Issues
- 10
- 0
- 9
- 3
How can we make KA fast on CPUs?
#509 opened by avik-pal - 3
- 0
[FR] Add nthreads argument to CPU backend
#507 opened by cncastillo - 0
- 0
Document/Decide on `adapt(Backend, ...)` API
#501 opened by vchuravy - 2
Add Enzyme CI for other backends than CPU and CUDA
#492 opened by vchuravy - 3
[EnzymeExt] tape_type error
#495 opened by michel2323 - 0
- 1
Atomic operations on complex numbers
#497 opened by nHackel - 2
External functions in GPU ODE example
#496 opened by H-Sax - 0
[Enzyme] NTuple index not working on CUDA backend
#494 opened by michel2323 - 0
- 6
LLVM error when using Atomix.@atomic with FP16
#486 opened by ymtoo - 2
KA.functional(::Backend) to mimick CUDA.functional
#489 opened by vchuravy - 1
`Base.rand(::Backend)`
#488 opened by vchuravy - 21
Printing results changes results in CPU kernel
#485 opened by jlk9 - 18
- 3
Small discrepancy
#482 opened by ymtoo - 2
- 2
Slow simple 2D copy kernel with Metal backend
#464 opened by LaurentPlagne - 0
`KA.copyto!` between gpu arrays not tested.
#474 opened by christiangnrd - 1
- 14
- 3
- 6
- 2
MAX support?
#460 opened by ManuelCostanzo - 9
- 2
Allow `return` or `return nothing` in kernels
#443 opened by eschnett - 2
- 5
- 7
CPU `__thread_run` could loop over CartesianIndices?
#448 opened by rafaqz - 7
Looping over high-dimensional arrays
#438 opened by b-fg - 0
Replace ROCArrays by AMDGPU in documentation?
#447 opened by eschnett - 1
On CPU always use `NoDynamicCheck()`, just finish the last partial workgroup with `DynamicCheck()`
#449 opened by rafaqz - 12
exposing warp-level semantics
#420 opened by leios - 2
- 1
Dynamic parallelism
#442 opened by simone-silvestri - 1
`inbounds=true` broken for CPU kernels
#430 opened by leios - 0
Julia 1.11 private tests are failing
#433 opened by vchuravy - 6
unsupported dynamic function invocation (call to `_cuprint`) on CUDA 5.0
#425 opened by charleskawczynski - 4
Index type
#396 opened by vchuravy - 7
`@unroll` unrolling the wrong loop
#411 opened by lcw - 1
Reverse CI for AMDGPU
#394 opened by vchuravy - 0
Enable Enzyme reverse mode rules for all backends
#408 opened by vchuravy - 4
Possible error due to KernelAbstractions while using VSCode debugger
#405 opened by sophia-wright-blue - 4
KA seems to be broken for CUDA
#400 opened by benedict-96 - 2
Add API for device selection
#395 opened by jkravs