-
This issue is a repro for correctness issue with the igemm pipeline.
1. Start with this branch https://github.com/Max191/iree/tree/llvmgpu-tile-and-fuse-wip
2. Input mlir:
```mlir
!input_type = …
-
# Summary
Precise in the doc if reorder is safe to use inplace, where src mem = dst mem
# URLs
https://oneapi-src.github.io/oneDNN/dev_guide_reorder.html#doxid-dev-guide-reorder
# Additional d…
-
I ran `conda remove --force mkl mkl-service` and PyMC3 seemed to still work fine. Is it still a requirement?
It's quite large, so it would be nice to get rid of.
-
Hello Hadrien,
I was looking for some nice map-reduce benchmarks to add to evaluate my multithreading runtime ([Weave](https://github.com/mratsim/weave)) and was thinking that histograms could be a…
-
I'm trying to install ctf on a linux box, first loading my openmpi module, then running this config command
./configure CXX=mpicxx --build-scalapack --build-hptt --with-hptt --with-scalapack --with-l…
-
Let's discuss new operations that we might like to add to BLIS, specifically those that would fall into level-1v or level-1m families (and perhaps level-2):
- [ ] element-wise vector/matrix multipl…
-
A core feature of any tensor library is being able to transpose tensors.
This is something that can be done efficiently when you have implemented the concept of strides and store those strides togeth…
-
**Describe the bug**
When loading multiple nrrd files with different ijk_to_ras matrices and resampling / concatenating them, results have improper relative positions / orientations
**To Reproduce…
-
When doing some analysis into the performance of the Transpose operator, I noticed that performance is significantly worse when the copy done in [`contiguous_data`](https://github.com/robertknight/rte…
-
TF provides the `TensorArray` to make automatic iteration and stacking efficient in `scan` or `while_loop`.
The naive variant with gathering and concatenating or dynamic updates would be inefficien…