [[SUM_I_LCSSA:%.*]], %[[FOR_J_SPLIT]] ], [ 0, %[[FOR_J_PREHEADER]] ]
|
|
; CHECK-NEXT: br label %[[FOR_I_HEADER_PREHEADER]]
|
|
; CHECK: [[FOR_J_SPLIT1]]:
|
|
; CHECK-NEXT: [[IDX:%.*]] = getelementptr inbounds [2 x [2 x i32]], ptr [[A]], i32 0, i32 [[J]], i32 [[I]]
|
|
; CHECK-NEXT: [[A:%.*]] = load i32, ptr [[IDX]], align 4
|
|
; CHECK-NEXT: [[SUM_J_NEXT]] = add i32 [[SUM_J]], [[A]]
|
|
; CHECK-NEXT: [[J_INC:%.*]] = add i32 [[J]], 1
|
|
; CHECK-NEXT: [[CMP_J:%.*]] = icmp slt i32 [[J_INC]], 2
|
|
; CHECK-NEXT: br label %[[FOR_I_LATCH]]
|
|
; CHECK: [[FOR_J_SPLIT]]:
|
|
; CHECK-NEXT: [[SUM_I_LCSSA]] = phi i32 [ [[SUM_J_NEXT]], %[[FOR_I_LATCH]] ]
|
|
; CHECK-NEXT: [[TMP0]] = add i32 [[J]], 1
|
|
; CHECK-NEXT: [[TMP1:%.*]] = icmp slt i32 [[TMP0]], 2
|
|
; CHECK-NEXT: br i1 [[TMP1]], label %[[FOR_J]], label %[[EXIT:.*]]
|
|
; CHECK: [[FOR_I_LATCH]]:
|
|
; CHECK-NEXT: [[I_INC]] = add i32 [[I]], 1
|
|
; CHECK-NEXT: [[CMP_I:%.*]] = icmp slt i32 [[I_INC]], 2
|
|
; CHECK-NEXT: br i1 [[CMP_I]], label %[[FOR_I_HEADER]], label %[[FOR_J_SPLIT]]
|
|
; CHECK: [[EXIT]]:
|
|
; CHECK-NEXT: ret void
|
|
;
|
|
entry:
|
|
br label %for.i.header
|
|
|
|
for.i.header:
|
|
%i = phi i32 [ 0, %entry ], [ %i.inc, %for.i.latch ]
|
|
%sum.i = phi i32 [ 0, %entry ], [ %sum.i.lcssa, %for.i.latch ]
|
|
br label %for.j
|
|
|
|
for.j:
|
|
%j = phi i32 [ 0, %for.i.header ], [ %j.inc, %for.j ]
|
|
%sum.j = phi i32 [ %sum.i, %for.i.header ], [ %sum.j.next, %for.j ]
|
|
%idx = getelementptr inbounds [2 x [2 x i32]], ptr %A, i32 0, i32 %j, i32 %i
|
|
%a = load i32, ptr %idx, align 4
|
|
%sum.j.next = add nuw nsw i32 %sum.j, %a
|
|
%j.inc = add i32 %j, 1
|
|
%cmp.j = icmp slt i32 %j.inc, 2
|
|
br i1 %cmp.j, label %for.j, label %for.i.latch
|
|
|
|
for.i.latch:
|
|
%sum.i.lcssa = phi i32 [ %sum.j.next, %for.j ]
|
|
%i.inc = add i32 %i, 1
|
|
%cmp.i = icmp slt i32 %i.inc, 2
|
|
br i1 %cmp.i, label %for.i.header, label %exit
|
|
|
|
exit:
|
|
ret void
|
|
}
|
|
|
|
; Check that nsw/nuw flags are dropped when interchanging loops.
