-
brisck.cc and ncrisck.cc (and erisc?) should issue a write_barrier to sync the noc before completing and telling dispatcher that everything is done. not doing so is a race, for example, a kernel writ…
-
https://github.com/josephbergevin/codebook-md/blob/5ebb4089a05705d4c6d359a6d3c3e117d2c74ddf/package.json#L64-L68
-
### 🐛 Describe the bug
Async NCCL comminucations from `torch.distributed` should run in parallel with CUDA computing kernels, but traces from `torch.profiler` shows it is not true for the first run. …
-
Might shrinker_to_text end up called with a null shrinker? Though I don't see how exactly.
CONFIG_SHRINKER_DEBUG was off.
From 1d875e4e9cf2ba1542c7279a97f64e5971804d1c (bcachefs-testing).
``…
-
Linux kernel console does not send `Ctrl`+`Fn` at all. In the same time, `Search and replace` is one of frequently used functions during system restore and config files edition. And there is on way to…
unxed updated
2 months ago
-
I've encountered a broken symbolic link for `/lib/modules/$kver` if the `--root` points to a rootfs directory that is not the parent of the directory where the kernel sources are located.
E.g. kernel…
-
hi, In some cases, the performance of CUDA Memcpy is better than that of GPU kernel.
is there any way to avoid gpu kernel but cuda memcpy in p2p sendrecv operation?
by the way, NCCL_P2P_USE_CUDA_ME…
-
Kernel outputs in the AVM are organised into columns based on a specific offset and are sorted by an incrementing `side_effect_counter`.
Changes to `NOTEHASHEXISTS` and `L1TOL2MSGEXISTS` means we …
-
I'd like to add support for `rust-gpu` in the not-so-distant future. I have some questions while I figure out the plan:
1. Would it make sense to have shaders written with `rust-gpu` to be hung off…
-
### Bug Description
image/kernel-tracking-stable checks if the snap is in stable channel but it doesn't make sense for testing snap in beta.
It should be removed as it produces redundant error.
Exa…