Split linalg tests into multiple files

[kshyatt]

Testing locally showed that we don't have undefined variable errors. More could likely be done here to decouple testsets within each level but this does allow us to run the two most expensive groups (level 2 and level 3) in parallel.

[kshyatt]

Sorry for the mega-diff btw.

[github-actions]

CUDA.jl Benchmarks

Benchmark suite	Current: eba40e5	Previous: a0c2f4b	Ratio
latency/precompile	45274142645.5 ns	45297385295 ns	1.00
latency/ttfp	6426012668 ns	6375596178 ns	1.01
latency/import	3053958391 ns	3036561495 ns	1.01
integration/volumerhs	9572279 ns	9567419 ns	1.00
integration/byval/slices=1	146444 ns	146746 ns	1.00
integration/byval/slices=3	425081 ns	425517.5 ns	1.00
integration/byval/reference	144620 ns	145010 ns	1.00
integration/byval/slices=2	285740 ns	286216 ns	1.00
integration/cudadevrt	103371 ns	103513 ns	1.00
kernel/indexing	14007 ns	14419 ns	0.97
kernel/indexing_checked	15224 ns	15499 ns	0.98
kernel/occupancy	696.9527027027027 ns	748.2734375 ns	0.93
kernel/launch	2103.4 ns	2194.6666666666665 ns	0.96
kernel/rand	14733 ns	17335 ns	0.85
array/reverse/1d	19399 ns	19412 ns	1.00
array/reverse/2d	24822 ns	24576 ns	1.01
array/reverse/1d_inplace	10033.333333333334 ns	11029 ns	0.91
array/reverse/2d_inplace	10790 ns	13223 ns	0.82
array/copy	20061 ns	20740 ns	0.97
array/iteration/findall/int	157486 ns	158179 ns	1.00
array/iteration/findall/bool	138323 ns	138583 ns	1.00
array/iteration/findfirst/int	152810 ns	153423 ns	1.00
array/iteration/findfirst/bool	154073.5 ns	154821 ns	1.00
array/iteration/scalar	76401 ns	77451 ns	0.99
array/iteration/logical	211056 ns	216735 ns	0.97
array/iteration/findmin/1d	40626 ns	41556.5 ns	0.98
array/iteration/findmin/2d	93579.5 ns	94128 ns	0.99
array/reductions/reduce/1d	38254 ns	42013 ns	0.91
array/reductions/reduce/2d	46399.5 ns	51911 ns	0.89
array/reductions/mapreduce/1d	36615 ns	39275 ns	0.93
array/reductions/mapreduce/2d	48354.5 ns	49505.5 ns	0.98
array/broadcast	21058 ns	21668 ns	0.97
array/copyto!/gpu_to_gpu	11424 ns	11569 ns	0.99
array/copyto!/cpu_to_gpu	209189 ns	211873 ns	0.99
array/copyto!/gpu_to_cpu	243384 ns	245423 ns	0.99
array/accumulate/1d	108525 ns	108388.5 ns	1.00
array/accumulate/2d	79667.5 ns	79823 ns	1.00
array/construct	1203.7 ns	1208.35 ns	1.00
array/random/randn/Float32	42407 ns	43873.5 ns	0.97
array/random/randn!/Float32	26229 ns	25937 ns	1.01
array/random/rand!/Int64	26884 ns	27271 ns	0.99
array/random/rand!/Float32	8450.666666666666 ns	8766.666666666666 ns	0.96
array/random/rand/Int64	29654 ns	29637 ns	1.00
array/random/rand/Float32	12550 ns	12723 ns	0.99
array/permutedims/4d	66640 ns	66923 ns	1.00
array/permutedims/2d	56190 ns	56439 ns	1.00
array/permutedims/3d	58685 ns	58867 ns	1.00
array/sorting/1d	2919790 ns	2933352 ns	1.00
array/sorting/by	3482848 ns	3500830 ns	0.99
array/sorting/2d	1079724 ns	1085059 ns	1.00
cuda/synchronization/stream/auto	1029.4 ns	1038.4 ns	0.99
cuda/synchronization/stream/nonblocking	6570.8 ns	6432 ns	1.02
cuda/synchronization/stream/blocking	812.8061224489796 ns	807.5918367346939 ns	1.01
cuda/synchronization/context/auto	1172.6 ns	1194.1 ns	0.98
cuda/synchronization/context/nonblocking	6811.6 ns	6649.8 ns	1.02
cuda/synchronization/context/blocking	888.8913043478261 ns	886.6415094339623 ns	1.00

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

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 73.59%. Comparing base (792aec5) to head (7f8cc3f).

Additional details and impacted files

@@             Coverage Diff             @@
##           master    #2609       +/-   ##
===========================================
+ Coverage    9.27%   73.59%   +64.31%     
===========================================
  Files         157      157               
  Lines       15025    15207      +182     
===========================================
+ Hits         1394    11191     +9797     
+ Misses      13631     4016     -9615     

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

[maleadt]

Moving them into a subdirectory.

---

Before:

Test                    (Worker) | Time (s) | GC (s) | GC % | Alloc (MB) | RSS (MB) | GC (s) | GC % | Alloc (MB) | RSS (MB) |
libraries/cublas  (2) |   183.98 |   0.02 |  0.0 |      43.14 |   612.00 |   3.16 |  1.7 |   22655.27 |  3237.44 |

After:

Test                    (Worker) | Time (s) | GC (s) | GC % | Alloc (MB) | RSS (MB) | GC (s) | GC % | Alloc (MB) | RSS (MB) |
libraries/cublas/level1      (5) |    31.63 |   0.00 |  0.0 |       0.03 |   512.00 |   0.56 |  1.8 |    3639.77 |  1714.91 |
libraries/cublas/level2      (3) |    61.55 |   0.01 |  0.0 |       1.06 |   588.00 |   0.93 |  1.5 |    5854.03 |  1758.51 |
libraries/cublas/level3      (2) |   116.81 |   0.01 |  0.0 |      14.20 |   534.00 |   2.11 |  1.8 |   14023.70 |  2564.25 |
libraries/cublas/extensions  (4) |    66.44 |   0.01 |  0.0 |      29.49 |   586.00 |   1.04 |  1.6 |    6527.81 |  1714.91 |