Update Buildkite pipeline for the new JuliaGPU cluster (release-0.9)

[vchuravy]

Cherry-picks #710 onto the release-0.9-ci branch.

Switches the Buildkite pipeline to the per-backend queues of the new JuliaGPU cluster:

• CUDA → cuda, oneAPI → oneapi, AMDGPU → rocm, Metal → metal
• New OpenCL job → cuda
• Drops the old juliagpu queue + per-backend tag scheme, and removes the now-unused pipeline-julia.yml

The release branch's CUDA Enzyme test is kept active (only its queue switched to cuda); that active/commented divergence from main predates #710 and is left unchanged.

🤖 Generated with Claude Code

[github-actions]

Benchmark Results

	main	e2c6e55...	main / e2c6e55...
saxpy/default/Float32/1024	0.0868 ± 0.0086 ms	0.634 ± 0.0089 μs	137 ± 14
saxpy/default/Float32/1048576	0.465 ± 0.027 ms	0.247 ± 0.014 ms	1.88 ± 0.15
saxpy/default/Float32/16384	0.0697 ± 0.034 ms	2.74 ± 0.15 μs	25.4 ± 13
saxpy/default/Float32/2048	0.0848 ± 0.027 ms	0.754 ± 0.043 μs	112 ± 37
saxpy/default/Float32/256	0.0873 ± 0.012 ms	0.567 ± 0.011 μs	154 ± 21
saxpy/default/Float32/262144	0.17 ± 0.031 ms	0.0633 ± 0.0025 ms	2.69 ± 0.51
saxpy/default/Float32/32768	0.0736 ± 0.031 ms	5.51 ± 0.59 μs	13.4 ± 5.8
saxpy/default/Float32/4096	0.0795 ± 0.03 ms	1.12 ± 0.072 μs	71.2 ± 27
saxpy/default/Float32/512	0.0864 ± 0.0098 ms	0.608 ± 0.011 μs	142 ± 16
saxpy/default/Float32/64	0.0869 ± 0.013 ms	0.557 ± 0.012 μs	156 ± 24
saxpy/default/Float32/65536	0.0899 ± 0.032 ms	13.3 ± 1.1 μs	6.73 ± 2.4
saxpy/default/Float64/1024	0.0873 ± 0.019 ms	0.801 ± 0.036 μs	109 ± 24
saxpy/default/Float64/1048576	0.6 ± 0.099 ms	0.503 ± 0.029 ms	1.19 ± 0.21
saxpy/default/Float64/16384	0.0702 ± 0.03 ms	5.65 ± 1.3 μs	12.4 ± 5.9
saxpy/default/Float64/2048	0.0842 ± 0.031 ms	1.16 ± 0.091 μs	72.5 ± 27
saxpy/default/Float64/256	0.0877 ± 0.011 ms	0.611 ± 0.0064 μs	144 ± 18
saxpy/default/Float64/262144	0.195 ± 0.037 ms	0.13 ± 0.014 ms	1.5 ± 0.33
saxpy/default/Float64/32768	0.0786 ± 0.029 ms	13.8 ± 1.4 μs	5.68 ± 2.1
saxpy/default/Float64/4096	0.0811 ± 0.029 ms	1.79 ± 0.23 μs	45.2 ± 17
saxpy/default/Float64/512	0.0872 ± 0.011 ms	0.664 ± 0.0084 μs	131 ± 16
saxpy/default/Float64/64	0.0876 ± 0.014 ms	0.59 ± 0.0058 μs	149 ± 24
saxpy/default/Float64/65536	0.0975 ± 0.03 ms	0.0329 ± 0.0031 ms	2.96 ± 0.95
saxpy/static workgroup=(1024,)/Float32/1024	0.0839 ± 0.0088 ms	2.08 ± 0.028 μs	40.3 ± 4.3
saxpy/static workgroup=(1024,)/Float32/1048576	0.463 ± 0.025 ms	0.258 ± 0.022 ms	1.79 ± 0.18
saxpy/static workgroup=(1024,)/Float32/16384	0.065 ± 0.032 ms	4.24 ± 0.26 μs	15.3 ± 7.6
saxpy/static workgroup=(1024,)/Float32/2048	0.0819 ± 0.028 ms	2.24 ± 0.075 μs	36.6 ± 12
