Support norm for Diagonal by kshyatt · Pull Request #2860 · JuliaGPU/CUDA.jl

kshyatt · 2025-08-27T14:52:33Z

No description provided.

github-actions · 2025-08-27T15:02:40Z

Your PR requires formatting changes to meet the project's style guidelines.
Please consider running Runic (git runic master) to apply these changes.

Click here to view the suggested changes.

diff --git a/lib/cublas/linalg.jl b/lib/cublas/linalg.jl
index 8ab9670a1..de9ca5a19 100644
--- a/lib/cublas/linalg.jl
+++ b/lib/cublas/linalg.jl
@@ -121,7 +121,7 @@ function LinearAlgebra.norm(x::DenseCuArray{<:Union{Float16, ComplexF16, CublasF
         return invoke(norm, Tuple{AbstractGPUArray, Real}, x, p)
     end
 end
-LinearAlgebra.norm(x::Diagonal{T, <:StridedCuVector{T}}, p::Real=2) where {T<:Union{Float16, ComplexF16, CublasFloat}} = norm(x.diag, p)
+LinearAlgebra.norm(x::Diagonal{T, <:StridedCuVector{T}}, p::Real = 2) where {T <: Union{Float16, ComplexF16, CublasFloat}} = norm(x.diag, p)
 LinearAlgebra.norm2(x::DenseCuArray{<:Union{Float16, ComplexF16, CublasFloat}}) = nrm2(x)
 
 LinearAlgebra.BLAS.asum(x::StridedCuArray{<:CublasFloat}) = asum(length(x), x)

codecov · 2025-08-28T20:42:26Z

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 89.62%. Comparing base (3da87de) to head (0479f19).
⚠️ Report is 1 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #2860      +/-   ##
==========================================
- Coverage   89.77%   89.62%   -0.16%     
==========================================
  Files         150      150              
  Lines       13249    13250       +1     
==========================================
- Hits        11894    11875      -19     
- Misses       1355     1375      +20

☔ View full report in Codecov by Sentry.
📢 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.

