Restore Enzyme to CI checks

[wsmoses]

now that @vchuravy fixed the GPUCompiler compat

[github-actions]

CUDA.jl Benchmarks

Benchmark suite	Current: 2afe6c6	Previous: bb0a27f	Ratio
latency/precompile	56628328674.5 ns	56531370914 ns	1.00
latency/ttfp	8465367383.5 ns	8348294090 ns	1.01
latency/import	4530861281 ns	4485476728 ns	1.01
integration/volumerhs	9606804.5 ns	9611273 ns	1.00
integration/byval/slices=1	146932 ns	146807 ns	1.00
integration/byval/slices=3	426049 ns	426057.5 ns	1.00
integration/byval/reference	145114 ns	145125 ns	1.00
integration/byval/slices=2	286406 ns	286538 ns	1.00
integration/cudadevrt	103457 ns	103588 ns	1.00
kernel/indexing	14186 ns	14206 ns	1.00
kernel/indexing_checked	15006.5 ns	14985 ns	1.00
kernel/occupancy	673.343949044586 ns	669.9113924050633 ns	1.01
kernel/launch	2168.6666666666665 ns	2201.5555555555557 ns	0.99
kernel/rand	17115 ns	15084 ns	1.13
array/reverse/1d	20177.5 ns	20085 ns	1.00
array/reverse/2dL_inplace	66812 ns	66806.5 ns	1.00
array/reverse/1dL	70343 ns	70229 ns	1.00
array/reverse/2d	21840 ns	21647 ns	1.01
array/reverse/1d_inplace	11477 ns	9989.5 ns	1.15
array/reverse/2d_inplace	13255 ns	13425 ns	0.99
array/reverse/2dL	73906 ns	73806 ns	1.00
array/reverse/1dL_inplace	66901 ns	66788 ns	1.00
array/copy	20827 ns	21013 ns	0.99
array/iteration/findall/int	156962 ns	158131 ns	0.99
array/iteration/findall/bool	139393.5 ns	139780 ns	1.00
array/iteration/findfirst/int	161049 ns	161255 ns	1.00
array/iteration/findfirst/bool	161777.5 ns	162292 ns	1.00
array/iteration/scalar	71086 ns	74249 ns	0.96
array/iteration/logical	214952 ns	215568.5 ns	1.00
array/iteration/findmin/1d	50399 ns	50621 ns	1.00
array/iteration/findmin/2d	96313 ns	96593 ns	1.00
array/reductions/reduce/Int64/1d	43812 ns	43717 ns	1.00
array/reductions/reduce/Int64/dims=1	45159.5 ns	47315 ns	0.95
array/reductions/reduce/Int64/dims=2	61368 ns	61512 ns	1.00
array/reductions/reduce/Int64/dims=1L	89073 ns	89114 ns	1.00
array/reductions/reduce/Int64/dims=2L	87814 ns	88127 ns	1.00
array/reductions/reduce/Float32/1d	36495 ns	38052 ns	0.96
array/reductions/reduce/Float32/dims=1	42038 ns	42212.5 ns	1.00
array/reductions/reduce/Float32/dims=2	59995 ns	59909 ns	1.00
array/reductions/reduce/Float32/dims=1L	52466 ns	52382 ns	1.00
array/reductions/reduce/Float32/dims=2L	72317 ns	72498 ns	1.00
array/reductions/mapreduce/Int64/1d	43193 ns	43848 ns	0.99
array/reductions/mapreduce/Int64/dims=1	47375 ns	55194.5 ns	0.86
array/reductions/mapreduce/Int64/dims=2	61631 ns	61712 ns	1.00
array/reductions/mapreduce/Int64/dims=1L	89051 ns	88933 ns	1.00
array/reductions/mapreduce/Int64/dims=2L	88042 ns	88319 ns	1.00
array/reductions/mapreduce/Float32/1d	37264.5 ns	37106 ns	1.00
array/reductions/mapreduce/Float32/dims=1	52288.5 ns	41792 ns	1.25
array/reductions/mapreduce/Float32/dims=2	60482 ns	60056 ns	1.01
array/reductions/mapreduce/Float32/dims=1L	53010 ns	52707 ns	1.01
array/reductions/mapreduce/Float32/dims=2L	72741 ns	72500 ns	1.00
array/broadcast	20175 ns	20185 ns	1.00
array/copyto!/gpu_to_gpu	11462 ns	13288 ns	0.86
array/copyto!/cpu_to_gpu	214887 ns	216328 ns	0.99
array/copyto!/gpu_to_cpu	283937 ns	284828 ns	1.00
array/accumulate/Int64/1d	124699 ns	124343 ns	1.00
array/accumulate/Int64/dims=1	83475 ns	83433 ns	1.00
array/accumulate/Int64/dims=2	158043.5 ns	157726 ns	1.00
array/accumulate/Int64/dims=1L	1708810 ns	1710225.5 ns	1.00
array/accumulate/Int64/dims=2L	966321.5 ns	966516 ns	1.00
array/accumulate/Float32/1d	109353 ns	109351 ns	1.00
array/accumulate/Float32/dims=1	80459 ns	80425 ns	1.00
array/accumulate/Float32/dims=2	147423.5 ns	147561 ns	1.00
array/accumulate/Float32/dims=1L	1618073 ns	1619125 ns	1.00
array/accumulate/Float32/dims=2L	697852.5 ns	698195 ns	1.00
array/construct	1300.7 ns	1266.3 ns	1.03
array/random/randn/Float32	47962.5 ns	45130 ns	1.06
array/random/randn!/Float32	25010 ns	25086 ns	1.00
array/random/rand!/Int64	27304 ns	27356 ns	1.00
array/random/rand!/Float32	8783 ns	9035 ns	0.97
array/random/rand/Int64	29896 ns	29946 ns	1.00
array/random/rand/Float32	13150 ns	13310 ns	0.99
array/permutedims/4d	59830 ns	59701 ns	1.00
array/permutedims/2d	54018 ns	53857.5 ns	1.00
array/permutedims/3d	54666 ns	54922.5 ns	1.00
array/sorting/1d	2756855 ns	2758155 ns	1.00
array/sorting/by	3343979.5 ns	3344572 ns	1.00
array/sorting/2d	1080937.5 ns	1081253.5 ns	1.00
cuda/synchronization/stream/auto	1021.9 ns	1040.7 ns	0.98
cuda/synchronization/stream/nonblocking	7116.9 ns	7682 ns	0.93
cuda/synchronization/stream/blocking	814.989247311828 ns	807.7010309278351 ns	1.01
cuda/synchronization/context/auto	1161.5 ns	1187.8 ns	0.98
cuda/synchronization/context/nonblocking	8780.9 ns	8672.3 ns	1.01
cuda/synchronization/context/blocking	895.1458333333334 ns	915.4772727272727 ns	0.98

