Skip to content

Document debugging POCL compilation with POCL_DEBUG#725

Merged
christiangnrd merged 1 commit into
mainfrom
pocl-debug-docs
Jul 17, 2026
Merged

Document debugging POCL compilation with POCL_DEBUG#725
christiangnrd merged 1 commit into
mainfrom
pocl-debug-docs

Conversation

@vchuravy

@vchuravy vchuravy commented Jul 4, 2026

Copy link
Copy Markdown
Member

Adds a docs page (under Extras) on debugging POCL compilation for the CPU backend (POCLBackend):

  • How to enable PoCL's debug output with POCL_DEBUG=err,warn, either in the shell or via ENV["POCL_DEBUG"] before the first kernel launch.
  • Other useful categories (llvm, cache, general, all) with a sample output line.
  • A note on the on-disk kernel cache (~/.cache/pocl/kcache, POCL_KERNEL_CACHE=0, POCL_CACHE_DIR), since cache hits can hide the compilation messages one is looking for.

Verified against the vendored pocl_standalone_jll (debug output confirmed for both ways of setting the variable) and that the docs build renders the page and nav entry.

🤖 Generated with Claude Code

@github-actions

github-actions Bot commented Jul 4, 2026

Copy link
Copy Markdown
Contributor

Benchmark Results

Show table
main f69d924... main / f69d924...
saxpy/default/Float32/1024 29.9 ± 18 μs 0.0318 ± 0.022 ms 0.939 ± 0.86
saxpy/default/Float32/1048576 0.359 ± 0.031 ms 0.343 ± 0.032 ms 1.05 ± 0.13
saxpy/default/Float32/16384 0.0452 ± 0.022 ms 0.0441 ± 0.022 ms 1.03 ± 0.72
saxpy/default/Float32/2048 0.0439 ± 0.021 ms 0.0431 ± 0.021 ms 1.02 ± 0.7
saxpy/default/Float32/256 29.1 ± 15 μs 26.3 ± 8.5 μs 1.11 ± 0.69
saxpy/default/Float32/262144 0.112 ± 0.017 ms 0.11 ± 0.017 ms 1.02 ± 0.22
saxpy/default/Float32/32768 0.0457 ± 0.023 ms 0.0464 ± 0.022 ms 0.986 ± 0.68
saxpy/default/Float32/4096 0.0537 ± 0.025 ms 0.0507 ± 0.023 ms 1.06 ± 0.69
saxpy/default/Float32/512 26.8 ± 10 μs 26.7 ± 11 μs 1 ± 0.58
saxpy/default/Float32/64 26.4 ± 7.5 μs 26.2 ± 7.8 μs 1.01 ± 0.42
saxpy/default/Float32/65536 0.0566 ± 0.024 ms 0.0546 ± 0.022 ms 1.04 ± 0.61
saxpy/default/Float64/1024 0.0371 ± 0.023 ms 0.0343 ± 0.024 ms 1.08 ± 1
saxpy/default/Float64/1048576 0.646 ± 0.057 ms 0.608 ± 0.048 ms 1.06 ± 0.13
saxpy/default/Float64/16384 0.0452 ± 0.023 ms 0.044 ± 0.022 ms 1.03 ± 0.73
saxpy/default/Float64/2048 0.0485 ± 0.024 ms 0.0459 ± 0.023 ms 1.06 ± 0.74
saxpy/default/Float64/256 26.2 ± 9.4 μs 25.4 ± 7.2 μs 1.03 ± 0.47
saxpy/default/Float64/262144 0.183 ± 0.023 ms 0.179 ± 0.023 ms 1.03 ± 0.18
saxpy/default/Float64/32768 0.0521 ± 0.024 ms 0.0511 ± 0.024 ms 1.02 ± 0.67
saxpy/default/Float64/4096 0.0538 ± 0.024 ms 0.0514 ± 0.023 ms 1.05 ± 0.66
saxpy/default/Float64/512 29.2 ± 13 μs 26.2 ± 9.7 μs 1.12 ± 0.65
