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
78 lines
2.5 KiB
LLVM
78 lines
2.5 KiB
LLVM
; RUN: llc < %s -mtriple=armv7a-eabi | FileCheck %s -check-prefix=CHECK -check-prefix=CHECK-AT2
|
|
; RUN: llc < %s -mtriple=thumbv7m-eabi | FileCheck %s -check-prefix=CHECK -check-prefix=CHECK-AT2
|
|
; RUN: llc < %s -mtriple=thumbv6m-eabi | FileCheck %s -check-prefix=CHECK -check-prefix=CHECK-T1
|
|
|
|
; This test checks that various kinds of getelementptr are all optimised to a
|
|
; simple multiply plus add, with the add being done by a register offset if the
|
|
; result is used in a load.
|
|
|
|
; CHECK-LABEL: calc_1d:
|
|
; CHECK: mov{{s?}} [[REG1:r[0-9]+]], #84
|
|
; CHECK-AT2: mla r0, r1, [[REG1]], r0
|
|
; CHECK-T1: muls [[REG2:r[0-9]+]], r1, [[REG1]]
|
|
; CHECK-T1: adds r0, r0, [[REG2]]
|
|
define ptr @calc_1d(ptr %p, i32 %n) {
|
|
entry:
|
|
%mul = mul nsw i32 %n, 21
|
|
%add.ptr = getelementptr inbounds i32, ptr %p, i32 %mul
|
|
ret ptr %add.ptr
|
|
}
|
|
|
|
; CHECK-LABEL: load_1d:
|
|
; CHECK: mov{{s?}} [[REG1:r[0-9]+]], #84
|
|
; CHECK: mul{{s?}} [[REG2:r[0-9]+]],{{( r1,)?}} [[REG1]]{{(, r1)?}}
|
|
; CHECK: ldr r0, [r0, [[REG2]]]
|
|
define i32 @load_1d(ptr %p, i32 %n) #1 {
|
|
entry:
|
|
%mul = mul nsw i32 %n, 21
|
|
%arrayidx = getelementptr inbounds i32, ptr %p, i32 %mul
|
|
%0 = load i32, ptr %arrayidx, align 4
|
|
ret i32 %0
|
|
}
|
|
|
|
; CHECK-LABEL: calc_2d_a:
|
|
; CHECK: mov{{s?}} [[REG1:r[0-9]+]], #84
|
|
; CHECK-AT2: mla r0, r1, [[REG1]], r0
|
|
; CHECK-T1: muls [[REG2:r[0-9]+]], r1, [[REG1]]
|
|
; CHECK-T1: adds r0, r0, [[REG2]]
|
|
define ptr @calc_2d_a(ptr %p, i32 %n) {
|
|
entry:
|
|
%mul = mul nsw i32 %n, 21
|
|
%arrayidx1 = getelementptr inbounds [100 x i32], ptr %p, i32 0, i32 %mul
|
|
ret ptr %arrayidx1
|
|
}
|
|
|
|
; CHECK-LABEL: load_2d_a:
|
|
; CHECK: mov{{s?}} [[REG1:r[0-9]+]], #84
|
|
; CHECK: mul{{s?}} [[REG2:r[0-9]+]],{{( r1,)?}} [[REG1]]{{(, r1)?}}
|
|
; CHECK: ldr r0, [r0, [[REG2]]]
|
|
define i32 @load_2d_a(ptr %p, i32 %n) #1 {
|
|
entry:
|
|
%mul = mul nsw i32 %n, 21
|
|
%arrayidx1 = getelementptr inbounds [100 x i32], ptr %p, i32 0, i32 %mul
|
|
%0 = load i32, ptr %arrayidx1, align 4
|
|
ret i32 %0
|
|
}
|
|
|
|
; CHECK-LABEL: calc_2d_b:
|
|
; CHECK: mov{{s?}} [[REG1:r[0-9]+]], #84
|
|
; CHECK-AT2: mla r0, r1, [[REG1]], r0
|
|
; CHECK-T1: muls [[REG2:r[0-9]+]], r1, [[REG1]]
|
|
; CHECK-T1: adds r0, r0, [[REG2]]
|
|
define ptr @calc_2d_b(ptr %p, i32 %n) {
|
|
entry:
|
|
%arrayidx1 = getelementptr inbounds [21 x i32], ptr %p, i32 %n, i32 0
|
|
ret ptr %arrayidx1
|
|
}
|
|
|
|
; CHECK-LABEL: load_2d_b:
|
|
; CHECK: mov{{s?}} [[REG1:r[0-9]+]], #84
|
|
; CHECK: mul{{s?}} [[REG2:r[0-9]+]],{{( r1,)?}} [[REG1]]{{(, r1)?}}
|
|
; CHECK: ldr r0, [r0, [[REG2]]]
|
|
define i32 @load_2d_b(ptr %p, i32 %n) {
|
|
entry:
|
|
%arrayidx1 = getelementptr inbounds [21 x i32], ptr %p, i32 %n, i32 0
|
|
%0 = load i32, ptr %arrayidx1, align 4
|
|
ret i32 %0
|
|
}
|