Enzyme support older versions

[wsmoses]

No description provided.

[vchuravy]

Can we name these files after the EnzymeCore version?

[wsmoses]

done

[vchuravy]

Can you also fix runic?

diff --git a/EnzymeCore08Ext.jl b/EnzymeCore08Ext.jl
index 265a897..8bdd8ee 100644
--- a/EnzymeCore08Ext.jl
+++ b/EnzymeCore08Ext.jl
@@ -12,7 +12,7 @@ function gpu_fwd(ctx, config, f, args...)
 end
 
 function EnzymeRules.forward(
-	config,
+        config,
         func::Const{<:Kernel{CPU}},
         ::Type{Const{Nothing}},
         args...;
@@ -27,7 +27,7 @@ function EnzymeRules.forward(
 end
 
 function EnzymeRules.forward(
-	config,
+        config,
         func::Const{<:Kernel{<:GPU}},
         ::Type{Const{Nothing}},
         args...;

[github-actions]

Benchmark Results

	main	46d2fcb...	main/46d2fcb10b5713...
saxpy/default/Float16/1024	2.81 ± 0.2 μs	2.79 ± 0.2 μs	1.01
saxpy/default/Float16/1048576	2.08 ± 0.0059 ms	2.08 ± 0.0091 ms	0.999
saxpy/default/Float16/16384	0.0328 ± 0.00015 ms	0.0328 ± 0.00015 ms	1
saxpy/default/Float16/2048	5.22 ± 0.047 μs	5.23 ± 0.092 μs	0.999
saxpy/default/Float16/256	0.967 ± 0.11 μs	0.979 ± 0.12 μs	0.988
saxpy/default/Float16/262144	0.524 ± 0.0094 ms	0.525 ± 0.0096 ms	1
saxpy/default/Float16/32768	0.065 ± 0.00018 ms	0.065 ± 0.00018 ms	1
saxpy/default/Float16/4096	10.1 ± 0.05 μs	10.1 ± 0.05 μs	1
saxpy/default/Float16/512	1.57 ± 0.16 μs	1.57 ± 0.16 μs	0.999
saxpy/default/Float16/64	0.631 ± 0.016 μs	0.62 ± 0.017 μs	1.02
saxpy/default/Float16/65536	0.129 ± 0.00035 ms	0.129 ± 0.00043 ms	1
saxpy/default/Float32/1024	1.02 ± 0.015 μs	1.02 ± 0.011 μs	1.01
saxpy/default/Float32/1048576	0.966 ± 0.0084 ms	0.964 ± 0.0091 ms	1
saxpy/default/Float32/16384	15.5 ± 0.13 μs	15.4 ± 0.1 μs	1
saxpy/default/Float32/2048	1.72 ± 0.023 μs	1.71 ± 0.02 μs	1
saxpy/default/Float32/256	0.535 ± 0.12 μs	0.526 ± 0.13 μs	1.02
saxpy/default/Float32/262144	0.238 ± 0.0096 ms	0.239 ± 0.0099 ms	0.998
saxpy/default/Float32/32768	30.3 ± 0.17 μs	30.3 ± 0.16 μs	1
saxpy/default/Float32/4096	3.01 ± 0.025 μs	3.02 ± 0.024 μs	0.999
saxpy/default/Float32/512	0.691 ± 0.11 μs	0.691 ± 0.12 μs	1
saxpy/default/Float32/64	0.41 ± 0.0061 μs	0.408 ± 0.0074 μs	1
saxpy/default/Float32/65536	0.0601 ± 0.00026 ms	0.0601 ± 0.00043 ms	1
saxpy/default/Float64/1024	1.07 ± 0.022 μs	1.06 ± 0.018 μs	1.01
saxpy/default/Float64/1048576	1.02 ± 0.032 ms	1 ± 0.026 ms	1.02
saxpy/default/Float64/16384	15.7 ± 0.13 μs	15.7 ± 0.12 μs	1
saxpy/default/Float64/2048	1.75 ± 0.028 μs	1.74 ± 0.019 μs	1
saxpy/default/Float64/256	0.516 ± 0.013 μs	0.518 ± 0.013 μs	0.995
saxpy/default/Float64/262144	0.243 ± 0.0096 ms	0.244 ± 0.01 ms	0.998
saxpy/default/Float64/32768	31.1 ± 0.75 μs	31.1 ± 0.59 μs	1
saxpy/default/Float64/4096	3.06 ± 0.093 μs	3.04 ± 0.078 μs	1.01
saxpy/default/Float64/512	0.699 ± 0.12 μs	0.702 ± 0.11 μs	0.997
saxpy/default/Float64/64	0.396 ± 0.011 μs	0.385 ± 0.0076 μs	1.03
saxpy/default/Float64/65536	0.0614 ± 0.00084 ms	0.0615 ± 0.00085 ms	0.998
saxpy/static workgroup=(1024,)/Float16/1024	2.1 ± 0.22 μs	2.11 ± 0.21 μs	0.994
