-
Similar to https://github.com/Capgemini/Apollo/blob/scaling/roles/haproxy/tasks/sysctl-tunning.yml
-
We should make use of a MPI_Info object.
### Some ressources:
- [HLRS/Jaguar Wiki](https://wickie.hlrs.de/platforms/index.php/MPI-IO) (MPI_Info)
- [ORNL Lustre Basics](https://www.olcf.ornl.gov/kb_art…
-
Cover how to identify a bottleneck (network input/output, CPU, disk I/O, memory) and how to improve Scrapy’s performance for any given bottleneck.
Cover `PYTHONOPTIMIZE` where relevant, but:
- Exp…
-
Using the mpich build 'mpich/20231026/icc-all-pmix-gpu' on sunspot and using darshan and vtune for performance analysis I am seeing what appears to be very bad performance in the messaging layer for t…
-
The use of `std::map` for the Event -> "operation pointer and metadata" lookup is much too slow. It needs to be switched over to a (concurrent) hash table of some sort.
-
## Description
it flings you km away now that theres less friction
-
### Question
I wonder about the performance when using Qwen2 as the LLM. Does it outperform the original LLaVA-v1.5?
By the way, are there any scripts for instruction tuning? I only found the scri…
-
[fixedtune.log](https://github.com/lightvector/KataGo/files/14325834/fixedtune.log)
[fixedtune11_gpugfx90c_x19_y19_c384_mv14.txt](https://github.com/lightvector/KataGo/files/14325836/fixedtune11_gpug…
-
### Bug summary
**First check**
- I added a descriptive title to this issue.
- I used the GitHub search to find a similar issue and didn't find it.
- I searched the Prefect documentation for …
-
hi, rocm expert,
wonder how 3000 this magic number is considered here ?
when bench matrix shape as [m, n], if m