-
```
➜ RUST_BACKTRACE=1 cargo run --example gelu -F cuda
Finished `dev` profile [optimized + debuginfo] target(s) in 0.22s
Running `target/debug/examples/gelu`
thread 'main' panicked at /…
-
Currently, specifying 'torch_cuda' as the backend appears to select the first CUDA device visible to PyTorch (cuda:0). However, in multi-gpu systems, it would be useful to specify a specific CUDA devi…
-
### What happened?
```
$ CUDA_VISIBLE_DEVICES=0 build/bin/Release/rpc-server -p 50052 --mem 10000
create_backend: using CUDA backend
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: G…
-
### Solution to issue cannot be found in the documentation.
- [X] I checked the documentation.
### Issue
## Steps to reproduce
1. Install the latest JAX and Keras versions with GPU support from …
-
### Description
I noticed in [Veros](https://github.com/team-ocean/veros) that we would sometimes observe solver divergences on GPU, but not on CPU. I've isolated the problem to a kernel that gives d…
-
I notice your repo use the gsplat which is from the old version of cuda backend. The newest version of gsplat have many good features such as mip-splatting, and memory efficient, would you update your…
-
While running JAX compiled with CUDA 12.6, I get error messages to the effect of
```
'+ptx86' is not a recognized feature for this target (ignoring feature)
```
Per the table [here](https://do…
-
How can I run de LoRA trainer on my second Cuda device ?
It seems to be working nice on the first cuda device :
```
...
08/22/2024 17:04:37 - INFO - __main__ - Distributed environment: DEEPSPEE…
-
### Describe the bug
Building VS project failed when the backend is CUDA 12.5.
### Steps to reproduce
1. Prerequisites:
1. Windows 11,
2. CUDA 12.5,
3. MSVC 19.41 (VS 2022 Preview),
…
-
The autotuner is currently tied closely to the CUDA backend: it takes a bunch of CUDA-specific parameters and then passes them to `do_bench` or `do_bench_cudagraph`, which both call into many CUDA-spe…