Get tests passing in Julia 10 #103
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## master #103 +/- ##
===========================================
+ Coverage 64.04% 89.28% +25.24%
===========================================
Files 34 34
Lines 2581 2595 +14
===========================================
+ Hits 1653 2317 +664
+ Misses 928 278 -650 ☔ View full report in Codecov by Sentry. |
mloubout
reviewed
Mar 26, 2024
.github/workflows/runtests.yml
Outdated
| env: | ||
| test_suite: ${{ matrix.tests }} | ||
|
|
||
| strategy: | ||
| fail-fast: false | ||
|
|
||
| matrix: | ||
| version: ['1.6', '1.7', '1.8'] | ||
| version: ['1.6', '1.7', '1.8', '1.10'] |
There was a problem hiding this comment.
Wanted to start with the superset of 1.10, and then add 1.9. In retrospect might have been smarter to go more incrementally and do 1.9 first. Will add 1.9 in a bit
| @@ -110,7 +110,10 @@ end | |||
| function mat_tens_i(out::AbstractVector{T}, Mat::AbstractArray{T, 2}, | |||
| Tens::AbstractArray{T, 3}, Mat2::AbstractArray{T, 2}) where T | |||
| # Computes sum( (Mat * tens) .* Mat2) for each element in the batch | |||
|
|
|||
| isa(Mat, CUDA.CuArray) && (Mat2 = CUDA.CuArray(Mat2)) #new Julia 1.10 subarrays require this | |||
There was a problem hiding this comment.
I think just Mat2 = collect(Mat2) should always work
There was a problem hiding this comment.
Darn looks like that puts the Mat2 on CPU:
Mat2 = collect(Mat2)
copyto!(out, map(i -> dot(Mat * Tens[i, :, :], Mat2) , 1:size(Tens, 1)))
LoadError: MethodError: no method matching dot(::Int64, ::CUDA.CuPtr{Float32}, ::Int64, ::Ptr{Float32}, ::Int64)
Closest candidates are:
dot(::Integer, ::Union{Ptr{Float32}, AbstractArray{Float32}}, ::Integer, ::Union{Ptr{Float32}, AbstractArray{Float32}}, ::Integer)
@ LinearAlgebra /net/coc-herrmann/GATechBundle/Julia/julia-1.10.0/share/julia/stdlib/v1.10/LinearAlgebra/src/blas.jl:344
dot(::Integer, ::Union{Ptr{Float64}, AbstractArray{Float64}}, ::Integer, ::Union{Ptr{Float64}, AbstractArray{Float64}}, ::Integer)
@ LinearAlgebra /net/coc-herrmann/GATechBundle/Julia/julia-1.10.0/share/julia/stdlib/v1.10/LinearAlgebra/src/blas.jl:344
Stacktrace:
[1] dot(x::CUDA.CuArray{Float32, 2, CUDA.Mem.DeviceBuffer}, y::Matrix{Float32})
@ LinearAlgebra.BLAS /net/coc-herrmann/GATechBundle/Julia/julia-1.10.0/share/julia/stdlib/v1.10/LinearAlgebra/src/blas.jl:395
[2] dot(x::CUDA.CuArray{Float32, 2, CUDA.Mem.DeviceBuffer}, y::Matrix{Float32})
@ LinearAlgebra /net/coc-herrmann/GATechBundle/Julia/julia-1.10.0/share/julia/stdlib/v1.10/LinearAlgebra/src/matmul.jl:15
[3] (::InvertibleNetworks.var"#249#250"{CUDA.CuArray{Float32, 2, CUDA.Mem.DeviceBuffer}, CUDA.CuArray{Float32, 3, CUDA.Mem.DeviceBuffer}, Matrix{Float32}})(i::Int64)
@ InvertibleNetworks ~/.julia/dev/InvertibleNetworks/src/layers/invertible_layer_conv1x1.jl:116
[4] iterate
@ Base ./generator.jl:47 [inlined]
[5] _collect
@ Base ./array.jl:854 [inlined]
[6] collect_similar
@ Base ./array.jl:763 [inlined]
[7] map
@ Base ./abstractarray.jl:3282 [inlined]
[8] mat_tens_i(out::CUDA.CuArray{Float32, 1, CUDA.Mem.DeviceBuffer}, Mat::CUDA.CuArray{Float32, 2, CUDA.Mem.DeviceBuffer}, Tens::CUDA.CuArray{Float32, 3, CUDA.Mem.DeviceBuffer}, Mat2::Base.ReshapedArray{Float32, 2, SubArray{Float32, 3, CUDA.CuArray{Float32, 4, CUDA.Mem.DeviceBuffer}, Tuple{Base.Slice{Base.OneTo{Int64}}, Base.Slice{Base.OneTo{Int64}}, Base.Slice{Base.OneTo{Int64}}, Int64}, true}, Tuple{}})
test/test_networks/test_glow.jl
Outdated
| @@ -175,7 +175,7 @@ for logdet_bool = [true,false] #logdet is not tested in Jacobian yet. | |||
| dX_, dθ_, _, _ = G.adjointJacobian(dY_, Y) | |||
| a = dot(dY, dY_) | |||
| b = dot(dX, dX_) + dot(dθ, dθ_) | |||
| @test isapprox(a, b; rtol=1f-3) | |||
| @test isapprox(a, b; rtol=1f-1) | |||
mloubout
reviewed
Mar 26, 2024
.github/workflows/runtests.yml
Outdated
| @@ -11,15 +11,15 @@ on: | |||
| jobs: | |||
| test: | |||
| name: Julia ${{ matrix.version }} - ${{ matrix.tests }} - ${{ matrix.os }} | |||
| runs-on: ${{ matrix.os }} | |||
| runs-on: self-hosted | |||
There was a problem hiding this comment.
That's overkill you can add an extra to the matrix to do just one run on GPU
matrix:
version: ['1.6', '1.7', '1.8', '1.10']
tests: ["basics", "layers", "networks"]
os: [ubuntu-latest]
include:
- os: self-hosted
tests: whichever
version: whichever
and then just needs
runs-on: ${{ matrix.os }}
There was a problem hiding this comment.
SMART thank you!
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Need to run tests on GPU before merging.