saxpy/default/Float64/64 28 ± 10 μs 25.9 ± 6.6 μs 1.08 ± 0.48
saxpy/default/Float64/65536 0.0681 ± 0.024 ms 0.0688 ± 0.025 ms 0.99 ± 0.5
saxpy/static workgroup=(1024,)/Float32/1024 0.0333 ± 0.018 ms 0.0327 ± 0.018 ms 1.02 ± 0.78
saxpy/static workgroup=(1024,)/Float32/1048576 0.444 ± 0.075 ms 0.429 ± 0.072 ms 1.04 ± 0.25
saxpy/static workgroup=(1024,)/Float32/16384 0.0483 ± 0.023 ms 0.0459 ± 0.021 ms 1.05 ± 0.69
saxpy/static workgroup=(1024,)/Float32/2048 0.047 ± 0.021 ms 0.0425 ± 0.019 ms 1.11 ± 0.71
saxpy/static workgroup=(1024,)/Float32/256 31.6 ± 12 μs 31.4 ± 13 μs 1.01 ± 0.56
saxpy/static workgroup=(1024,)/Float32/262144 0.147 ± 0.015 ms 0.141 ± 0.015 ms 1.04 ± 0.15
saxpy/static workgroup=(1024,)/Float32/32768 0.0513 ± 0.021 ms 0.0513 ± 0.021 ms 1 ± 0.58
saxpy/static workgroup=(1024,)/Float32/4096 0.0509 ± 0.02 ms 0.0529 ± 0.02 ms 0.962 ± 0.52
saxpy/static workgroup=(1024,)/Float32/512 0.0316 ± 0.013 ms 0.032 ± 0.014 ms 0.989 ± 0.59
saxpy/static workgroup=(1024,)/Float32/64 0.0316 ± 0.011 ms 31.4 ± 9.9 μs 1.01 ± 0.47
saxpy/static workgroup=(1024,)/Float32/65536 0.0661 ± 0.021 ms 0.0643 ± 0.02 ms 1.03 ± 0.46
saxpy/static workgroup=(1024,)/Float64/1024 0.0396 ± 0.022 ms 0.0342 ± 0.02 ms 1.16 ± 0.93
saxpy/static workgroup=(1024,)/Float64/1048576 0.677 ± 0.1 ms 0.668 ± 0.1 ms 1.01 ± 0.22
saxpy/static workgroup=(1024,)/Float64/16384 0.0457 ± 0.021 ms 0.0446 ± 0.02 ms 1.03 ± 0.66
saxpy/static workgroup=(1024,)/Float64/2048 0.0478 ± 0.024 ms 0.0483 ± 0.024 ms 0.989 ± 0.69
saxpy/static workgroup=(1024,)/Float64/256 0.0318 ± 0.011 ms 31.5 ± 11 μs 1.01 ± 0.49
saxpy/static workgroup=(1024,)/Float64/262144 0.192 ± 0.019 ms 0.193 ± 0.017 ms 0.995 ± 0.13
saxpy/static workgroup=(1024,)/Float64/32768 0.0538 ± 0.022 ms 0.0544 ± 0.021 ms 0.989 ± 0.56
saxpy/static workgroup=(1024,)/Float64/4096 0.0513 ± 0.019 ms 0.0527 ± 0.019 ms 0.973 ± 0.5
saxpy/static workgroup=(1024,)/Float64/512 0.0331 ± 0.015 ms 31.2 ± 9.8 μs 1.06 ± 0.59
saxpy/static workgroup=(1024,)/Float64/64 0.032 ± 0.012 ms 0.0328 ± 0.012 ms 0.976 ± 0.51
saxpy/static workgroup=(1024,)/Float64/65536 0.0742 ± 0.02 ms 0.0721 ± 0.02 ms 1.03 ± 0.39
time_to_load 0.976 ± 0.015 s 0.957 ± 0.013 s 1.02 ± 0.021

Benchmark Plots

A plot of the benchmark results have been uploaded as an artifact to the workflow run for this PR.
Go to "Actions"->"Benchmark a pull request"->[the most recent run]->"Artifacts" (at the bottom).

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
@christiangnrd
christiangnrd merged commit 5b47478 into main Jul 17, 2026
27 of 46 checks passed
@christiangnrd
christiangnrd deleted the pocl-debug-docs branch July 17, 2026 21:27
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants