Fix spdiagm with specified pairs

[ErikQQY]

Fix: #2783

Provide the more general way to define a sparse diagonal matrix with CUDA.

[github-actions]

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/cusparse/array.jl b/lib/cusparse/array.jl
index 4c11e0b5d..3f4df8f8d 100644
--- a/lib/cusparse/array.jl
+++ b/lib/cusparse/array.jl
@@ -336,20 +336,20 @@ function SparseArrays.sparsevec(I::CuArray{Ti}, V::CuArray{Tv}, n::Integer) wher
     CuSparseVector(I, V, n)
 end
 
-SparseArrays.spdiagm(kv::Pair{<:Integer,<:CuVector}...) = _cuda_spdiagm(nothing, kv...)
-SparseArrays.spdiagm(m::Integer, n::Integer, kv::Pair{<:Integer,<:CuVector}...) = _cuda_spdiagm((Int(m),Int(n)), kv...)
+SparseArrays.spdiagm(kv::Pair{<:Integer, <:CuVector}...) = _cuda_spdiagm(nothing, kv...)
+SparseArrays.spdiagm(m::Integer, n::Integer, kv::Pair{<:Integer, <:CuVector}...) = _cuda_spdiagm((Int(m), Int(n)), kv...)
 SparseArrays.spdiagm(v::CuVector) = _cuda_spdiagm(nothing, 0 => v)
 SparseArrays.spdiagm(m::Integer, n::Integer, v::CuVector) = _cuda_spdiagm((Int(m), Int(n)), 0 => v)
 
 function _cuda_spdiagm(size, kv::Pair{<:Integer, <:CuVector}...)
     I, J, V, mmax, nmax = _cuda_spdiagm_internal(kv...)
     mnmax = max(mmax, nmax)
-    m, n = something(size, (mnmax,mnmax))
+    m, n = something(size, (mnmax, mnmax))
     (m ≥ mmax && n ≥ nmax) || throw(DimensionMismatch("invalid size=$size"))
     return sparse(CuVector(I), CuVector(J), V, m, n)
 end
 
-function _cuda_spdiagm_internal(kv::Pair{T,<:CuVector}...) where {T<:Integer}
+function _cuda_spdiagm_internal(kv::Pair{T, <:CuVector}...) where {T <: Integer}
     ncoeffs = 0
     for p in kv
         ncoeffs += SparseArrays._nnz(p.second)
@@ -374,7 +374,7 @@ function _cuda_spdiagm_internal(kv::Pair{T,<:CuVector}...) where {T<:Integer}
             col = 0
         end
         numel = SparseArrays._nnz(v)
-        r = 1+i:numel+i
+        r = (1 + i):(numel + i)
         I_r, J_r = SparseArrays._indices(v, row, col)
         copyto!(view(I, r), I_r)
         copyto!(view(J, r), J_r)

[codecov]

Codecov Report

Attention: Patch coverage is 90.90909% with 4 lines in your changes missing coverage. Please review.

Project coverage is 89.71%. Comparing base (a4a7af4) to head (f3322ad).
Report is 1 commits behind head on master.

Files with missing lines	Patch %	Lines
lib/cusparse/array.jl	90.90%	4 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #2784      +/-   ##
==========================================
- Coverage   89.72%   89.71%   -0.01%     
==========================================
  Files         153      153              
  Lines       13232    13274      +42     
==========================================
+ Hits        11872    11909      +37     
- Misses       1360     1365       +5     

