Skip to content

Pull requests: JuliaGPU/CUDA.jl

New pull request New
52 Open 1,420 Closed
52 Open 1,420 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Reviews
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

Enzyme: mark device_sync as non-differentiable [only downstream]
#2605 opened Dec 26, 2024 by wsmoses Loading…
1
Move strided batch pointer conversion to GPU
#2601 opened Dec 20, 2024 by THargreaves Loading…
5
Use GPUArrays caching allocator cuda array Stuff about CuArray. performance How fast can we go?
#2593 opened Dec 15, 2024 by pxl-th • Draft
make CUDA randn work with Zygote
#2581 opened Dec 9, 2024 by bgctw Loading…
5
Directed rounding cuda kernels Stuff about writing CUDA kernels. enhancement New feature or request needs tests Tests are requested.
#2576 opened Dec 3, 2024 by orkolorko Loading…
15
[CUSPARSE] Fix constructor of sparse empty matrices bugfix This gets something working again. cuda libraries Stuff about CUDA library wrappers.
#2575 opened Dec 2, 2024 by amontoison Loading…
1
Add function for recursively printing parameter memory cuda kernels Stuff about writing CUDA kernels. enhancement New feature or request speculative Not sure about this one yet.
#2560 opened Nov 21, 2024 by charleskawczynski Loading…
8
WIP: Native I/O. cuda kernels Stuff about writing CUDA kernels. speculative Not sure about this one yet.
#2485 opened Sep 5, 2024 by maleadt • Draft
Bugfix for batched gemv bugfix This gets something working again. needs changes Changes are needed.
#2481 opened Aug 28, 2024 by kose-y • Draft
14
High Level Wrapper for Fused Matmul + Bias + Activation cuda libraries Stuff about CUDA library wrappers. enhancement New feature or request
#2360 opened May 4, 2024 by avik-pal • Draft
Mark more methods as device methods enhancement New feature or request
#2336 opened Apr 19, 2024 by vchuravy • Draft
4
Use PrecompileTools to warmup CUDA.jl enhancement New feature or request needs changes Changes are needed.
#2325 opened Apr 15, 2024 by vchuravy • Draft
5
Add a dispatch for LinearAlgebra.norm2 cuda array Stuff about CuArray. needs changes Changes are needed.
#2302 opened Mar 22, 2024 by sharanry • Draft
1
1
docs: perf tips: deemphasize assume in favor of UnsafeAssume.jl
#2181 opened Nov 29, 2023 by nsajko Loading…
8
KA 0.10 API changes
#2084 opened Sep 19, 2023 by vchuravy • Draft
Use TaskLocalValues
#2075 opened Sep 8, 2023 by vchuravy • Draft
Support FFT adjoint plans and test cuda libraries Stuff about CUDA library wrappers. enhancement New feature or request
#2073 opened Sep 4, 2023 by gaurav-arya • Draft
Add contract through FastmathOverlays.jl cuda kernels Stuff about writing CUDA kernels. enhancement New feature or request
#2037 opened Aug 16, 2023 by vchuravy • Draft
1
WIP: Add an index typevar to CuDeviceArray. enhancement New feature or request help wanted Extra attention is needed performance How fast can we go?
#1895 opened May 3, 2023 by maleadt • Draft
11
Add an experimental opaque closure type. cuda kernels Stuff about writing CUDA kernels. enhancement New feature or request speculative Not sure about this one yet.
#1853 opened Apr 4, 2023 by maleadt • Draft
5
Attempted to add method definitions for supporting the SubArray type …
#1830 opened Mar 24, 2023 by JamieMair • Draft
1
Add Julian API for NVPerf and CUPTI Profiler
#1826 opened Mar 23, 2023 by vchuravy • Draft
1
Add wrappers for NVPERF
#1823 opened Mar 22, 2023 by vchuravy • Draft
1
Use Atomix
#1790 opened Mar 10, 2023 by vchuravy • Draft
4 tasks
4
Adding copyto for non-contiguous matrices and vectors cuda array Stuff about CuArray.
#1778 opened Feb 22, 2023 by evelyne-ringoot • Draft
3
ProTip! Find all pull requests that aren't related to any open issues with -linked:issue.