|
|
;
|
|
; int prod = 1;
|
|
; for (int i = 0; i < 2; i++)
|
|
; for (int j = 0; j < 2; j++)
|
|
; prod *= A[j][i];
|
|
;
|
|
define void @reduction_mul(ptr %A) {
|
|
; CHECK-LABEL: define void @reduction_mul(
|
|
; CHECK-SAME: ptr [[A:%.*]]) {
|
|
; CHECK-NEXT: [[ENTRY:.*:]]
|
|
; CHECK-NEXT: br label %[[FOR_J_PREHEADER:.*]]
|
|
; CHECK: [[FOR_I_HEADER_PREHEADER:.*]]:
|
|
; CHECK-NEXT: br label %[[FOR_I_HEADER:.*]]
|
|
; CHECK: [[FOR_I_HEADER]]:
|
|
; CHECK-NEXT: [[I:%.*]] = phi i32 [ [[I_INC:%.*]], %[[FOR_I_LATCH:.*]] ], [ 0, %[[FOR_I_HEADER_PREHEADER]] ]
|
|
; CHECK-NEXT: [[PROD_J:%.*]] = phi i32 [ [[PROD_J_NEXT:%.*]], %[[FOR_I_LATCH]] ], [ [[PROD_I:%.*]], %[[FOR_I_HEADER_PREHEADER]] ]
|
|
; CHECK-NEXT: br label %[[FOR_J_SPLIT1:.*]]
|
|
; CHECK: [[FOR_J_PREHEADER]]:
|
|
; CHECK-NEXT: br label %[[FOR_J:.*]]
|
|
; CHECK: [[FOR_J]]:
|
|
; CHECK-NEXT: [[J:%.*]] = phi i32 [ [[TMP0:%.*]], %[[FOR_J_SPLIT:.*]] ], [ 0, %[[FOR_J_PREHEADER]] ]
|
|
; CHECK-NEXT: [[PROD_I]] = phi i32 [ [[PROD_I_LCSSA:%.*]], %[[FOR_J_SPLIT]] ], [ 1, %[[FOR_J_PREHEADER]] ]
|
|
; CHECK-NEXT: br label %[[FOR_I_HEADER_PREHEADER]]
|
|
; CHECK: [[FOR_J_SPLIT1]]:
|
|
; CHECK-NEXT: [[IDX:%.*]] = getelementptr inbounds [2 x [2 x i32]], ptr [[A]], i32 0, i32 [[J]], i32 [[I]]
|
|
; CHECK-NEXT: [[A:%.*]] = load i32, ptr [[IDX]], align 4
|
|
; CHECK-NEXT: [[PROD_J_NEXT]] = mul i32 [[PROD_J]], [[A]]
|
|
; CHECK-NEXT: [[J_INC:%.*]] = add i32 [[J]], 1
|
|
; CHECK-NEXT: [[CMP_J:%.*]] = icmp slt i32 [[J_INC]], 2
|
|
; CHECK-NEXT: br label %[[FOR_I_LATCH]]
|
|
; CHECK: [[FOR_J_SPLIT]]:
|
|
; CHECK-NEXT: [[PROD_I_LCSSA]] = phi i32 [ [[PROD_J_NEXT]], %[[FOR_I_LATCH]] ]
|
|
; CHECK-NEXT: [[TMP0]] = add i32 [[J]], 1
|
|
; CHECK-NEXT: [[TMP1:%.*]] = icmp slt i32 [[TMP0]], 2
|
|
; CHECK-NEXT: br i1 [[TMP1]], label %[[FOR_J]], label %[[EXIT:.*]]
|
|
; CHECK: [[FOR_I_LATCH]]:
|
|
; CHECK-NEXT: [[I_INC]] = add i32 [[I]], 1
|
|
; CHECK-NEXT: [[CMP_I:%.*]] = icmp slt i32 [[I_INC]], 2
|
|
; CHECK-NEXT: br i1 [[CMP_I]], label %[[FOR_I_HEADER]], label %[[FOR_J_SPLIT]]
|
|
; CHECK: [[EXIT]]:
|
|
; CHECK-NEXT: ret void
|
|
;
|
|
entry:
|
|
br label %for.i.header
|
|
|
|
for.i.header:
|
|
%i = phi i32 [ 0, %entry ], [ %i.inc, %for.i.latch ]
|
|
%prod.i = phi i32 [ 1, %entry ], [ %prod.i.lcssa, %for.i.latch ]
|
|
br label %for.j
|
|
|
|
for.j:
|
|
%j = phi i32 [ 0, %for.i.header ], [ %j.inc, %for.j ]
|
|
%prod.j = phi i32 [ %prod.i, %for.i.header ], [ %prod.j.next, %for.j ]
|
|
%idx = getelementptr inbounds [2 x [2 x i32]], ptr %A, i32 0, i32 %j, i32 %i
|
|
%a = load i32, ptr %idx, align 4
|
|
%prod.j.next = mul nsw nuw i32 %prod.j, %a
|
|
%j.inc = add i32 %j, 1
|
|
%cmp.j = icmp slt i32 %j.inc, 2
|
|
br i1 %cmp.j, label %for.j, label %for.i.latch
|
|
|
|
for.i.latch:
|
|
%prod.i.lcssa = phi i32 [ %prod.j.next, %for.j ]
|
|
%i.inc = add i32 %i, 1
|
|
%cmp.i = icmp slt i32 %i.inc, 2
|
|
br i1 %cmp.i, label %for.i.header, label %exit
|
|
|
|
exit:
|
|
ret void
|
|
}
|