☔ 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

Benchmark suite	Current: 2523b11	Previous: a4a7af4	Ratio
latency/precompile	43078512320.5 ns	43227936166.5 ns	1.00
latency/ttfp	7235008141 ns	7121275224 ns	1.02
latency/import	3450098826 ns	3420695643 ns	1.01
integration/volumerhs	9609573 ns	9610749.5 ns	1.00
integration/byval/slices=1	147170 ns	146856 ns	1.00
integration/byval/slices=3	425740 ns	425444 ns	1.00
integration/byval/reference	145312 ns	145104 ns	1.00
integration/byval/slices=2	286277.5 ns	286260 ns	1.00
integration/cudadevrt	103616 ns	103521 ns	1.00
kernel/indexing	14397.5 ns	14402 ns	1.00
kernel/indexing_checked	15277 ns	14989.5 ns	1.02
kernel/occupancy	755.1029411764706 ns	722.1785714285714 ns	1.05
kernel/launch	2430.6666666666665 ns	2276.3333333333335 ns	1.07
kernel/rand	17888 ns	17130 ns	1.04
array/reverse/1d	20188 ns	20044.5 ns	1.01
array/reverse/2d	25118.5 ns	24062 ns	1.04
array/reverse/1d_inplace	11233 ns	10376 ns	1.08
array/reverse/2d_inplace	13428 ns	11904 ns	1.13
array/copy	21288 ns	21530 ns	0.99
array/iteration/findall/int	160244 ns	159043 ns	1.01
array/iteration/findall/bool	140284 ns	139211 ns	1.01
array/iteration/findfirst/int	164195 ns	160978 ns	1.02
array/iteration/findfirst/bool	165257.5 ns	162266 ns	1.02
array/iteration/scalar	73406 ns	71832 ns	1.02
array/iteration/logical	219938 ns	217950.5 ns	1.01
array/iteration/findmin/1d	46937.5 ns	47357.5 ns	0.99
array/iteration/findmin/2d	97423 ns	97468 ns	1.00
array/reductions/reduce/1d	36616 ns	36864 ns	0.99
array/reductions/reduce/2d	42192.5 ns	41394 ns	1.02
array/reductions/mapreduce/1d	34847 ns	34541 ns	1.01
array/reductions/mapreduce/2d	41664.5 ns	41294 ns	1.01
array/broadcast	21469 ns	21139 ns	1.02
array/copyto!/gpu_to_gpu	12489 ns	12850.5 ns	0.97
array/copyto!/cpu_to_gpu	217160 ns	217756 ns	1.00
array/copyto!/gpu_to_cpu	284338 ns	285934 ns	0.99
array/accumulate/1d	110000 ns	108976.5 ns	1.01
array/accumulate/2d	81141 ns	80598 ns	1.01
array/construct	1262.5 ns	1307.6 ns	0.97
array/random/randn/Float32	44785.5 ns	47928 ns	0.93
array/random/randn!/Float32	25168 ns	25275 ns	1.00
array/random/rand!/Int64	27407 ns	27404 ns	1.00
array/random/rand!/Float32	8859 ns	8877.666666666666 ns	1.00
array/random/rand/Int64	30532 ns	30410 ns	1.00
array/random/rand/Float32	13334 ns	13403 ns	0.99
array/permutedims/4d	61801 ns	61384.5 ns	1.01
array/permutedims/2d	56099.5 ns	55567 ns	1.01
array/permutedims/3d	56622 ns	56298 ns	1.01
array/sorting/1d	2776621.5 ns	2778057 ns	1.00
array/sorting/by	3367591.5 ns	3369635 ns	1.00
array/sorting/2d	1087061.5 ns	1086371 ns	1.00
cuda/synchronization/stream/auto	1062.1 ns	1016.0714285714286 ns	1.05
cuda/synchronization/stream/nonblocking	7424.4 ns	7897.7 ns	0.94
cuda/synchronization/stream/blocking	851.4489795918367 ns	805.7272727272727 ns	1.06
cuda/synchronization/context/auto	1152.2 ns	1176.3 ns	0.98
cuda/synchronization/context/nonblocking	7721.1 ns	7510.1 ns	1.03
cuda/synchronization/context/blocking	914.1041666666666 ns	915.85 ns	1.00

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

[amontoison]

Except the file .DS_Store that should be removed, everything looks correct!
Thanks for the PR.

[ErikQQY]

This PR seems good to be merged.

[ErikQQY]

@amontoison Thanks a lot! Can we cut a release with this patch? I need it in my package testing.

[amontoison]

I need a green light from @maleadt if you want a new release.
You probably also want #2786 for the new release?

[ErikQQY]

Yeah, thanks, a new release with #2786 would be great