-
**Originally by _gropp_ on 2012-03-07 09:30:07 -0600**
---
I ran the code checker over the current version of the MPI-3 standard. Attached are my recommended fixes for the collective chapter.
This …
-
**Originally by _htor_ on 2011-01-17 13:04:26 -0600**
---
Author: Torsten Hoefler, Christian Siebert
## Description
In the existing vector collectives, the user needs to specify all displacements a…
-
Could you explain what does above argument means and if that help improve performance?
Thanks
Amit Ruhela
-
As a replacement for [Monthly Reports](https://github.com/opencollective/opencollective/issues/6712).
To consider:
- https://github.com/opencollective/opencollective/issues/6037
_Issue to be co…
-
Hello
After some confusement about collectives in Modality I had a chat with Alberto Decampo (@adcxyz) on the mailing list and we seem to have discovered a bug with collectives.
My goal was to…
-
**Is your feature request related to a problem? Please describe.**
Attaching an image (say, in Collectives) will embed it, but with no possibility to further format the layout - alignment, scaling, e…
-
#3655 adds the ability to pass errors up and down the call stack for collectives, in addition too a bunch more tests that are similar to `bcastlength` to validate that the various collectives are corr…
-
### 🚀 The feature, motivation and pitch
# Motivation
This Request for Comments (RFC) document aims to propose and discuss Intel GPU distributed support in PyTorch, introducing Intel GPU distribu…
-
How does NCCL decide which NIC to use for inter-node send / recv operations over sockets?
I'm running benchmarks where multiple GPUs on a node are sending data to the corresponding GPU on another nod…
-