saxpy/static workgroup=(1024,)/Float16/1048576	0.166 ± 0.012 ms	0.161 ± 0.0098 ms	1.03
saxpy/static workgroup=(1024,)/Float16/16384	4.28 ± 0.21 μs	4.28 ± 0.21 μs	1
saxpy/static workgroup=(1024,)/Float16/2048	2.13 ± 0.23 μs	2.14 ± 0.22 μs	0.993
saxpy/static workgroup=(1024,)/Float16/256	2.65 ± 0.039 μs	2.64 ± 0.038 μs	1
saxpy/static workgroup=(1024,)/Float16/262144	0.0433 ± 0.0024 ms	0.0431 ± 0.0019 ms	1
saxpy/static workgroup=(1024,)/Float16/32768	6.54 ± 0.17 μs	6.6 ± 0.21 μs	0.992
saxpy/static workgroup=(1024,)/Float16/4096	2.42 ± 0.033 μs	2.45 ± 0.043 μs	0.987
saxpy/static workgroup=(1024,)/Float16/512	3.14 ± 0.095 μs	3.16 ± 0.097 μs	0.996
saxpy/static workgroup=(1024,)/Float16/64	2.27 ± 0.026 μs	2.24 ± 0.021 μs	1.01
saxpy/static workgroup=(1024,)/Float16/65536	12.2 ± 0.34 μs	12.3 ± 0.29 μs	0.994
saxpy/static workgroup=(1024,)/Float32/1024	1.96 ± 0.026 μs	1.97 ± 0.031 μs	0.996
saxpy/static workgroup=(1024,)/Float32/1048576	0.253 ± 0.019 ms	0.251 ± 0.019 ms	1.01
saxpy/static workgroup=(1024,)/Float32/16384	4.11 ± 0.89 μs	4.1 ± 0.95 μs	1
saxpy/static workgroup=(1024,)/Float32/2048	2.29 ± 0.31 μs	2.31 ± 0.23 μs	0.994
saxpy/static workgroup=(1024,)/Float32/256	2.8 ± 1.6 μs	2.84 ± 1 μs	0.984
saxpy/static workgroup=(1024,)/Float32/262144	0.0664 ± 0.0036 ms	0.066 ± 0.004 ms	1.01
saxpy/static workgroup=(1024,)/Float32/32768	6.98 ± 0.41 μs	6.95 ± 0.24 μs	1
saxpy/static workgroup=(1024,)/Float32/4096	2.57 ± 0.19 μs	2.56 ± 0.22 μs	1
saxpy/static workgroup=(1024,)/Float32/512	2.51 ± 0.22 μs	2.49 ± 0.23 μs	1.01
saxpy/static workgroup=(1024,)/Float32/64	2.45 ± 0.051 μs	2.45 ± 0.056 μs	0.999
saxpy/static workgroup=(1024,)/Float32/65536	17.7 ± 2.1 μs	17.6 ± 1.7 μs	1.01
saxpy/static workgroup=(1024,)/Float64/1024	2.05 ± 0.03 μs	2.05 ± 0.03 μs	1
saxpy/static workgroup=(1024,)/Float64/1048576	0.552 ± 0.058 ms	0.529 ± 0.047 ms	1.04
saxpy/static workgroup=(1024,)/Float64/16384	6.91 ± 0.78 μs	6.99 ± 1.4 μs	0.988
saxpy/static workgroup=(1024,)/Float64/2048	2.54 ± 0.28 μs	2.54 ± 0.26 μs	1
saxpy/static workgroup=(1024,)/Float64/256	2.4 ± 0.055 μs	2.41 ± 0.053 μs	0.997
saxpy/static workgroup=(1024,)/Float64/262144	0.123 ± 0.0089 ms	0.123 ± 0.0087 ms	0.999
saxpy/static workgroup=(1024,)/Float64/32768	17.4 ± 1.7 μs	17.1 ± 1.9 μs	1.02
saxpy/static workgroup=(1024,)/Float64/4096	3.1 ± 0.34 μs	3.11 ± 0.35 μs	0.997
saxpy/static workgroup=(1024,)/Float64/512	2.39 ± 0.045 μs	2.4 ± 0.049 μs	0.998
saxpy/static workgroup=(1024,)/Float64/64	2.38 ± 0.087 μs	2.37 ± 0.1 μs	1
saxpy/static workgroup=(1024,)/Float64/65536	0.0356 ± 0.003 ms	0.0352 ± 0.0028 ms	1.01
time_to_load	0.313 ± 0.0015 s	0.311 ± 0.0015 s	1.01

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).

[wsmoses]

is there a similar clang-format like command I can run to auto fix?

[vchuravy]

Yeah you can https://github.com/fredrikekre/Runic.jl

julia --project=@runic -e 'using Pkg; Pkg.add(url = "https://github.com/fredrikekre/Runic.jl")'
alias runic="julia --project=@runic -e 'using Runic; exit(Runic.main(ARGS))' --"
 runic -i ext src