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
88 lines
2.9 KiB
LLVM
88 lines
2.9 KiB
LLVM
; Check that the loop body frequency is maintained when LoopPeel both:
|
|
; - Peels one unconditional iteration after the loop.
|
|
; - Does not add a guard to sometimes skip the remaining loop because it has
|
|
; proven the original loop always executes at least two iterations, which
|
|
; become the initial iteration and the peeled iteration.
|
|
|
|
; DEFINE: %{exitWeight} =
|
|
; DEFINE: %{loopWeight} =
|
|
; DEFINE: %{loopFreqOld} =
|
|
; DEFINE: %{loopFreqNew} =
|
|
|
|
; DEFINE: %{run} = \
|
|
; DEFINE: cp %s %t.ll && chmod +w %t.ll && \
|
|
; DEFINE: echo '!0 = !{!"branch_weights", i32 %{exitWeight}, ' \
|
|
; DEFINE: 'i32 %{loopWeight}}' >> %t.ll && \
|
|
; DEFINE: opt -p "print<block-freq>,loop-unroll,print<block-freq>" \
|
|
; DEFINE: -unroll-full-max-count=0 -S %t.ll 2>&1 | \
|
|
; DEFINE: FileCheck -DLOOP_FREQ_OLD='%{loopFreqOld}' \
|
|
; DEFINE: -DLOOP_FREQ_NEW='%{loopFreqNew}' %s
|
|
|
|
; Branch weights give the original loop 10 iterations. We expect that
|
|
; loopFreqOld = loopFreqNew + 1.
|
|
; REDEFINE: %{exitWeight} = 1
|
|
; REDEFINE: %{loopWeight} = 9
|
|
; REDEFINE: %{loopFreqOld} = 10.0
|
|
; REDEFINE: %{loopFreqNew} = 9.0
|
|
; RUN: %{run}
|
|
|
|
; Branch weights give the original loop 2 iterations. We expect that
|
|
; loopFreqOld = loopFreqNew + 1.
|
|
; REDEFINE: %{exitWeight} = 1
|
|
; REDEFINE: %{loopWeight} = 1
|
|
; REDEFINE: %{loopFreqOld} = 2.0
|
|
; REDEFINE: %{loopFreqNew} = 1.0
|
|
; RUN: %{run}
|
|
|
|
; Branch weights give the original loop 1 iteration, but LoopPeel proved it has
|
|
; at least 2. There is no loop probability that produces a frequency below 1,
|
|
; so the original total frequency cannot be maintained.
|
|
; REDEFINE: %{exitWeight} = 1
|
|
; REDEFINE: %{loopWeight} = 0
|
|
; REDEFINE: %{loopFreqOld} = 1.0
|
|
; REDEFINE: %{loopFreqNew} = 1.0
|
|
; RUN: %{run}
|
|
|
|
; Branch weights say the original loop is infinite, maximizing the frequency,
|
|
; so LoopPeel does not try to decrement it.
|
|
; REDEFINE: %{exitWeight} = 0
|
|
; REDEFINE: %{loopWeight} = 1
|
|
; REDEFINE: %{loopFreqOld} = 2147483647.8
|
|
; REDEFINE: %{loopFreqNew} = 2147483647.8
|
|
; RUN: %{run}
|
|
|
|
; Everything other than loop should be 1.0 because it is reached once.
|
|
;
|
|
; CHECK: block-frequency-info: test
|
|
; CHECK-NEXT: - entry: float = 1.0,
|
|
; CHECK-NEXT: - loop: float = [[LOOP_FREQ_OLD]],
|
|
; CHECK-NEXT: - exit: float = 1.0,
|
|
;
|
|
; CHECK: block-frequency-info: test
|
|
; CHECK-NEXT: - entry: float = 1.0,
|
|
; CHECK-NEXT: - loop: float = [[LOOP_FREQ_NEW]],
|
|
; CHECK-NEXT: - exit.peel.begin: float = 1.0,
|
|
; CHECK-NEXT: - loop.peel: float = 1.0,
|
|
; CHECK-NEXT: - exit.peel.next: float = 1.0,
|
|
; CHECK-NEXT: - loop.peel.next: float = 1.0,
|
|
; CHECK-NEXT: - exit: float = 1.0,
|
|
|
|
declare void @f(i32)
|
|
|
|
define void @test() {
|
|
entry:
|
|
br label %loop
|
|
|
|
loop:
|
|
%i = phi i32 [ 0, %entry ], [ %inc, %loop ]
|
|
%isLast = icmp eq i32 %i, 20
|
|
%sel = select i1 %isLast, i32 1, i32 0
|
|
call void @f(i32 %sel)
|
|
%inc = add i32 %i, 1
|
|
%isLast1 = icmp eq i32 %i, 20
|
|
br i1 %isLast1, label %exit, label %loop, !prof !0
|
|
|
|
exit:
|
|
ret void
|
|
}
|