github-actions

CUDA.jl Benchmarks

Details

Benchmark suite	Current: `0479f19`	Previous: `3da87de`	Ratio
`latency/precompile`	`43115418381` ns	`43091148296.5` ns	`1.00`
`latency/ttfp`	`7012699873` ns	`7011527319` ns	`1.00`
`latency/import`	`3552835755` ns	`3576055301` ns	`0.99`
`integration/volumerhs`	`9627355` ns	`9615926` ns	`1.00`
`integration/byval/slices=1`	`146856` ns	`147316` ns	`1.00`
`integration/byval/slices=3`	`425893` ns	`426264` ns	`1.00`
`integration/byval/reference`	`145084` ns	`145134` ns	`1.00`
`integration/byval/slices=2`	`286617` ns	`286886` ns	`1.00`
`integration/cudadevrt`	`103525.5` ns	`103713` ns	`1.00`
`kernel/indexing`	`14313` ns	`14307` ns	`1.00`
`kernel/indexing_checked`	`14929` ns	`15001` ns	`1.00`
`kernel/occupancy`	`673.4649681528663` ns	`668.811320754717` ns	`1.01`
`kernel/launch`	`2192.6666666666665` ns	`2183.1111111111113` ns	`1.00`
`kernel/rand`	`14693` ns	`17872` ns	`0.82`
`array/reverse/1d`	`20009` ns	`20122` ns	`0.99`
`array/reverse/2d`	`24618` ns	`25216` ns	`0.98`
`array/reverse/1d_inplace`	`10447` ns	`10612` ns	`0.98`
`array/reverse/2d_inplace`	`11895` ns	`12142` ns	`0.98`
`array/copy`	`21104` ns	`20890` ns	`1.01`
`array/iteration/findall/int`	`158416` ns	`157549.5` ns	`1.01`
`array/iteration/findall/bool`	`140158` ns	`140008` ns	`1.00`
`array/iteration/findfirst/int`	`163128` ns	`161714.5` ns	`1.01`
`array/iteration/findfirst/bool`	`164591` ns	`158865` ns	`1.04`
`array/iteration/scalar`	`73785` ns	`70461` ns	`1.05`
`array/iteration/logical`	`217037` ns	`215555` ns	`1.01`
`array/iteration/findmin/1d`	`47142` ns	`46851` ns	`1.01`
`array/iteration/findmin/2d`	`97085` ns	`96656` ns	`1.00`
`array/reductions/reduce/Int64/1d`	`43124.5` ns	`43153` ns	`1.00`
`array/reductions/reduce/Int64/dims=1`	`47359` ns	`50998.5` ns	`0.93`
`array/reductions/reduce/Int64/dims=2`	`62571` ns	`61632.5` ns	`1.02`
`array/reductions/reduce/Int64/dims=1L`	`89134` ns	`89235.5` ns	`1.00`
`array/reductions/reduce/Int64/dims=2L`	`88047` ns	`87152.5` ns	`1.01`
`array/reductions/reduce/Float32/1d`	`35577` ns	`34743` ns	`1.02`
`array/reductions/reduce/Float32/dims=1`	`52178` ns	`42152.5` ns	`1.24`
`array/reductions/reduce/Float32/dims=2`	`59840` ns	`59901` ns	`1.00`
`array/reductions/reduce/Float32/dims=1L`	`52403` ns	`52601` ns	`1.00`
`array/reductions/reduce/Float32/dims=2L`	`70523.5` ns	`70469` ns	`1.00`
`array/reductions/mapreduce/Int64/1d`	`43530` ns	`42220.5` ns	`1.03`
`array/reductions/mapreduce/Int64/dims=1`	`55572` ns	`47475.5` ns	`1.17`
`array/reductions/mapreduce/Int64/dims=2`	`62620` ns	`61905` ns	`1.01`
`array/reductions/mapreduce/Int64/dims=1L`	`89092.5` ns	`89183` ns	`1.00`
`array/reductions/mapreduce/Int64/dims=2L`	`87357` ns	`87008.5` ns	`1.00`
`array/reductions/mapreduce/Float32/1d`	`35077` ns	`34867` ns	`1.01`
`array/reductions/mapreduce/Float32/dims=1`	`41967` ns	`44301` ns	`0.95`
`array/reductions/mapreduce/Float32/dims=2`	`60364` ns	`60109` ns	`1.00`
`array/reductions/mapreduce/Float32/dims=1L`	`52683` ns	`52926` ns	`1.00`
`array/reductions/mapreduce/Float32/dims=2L`	`70674` ns	`70774` ns	`1.00`
`array/broadcast`	`20174` ns	`20143` ns	`1.00`
`array/copyto!/gpu_to_gpu`	`12808` ns	`11225` ns	`1.14`
`array/copyto!/cpu_to_gpu`	`218810` ns	`217050` ns	`1.01`
`array/copyto!/gpu_to_cpu`	`285490` ns	`284382` ns	`1.00`
`array/accumulate/Int64/1d`	`125173` ns	`124833` ns	`1.00`
`array/accumulate/Int64/dims=1`	`83969` ns	`83599` ns	`1.00`
`array/accumulate/Int64/dims=2`	`158389` ns	`158301` ns	`1.00`
`array/accumulate/Int64/dims=1L`	`1709772` ns	`1720861` ns	`0.99`
`array/accumulate/Int64/dims=2L`	`966803` ns	`968047` ns	`1.00`
`array/accumulate/Float32/1d`	`109630` ns	`109266` ns	`1.00`
`array/accumulate/Float32/dims=1`	`81289` ns	`80566` ns	`1.01`
`array/accumulate/Float32/dims=2`	`148717` ns	`148022` ns	`1.00`
`array/accumulate/Float32/dims=1L`	`1619167` ns	`1618691` ns	`1.00`
`array/accumulate/Float32/dims=2L`	`699103` ns	`701771` ns	`1.00`
`array/construct`	`1297.2` ns	`1218.2` ns	`1.06`
`array/random/randn/Float32`	`45133` ns	`44659` ns	`1.01`
`array/random/randn!/Float32`	`24981` ns	`25047` ns	`1.00`
`array/random/rand!/Int64`	`27605` ns	`27447` ns	`1.01`
`array/random/rand!/Float32`	`8796.666666666666` ns	`8705` ns	`1.01`
`array/random/rand/Int64`	`30031` ns	`38105` ns	`0.79`
`array/random/rand/Float32`	`13393` ns	`13236` ns	`1.01`
`array/permutedims/4d`	`60613` ns	`60763.5` ns	`1.00`
`array/permutedims/2d`	`54395` ns	`54489` ns	`1.00`
`array/permutedims/3d`	`55530` ns	`55278` ns	`1.00`
`array/sorting/1d`	`2758253` ns	`2757119` ns	`1.00`
`array/sorting/by`	`3345319.5` ns	`3355098.5` ns	`1.00`
`array/sorting/2d`	`1081276` ns	`1084658` ns	`1.00`
`cuda/synchronization/stream/auto`	`1049.7` ns	`1012.4166666666666` ns	`1.04`
`cuda/synchronization/stream/nonblocking`	`7811.9` ns	`7297.700000000001` ns	`1.07`
`cuda/synchronization/stream/blocking`	`802.5869565217391` ns	`809.6836734693877` ns	`0.99`
`cuda/synchronization/context/auto`	`1168.8` ns	`1166` ns	`1.00`
`cuda/synchronization/context/nonblocking`	`9161.3` ns	`7817.700000000001` ns	`1.17`
`cuda/synchronization/context/blocking`	`897.4629629629629` ns	`915.4561403508771` ns	`0.98`

This comment was automatically generated by workflow using github-action-benchmark.

kshyatt requested a review from amontoison August 27, 2025 14:52

kshyatt added cuda array Stuff about CuArray. cuda libraries Stuff about CUDA library wrappers. labels Aug 27, 2025

kshyatt force-pushed the ksh/diag_norm branch from 2eba5cb to f40cfe0 Compare August 27, 2025 15:01

amontoison approved these changes Aug 27, 2025

View reviewed changes

kshyatt added 2 commits August 28, 2025 19:09

Support norm for Diagonal

65ce5ab

Fix type sig

0479f19

kshyatt force-pushed the ksh/diag_norm branch from 55bc816 to 0479f19 Compare August 28, 2025 17:09

kshyatt enabled auto-merge (squash) August 28, 2025 17:09

kshyatt merged commit 8ff92f9 into master Aug 28, 2025
3 checks passed

kshyatt deleted the ksh/diag_norm branch August 28, 2025 20:42

github-actions Bot reviewed Aug 28, 2025

View reviewed changes

Navigation Menu

Repository navigation

Support norm for Diagonal - #2860

Support norm for Diagonal#2860
kshyatt merged 2 commits into
masterfrom
ksh/diag_norm

kshyatt commented Aug 27, 2025

Uh oh!

github-actions Bot commented Aug 27, 2025 •

edited

Loading

Uh oh!

Uh oh!

codecov Bot commented Aug 28, 2025

Uh oh!

github-actions Bot left a comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Conversation

kshyatt commented Aug 27, 2025

Uh oh!

github-actions Bot commented Aug 27, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

codecov Bot commented Aug 28, 2025

Codecov Report

Uh oh!

github-actions Bot left a comment

Choose a reason for hiding this comment

CUDA.jl Benchmarks

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

github-actions Bot commented Aug 27, 2025 •

edited

Loading