-
Currently, its realistic usage involves a macro (`__COUNTER__`):
`Type & var = block::shared::st::allocVar(acc);`
Perhaps it would be more intuitive if we provide a macro for the whole expression …
-
For example, when I send the program `H 1`, the QVM rewrites 1 -> 0 as is expected behavior for using the least number of qubits possible, but for shared memory mode I actually want to apply a Hadamar…
-
## 🚀 Feature
Tensor ctors should support directly creating in shared memory. The current workaround is either the inefficient `.share_memory_()` after creation or private APIs like
https://githu…
-
## 🐛 Bug
Any operation (read or write) on a shared memory tensor hangs if it involves more than `at::GRAIN_SIZE` elements (131076 bytes) because it gets dispatched on multiple threads using `parall…
-
## 🐛 Bug
DataLoader consumes extremely large shared memory (shm) in its initialization (when calling `iter(dataloader)`) compared to following training steps. For example, it consumes up to 150GB s…
-
# Description of problem
postgres uses the kernel shared memory achitecture `shm`.
It looks like clear containers don't properly initialise this value, causing an error like:
```
docker run -d…
-
Trying to run Remix on Kane and Lynch 2 Dog Days, I get following error after launching game.
> Error : This game requires a display driver that supports DirectX9.0c 3D acceleration.
- Remix-0.1.1…
-
### Information
- **Qiskit Metal version - latest from the source as per 22.10.2024**:
- **Python version - the one that comes along as specified in environment.yml**:
- **Operating system …
-
### Description
**Hi, there. While applying FusedAttention with jax-triton, the following XLA error happens on Nvidia-A100:**
2023-08-28 03:06:51.319566: W external/xla/xla/service/gpu/runtime/s…
-
After some time system may hang.
### Steps to reproduce
The problem appears on singleboard computer as RaspberryPi/OrangePi with Ubuntu, Debian, Armbian distros right up to kernel 6.1
for examp…