-
**Is your feature request related to a problem? Please describe.**
Currently, all the code profiling is custom based on internal functions from psutil and/or pytorch. It would be great to have a more…
-
BPO | [31190](https://bugs.python.org/issue31190)
--- | :---
*Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.*
Show more detail…
-
Profiling at the ML framework level, extract some key performance indicators to reflect how the training process goes, e.g. time spent on data loading, preprocessing, forward pass, backward pass, each…
-
```
OS: Ubuntu 24.04 LTS
Kernel: 6.8.0-40-generic
GPU: AMD Radeon RX 6900 XT, AMD Radeon Instinct MI 50
ROCk module version: 6.8.5
ROCm: 6.2.0
```
With the latest ROCm installation, AMD…
-
## Bug Report
### Version
v0.23.0
This also affects all previous versions.
### Description
When timing information is computed (`idle_ns` and `busy_ns`) the decision of whether the la…
-
Our performance profiling tool intercepts application SHMEM functions, does some work, and then calls the related pshmem function in the OSHMEM library. OSHMEM is missing pshmem_info_get_version and …
-
For now imports seem really slow, make it faster for faster debug cycles
-
Hi
Thanks for your work in DIA-NN.
I used DIANN-1.8.1 to analyze the DIA files in library-based mode. The library contained 6 fragments for each precursor. Following is the cmd:
diann-1.8.1 --d…
-
### Describe the issue
I am testing Meta's Segment Anything (SAM) encoder model, both on Linux (CUDA) and on Windows (DirectML). When testing the model on the two platforms, using identical hardwar…
-
Hi!
Iv'e ran the crawler of the android repo json file, and downloaded the libs.
Since there are new versions of the libraries, i want to create new profiles for each, and then run the match command…