Update gemmStridedBatchedEx! size checks

[lpawela]

This changes the size checks in gemmStridedBatchedEx! to match the ones in gemm_strided_batched!. Now this works

using CUDA

N = 3
nbatch = 4

A = CUDA.rand(N, N, nbatch)
B = CUDA.rand(N, N)
C = CUDA.zeros(N, N, nbatch)
CUDA.CUBLAS.gemmStridedBatchedEx!('N', 'N', 1, A, reshape(B, size(B)..., 1), 0, C)
all(A[:, :, i] * B ≈ C[:, :, i] for i=1:nbatch)

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 89.45%. Comparing base (f7deec6) to head (0591ebd).

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #2935      +/-   ##
==========================================
+ Coverage   88.98%   89.45%   +0.47%     
==========================================
  Files         150      150              
  Lines       13078    13078              
==========================================
+ Hits        11637    11699      +62     
+ Misses       1441     1379      -62     

☔ 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: 0591ebd	Previous: f7deec6	Ratio
latency/precompile	56884307796 ns	56924777734.5 ns	1.00
latency/ttfp	8391774083.5 ns	8417873332.5 ns	1.00
latency/import	4506835308 ns	4531361015 ns	0.99
integration/volumerhs	9610638 ns	9625377 ns	1.00
integration/byval/slices=1	146968 ns	146827 ns	1.00
integration/byval/slices=3	426103 ns	425931 ns	1.00
integration/byval/reference	145020 ns	144949 ns	1.00
integration/byval/slices=2	286452 ns	286317 ns	1.00
integration/cudadevrt	103608 ns	103600 ns	1.00
kernel/indexing	14410 ns	14225 ns	1.01
kernel/indexing_checked	15051 ns	15087 ns	1.00
kernel/occupancy	716.9236111111111 ns	679.8954248366013 ns	1.05
kernel/launch	2149.4444444444443 ns	2150.3333333333335 ns	1.00
kernel/rand	15291 ns	14810 ns	1.03
array/reverse/1d	20117.5 ns	20182 ns	1.00
array/reverse/2dL_inplace	66853 ns	66832.5 ns	1.00
array/reverse/1dL	70355 ns	70358 ns	1.00
array/reverse/2d	22980 ns	21865 ns	1.05
array/reverse/1d_inplace	11496 ns	11480 ns	1.00
array/reverse/2d_inplace	13376 ns	13272 ns	1.01
array/reverse/2dL	74966 ns	73906 ns	1.01
array/reverse/1dL_inplace	66815 ns	66817 ns	1.00
array/copy	21276 ns	20949 ns	1.02
array/iteration/findall/int	158488.5 ns	157295 ns	1.01
array/iteration/findall/bool	140199 ns	139923.5 ns	1.00
array/iteration/findfirst/int	162221 ns	161193 ns	1.01
array/iteration/findfirst/bool	162961 ns	162272 ns	1.00
array/iteration/scalar	73293 ns	73738 ns	0.99
array/iteration/logical	218788 ns	214452.5 ns	1.02
array/iteration/findmin/1d	53667 ns	50889.5 ns	1.05
array/iteration/findmin/2d	97313 ns	96643 ns	1.01
array/reductions/reduce/Int64/1d	44330.5 ns	43989 ns	1.01
array/reductions/reduce/Int64/dims=1	45198 ns	44879 ns	1.01
array/reductions/reduce/Int64/dims=2	62058.5 ns	61825 ns	1.00
array/reductions/reduce/Int64/dims=1L	89454 ns	89232 ns	1.00
array/reductions/reduce/Int64/dims=2L	88685 ns	88384 ns	1.00
array/reductions/reduce/Float32/1d	38994 ns	37163 ns	1.05
array/reductions/reduce/Float32/dims=1	43425.5 ns	47666 ns	0.91
array/reductions/reduce/Float32/dims=2	60596 ns	59848 ns	1.01
array/reductions/reduce/Float32/dims=1L	53007 ns	52408 ns	1.01
array/reductions/reduce/Float32/dims=2L	73475 ns	72122.5 ns	1.02
array/reductions/mapreduce/Int64/1d	44846 ns	43666 ns	1.03
array/reductions/mapreduce/Int64/dims=1	48412.5 ns	47028 ns	1.03
array/reductions/mapreduce/Int64/dims=2	62230 ns	61661 ns	1.01
array/reductions/mapreduce/Int64/dims=1L	89379 ns	88863 ns	1.01
array/reductions/mapreduce/Int64/dims=2L	89365 ns	88192 ns	1.01
array/reductions/mapreduce/Float32/1d	38872.5 ns	37065 ns	1.05
array/reductions/mapreduce/Float32/dims=1	46520 ns	42446.5 ns	1.10
array/reductions/mapreduce/Float32/dims=2	60533 ns	60229 ns	1.01
array/reductions/mapreduce/Float32/dims=1L	53350 ns	52761 ns	1.01
array/reductions/mapreduce/Float32/dims=2L	73581 ns	72561 ns	1.01
array/broadcast	20586 ns	20011 ns	1.03
array/copyto!/gpu_to_gpu	13319 ns	11355.5 ns	1.17
array/copyto!/cpu_to_gpu	216698 ns	216192 ns	1.00
array/copyto!/gpu_to_cpu	284398 ns	283975.5 ns	1.00
array/accumulate/Int64/1d	125218 ns	125034.5 ns	1.00
array/accumulate/Int64/dims=1	84241 ns	83398 ns	1.01
array/accumulate/Int64/dims=2	158674 ns	157817 ns	1.01
array/accumulate/Int64/dims=1L	1709855.5 ns	1708490 ns	1.00
array/accumulate/Int64/dims=2L	966521 ns	966251 ns	1.00
array/accumulate/Float32/1d	109885 ns	109114 ns	1.01
array/accumulate/Float32/dims=1	80855 ns	80351 ns	1.01
array/accumulate/Float32/dims=2	148692 ns	147295.5 ns	1.01
array/accumulate/Float32/dims=1L	1618915 ns	1618020.5 ns	1.00
array/accumulate/Float32/dims=2L	698676.5 ns	698067 ns	1.00
array/construct	1313.4 ns	1296.7 ns	1.01
array/random/randn/Float32	45969.5 ns	48838.5 ns	0.94
array/random/randn!/Float32	25242 ns	24912 ns	1.01
array/random/rand!/Int64	27490 ns	27275 ns	1.01
array/random/rand!/Float32	8972.333333333334 ns	8805.333333333334 ns	1.02
array/random/rand/Int64	30129 ns	30044 ns	1.00
array/random/rand/Float32	13306 ns	13354 ns	1.00
array/permutedims/4d	60499 ns	60446 ns	1.00
array/permutedims/2d	54612 ns	54105.5 ns	1.01
array/permutedims/3d	55720 ns	54893 ns	1.02
array/sorting/1d	2759021 ns	2756483.5 ns	1.00
array/sorting/by	3345765 ns	3368977 ns	0.99
array/sorting/2d	1082960 ns	1088064.5 ns	1.00
cuda/synchronization/stream/auto	1015.8 ns	1030.1 ns	0.99
cuda/synchronization/stream/nonblocking	7615 ns	7504.6 ns	1.01
cuda/synchronization/stream/blocking	810.054347826087 ns	801.2842105263157 ns	1.01
cuda/synchronization/context/auto	1178.7 ns	1179.3 ns	1.00
cuda/synchronization/context/nonblocking	7358.2 ns	7293.5 ns	1.01
cuda/synchronization/context/blocking	909.88 ns	909.9636363636364 ns	1.00

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