[christiangnrd]

Does this need tests or are the GPUArrays tests (that don't run here) sufficient?

[github-actions]

Benchmark results

Compared with 7d2ea0f46, the merge base.
The environment differs from the baseline's: JSON 1.13.0 → 1.13.4, Parsers 3.0.3 → 3.0.5, SPIRVIntrinsics 1.5.1 → 1.6.0, StructUtils 2.11.2 → 2.12.0, StructUtilsStaticArraysCoreExt 2.11.2 → 2.12.0.

Times are the minimum time of every benchmark. A change is unusual when it exceeds both three times the benchmark's typical change between consecutive runs on the default branch and a minimum per category (performance 3%, latency 5%, applications 3%, other 3%). A benchmark that starts or stops allocating is listed however few the allocations. A single run can still be a fluke, so rerun before drawing conclusions.

	Geometric mean	Unusual: slower	Unusual: faster	Compared
Performance	+0.1%	0	0	32
Latency	+0.1%	0	0	13
Applications	+1.6%	0	0	5

No unusual changes.

All benchmarks

Benchmark	Baseline	This PR	Change
Performance
julia/serial/arithmetic/muladd-reference	655 µs	655 µs	−0.0%	noise unknown
julia/serial/math/exp-f64-reference	290 µs	284 µs	−1.9%	noise unknown
julia/serial/narrow/masked-reference	711 µs	700 µs	−1.5%	noise unknown
pocl/parallel/atomics/histogram	5.16 ms	5.15 ms	−0.2%	noise unknown
pocl/parallel/groupreduce/add-f64	1.66 ms	1.65 ms	−0.3%	noise unknown
pocl/parallel/groupreduce/fallback	18.1 ms	18.1 ms	−0.0%	noise unknown
pocl/parallel/groupscan/add-i32	2.53 ms	2.52 ms	−0.5%	noise unknown
pocl/parallel/indexing/cartesian3d	77.8 µs	80.4 µs	+3.4%	noise unknown
pocl/parallel/indexing/linear3d	78.4 µs	81.3 µs	+3.7%	noise unknown
pocl/parallel/indexing/ntuple3d	79.5 µs	77.4 µs	−2.6%	noise unknown
pocl/parallel/indexing/ntuple4d	165 µs	163 µs	−1.2%	noise unknown
pocl/parallel/localmem/transpose-naive	457 µs	460 µs	+0.6%	noise unknown
pocl/parallel/localmem/transpose-tiled	104 µs	105 µs	+1.1%	noise unknown
pocl/parallel/private/across-barrier	165 µs	161 µs	−2.4%	noise unknown
pocl/parallel/scheduling/default	182 µs	181 µs	−0.2%	noise unknown
pocl/parallel/scheduling/explicit	181 µs	181 µs	−0.3%	noise unknown
pocl/parallel/stream/copy/cache	784 ns	684 ns	−12.8%	noise unknown
pocl/parallel/stream/copy/dram	6.25 ms	6.52 ms	+4.2%	noise unknown
pocl/parallel/stream/triad/cache	7.54 µs	7.32 µs	−2.9%	noise unknown
pocl/parallel/stream/triad/dram	8.73 ms	9.35 ms	+7.2%	noise unknown
pocl/parallel/util/copyto-small	24.0 ns	26.0 ns	+8.3%	noise unknown
pocl/parallel/util/zeros-large	999 µs	1.02 ms	+1.7%	noise unknown
pocl/parallel/util/zeros-small	632 ns	629 ns	−0.5%	noise unknown
pocl/serial/arithmetic/contract	1.91 ms	1.91 ms	−0.2%	noise unknown
pocl/serial/arithmetic/muladd	765 µs	761 µs	−0.5%	noise unknown
pocl/serial/const/marked	727 µs	718 µs	−1.2%	noise unknown
pocl/serial/const/unmarked	914 µs	923 µs	+1.0%	noise unknown
pocl/serial/math/exp-f64	146 µs	147 µs	+0.6%	noise unknown
pocl/serial/math/sin-f64	191 µs	195 µs	+2.1%	noise unknown
pocl/serial/narrow/masked	815 µs	831 µs	+2.0%	noise unknown
pocl/serial/narrow/unmasked	788 µs	777 µs	−1.5%	noise unknown
pocl/serial/stream/triad/dram	8.61 ms	8.66 ms	+0.6%	noise unknown
Latency
host/import/ka	509 ms	502 ms	−1.3%	noise unknown
host/precompile/ka	44.8 s	44.7 s	−0.3%	noise unknown
host/startup	131 ms	131 ms	−0.1%	noise unknown
pocl/parallel/launch/dynamic-default	4.36 µs	4.34 µs	−0.3%	noise unknown
pocl/parallel/launch/groupreduce-f64	6.59 µs	6.60 µs	+0.2%	noise unknown
pocl/parallel/launch/many-args	12.1 µs	12.1 µs	−0.1%	noise unknown
pocl/parallel/launch/partial-workgroup	4.25 µs	4.26 µs	+0.4%	noise unknown
pocl/serial/compile/groupreduce	1.86 s	1.89 s	+1.4%	noise unknown
pocl/serial/compile/localmem-barrier	718 ms	710 ms	−1.0%	noise unknown
pocl/serial/compile/new-kernel	109 ms	110 ms	+0.4%	noise unknown
pocl/serial/compile/new-workgroupsize	13.6 ms	13.8 ms	+1.4%	noise unknown
pocl/serial/first-kernel/cold-cache	2.05 s	2.06 s	+0.5%	noise unknown
pocl/serial/first-kernel/warm-cache	386 ms	385 ms	−0.4%	noise unknown
Applications
pocl/parallel/dgsem/fluxdiff-node	295 µs	291 µs	−1.2%	noise unknown
pocl/parallel/dgsem/weak-element	160 µs	159 µs	−0.5%	noise unknown
pocl/parallel/dgsem/weak-node-localmem	119 µs	117 µs	−1.4%	noise unknown
pocl/parallel/stencil/timestep	7.65 ms	8.75 ms	+14.3%	noise unknown
pocl/serial/stencil/timestep	11.5 ms	11.2 ms	−2.2%	noise unknown

History: https://juliagpu.org/perf/?pkg=kernelabstractions

[codecov]

Codecov Report

❌ Patch coverage is 94.44444% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 83.85%. Comparing base (7d2ea0f) to head (81aca7d).

Files with missing lines	Patch %	Lines
src/pocl/pocl.jl	90.00%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #943      +/-   ##
==========================================
+ Coverage   83.17%   83.85%   +0.68%     
==========================================
  Files          31       31              
  Lines        2567     2533      -34     
==========================================
- Hits         2135     2124      -11     
+ Misses        432      409      -23     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.