saxpy/static workgroup=(1024,)/Float32/256	0.0853 ± 0.012 ms	2.65 ± 0.02 μs	32.2 ± 4.7
saxpy/static workgroup=(1024,)/Float32/262144	0.166 ± 0.029 ms	0.0662 ± 0.0024 ms	2.51 ± 0.45
saxpy/static workgroup=(1024,)/Float32/32768	0.0691 ± 0.028 ms	7.39 ± 0.48 μs	9.35 ± 3.9
saxpy/static workgroup=(1024,)/Float32/4096	0.0812 ± 0.03 ms	2.54 ± 0.098 μs	32 ± 12
saxpy/static workgroup=(1024,)/Float32/512	0.0838 ± 0.011 ms	2.76 ± 0.078 μs	30.3 ± 4
saxpy/static workgroup=(1024,)/Float32/64	0.0844 ± 0.013 ms	2.58 ± 0.018 μs	32.7 ± 5.2
saxpy/static workgroup=(1024,)/Float32/65536	0.0831 ± 0.03 ms	16.7 ± 1.7 μs	4.98 ± 1.9
saxpy/static workgroup=(1024,)/Float64/1024	0.0828 ± 0.014 ms	2.27 ± 0.072 μs	36.5 ± 6.4
saxpy/static workgroup=(1024,)/Float64/1048576	0.554 ± 0.078 ms	0.527 ± 0.031 ms	1.05 ± 0.16
saxpy/static workgroup=(1024,)/Float64/16384	0.0666 ± 0.029 ms	7.38 ± 0.49 μs	9.04 ± 4
saxpy/static workgroup=(1024,)/Float64/2048	0.0819 ± 0.03 ms	2.57 ± 0.093 μs	31.8 ± 12
saxpy/static workgroup=(1024,)/Float64/256	0.0838 ± 0.013 ms	2.6 ± 0.023 μs	32.3 ± 4.9
saxpy/static workgroup=(1024,)/Float64/262144	0.193 ± 0.033 ms	0.127 ± 0.012 ms	1.52 ± 0.3
saxpy/static workgroup=(1024,)/Float64/32768	0.0751 ± 0.027 ms	16.2 ± 1.4 μs	4.64 ± 1.7
saxpy/static workgroup=(1024,)/Float64/4096	0.0793 ± 0.029 ms	3.11 ± 0.14 μs	25.5 ± 9.3
saxpy/static workgroup=(1024,)/Float64/512	0.0839 ± 0.0078 ms	2.76 ± 0.036 μs	30.4 ± 2.8
saxpy/static workgroup=(1024,)/Float64/64	0.0844 ± 0.013 ms	2.49 ± 0.02 μs	33.9 ± 5.2
saxpy/static workgroup=(1024,)/Float64/65536	0.0925 ± 0.029 ms	0.0347 ± 0.0023 ms	2.66 ± 0.86
time_to_load	1.05 ± 0.011 s	0.282 ± 0.0013 s	3.74 ± 0.041

Benchmark Plots

A plot of the benchmark results have been uploaded as an artifact to the workflow run for this PR.
Go to "Actions"->"Benchmark a pull request"->[the most recent run]->"Artifacts" (at the bottom).

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 78.80%. Comparing base (f6c1744) to head (e2c6e55).
⚠️ Report is 1 commits behind head on release-0.9.

Additional details and impacted files

@@               Coverage Diff               @@
##           release-0.9     #717      +/-   ##
===============================================
+ Coverage        71.85%   78.80%   +6.95%     
===============================================
  Files               14       14              
  Lines              906      906              
===============================================
+ Hits               651      714      +63     
+ Misses             255      192      -63     

☔ 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.