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

[maleadt]

Enzyme CI fails.

[wsmoses]

@vchuravy looks like your fix missed the tape_type function?

extensions/enzyme: Error During Test at /var/lib/buildkite-agent/builds/gpuci-3/julialang/cuda-dot-jl/test/extensions/enzyme.jl:51
--
  | Got exception outside of a @test
  | UndefVarError: `codegen` not defined in `Enzyme.Compiler`
  | Stacktrace:
  | [1] #91
  | @ ~/.cache/julia-buildkite-plugin/depots/3cc01fab-3357-4a7a-9294-cde2d3115a97/dev/Enzyme/src/Enzyme.jl:1311
  | [2] #JuliaContext#152
  | @ ~/.cache/julia-buildkite-plugin/depots/3cc01fab-3357-4a7a-9294-cde2d3115a97/packages/GPUCompiler/Ecaql/src/driver.jl:34
  | [3] JuliaContext
  | @ ~/.cache/julia-buildkite-plugin/depots/3cc01fab-3357-4a7a-9294-cde2d3115a97/packages/GPUCompiler/Ecaql/src/driver.jl:25 [inlined]
  | [4] tape_type
  | @ ~/.cache/julia-buildkite-plugin/depots/3cc01fab-3357-4a7a-9294-cde2d3115a97/dev/Enzyme/src/Enzyme.jl:1310 [inlined]