-
On the PR https://github.com/ros2/rclcpp/pull/2624 I provide a fix to allow the user to make copies of loaned messages, keeping them beyond the subscription callback scope. With the fix, the loaned me…
-
### What happened?
When forcing llama.cpp to use "GPU + CUDA + VRAM + shared memory (UMA)", we noticed:
- High CPU load (even when only GPU should be used)
- Worse performance than using "CPU + RAM…
-
### Description
When testing Velox spilling functionality, we found that the velox spilling would have performance issue.
Although the spilling has a thread pool up to 8 threads, the actual CPU co…
-
**Description**
https://github.com/triton-inference-server/server/pull/7567 introduced a reference counter to shared memory region to prevent users from releasing shared memory regions while any infe…
hcho3 updated
3 hours ago
-
- [✓] I was not able to find an [open](https://github.com/git-for-windows/git/issues?q=is%3Aopen) or [closed](https://github.com/git-for-windows/git/issues?q=is%3Aclosed) issue matching what I'm seei…
-
### Problem Description
While testing memory allocation using `hipMalloc()` in a HIP environment, I observed an unexpected behavior when attempting to allocate large chunks of memory.
Key Observ…
-
# Crash report
### What happened?
It's possible to abort the interpreter by calling `multiprocessing.shared_memory.ShareableList.count` in threads with `PYTHON_GIL=0` in a debug build:
```python
im…
-
zenoh in Rust implement a zero-copy shared memory transport. See examples of use:
- [zn_pub_shm.rs](https://github.com/eclipse-zenoh/zenoh/blob/9de48132a8780f853f0de38ab564a57b2687c968/zenoh/example…
-
Only create link when object is created?
Or create one big chunk of memory that gets build each frame?
Creating a link has the benefit of dynamic size.
Reading waveforms!? 🤘
-
https://s3.amazonaws.com/clickhouse-test-reports/72104/4563e796887121bf6ef6d9f64a3e973cb67becac/stress_test__debug_.html
```
[ 56227 ] {407a3079-a7fb-4a0d-a214-c6dd5248ad9d} : Logical error: \'fa…