-
### 🐛 Describe the bug
The following code some times hung and sometimes get `Native API failed. Native API returns: -2 (PI_ERROR_DEVICE_NOT_AVAILABLE) -2 `
Reproduce:
stock Pytorch main branch
…
-
### 🐛 Describe the bug
In theory, the output of any convolutional layer should be independent of the batch_size. For example, when I have a convolutional layer, once the input features [ci, h, w] are…
-
I want to evaluate inference performance using AMX-FP16 on new CPU while default docker or AI tool package not supported AMX-FP16 because of old version of onednn(3.2.0), so building onednn-3.3.0 and …
-
### Request Description
I found in the CMakelist file of GPU plugin has include the building macro for oneDNN, as shown below.
# # oneDNN for GPU plugin #
if(ENABLE_ONEDNN_FOR_GPU)
how do we che…
-
### 🚀 The feature, motivation and pitch
MKLDNN was renamed to oneDNN a long time ago and the cmake name is also different.
The internal copy of MKLDNN in PyTorch is very outdated.
While PyTorch c…
-
https://github.com/oneapi-src/oneDNN/ aka MKLDNN aka DNNL now has better performance for MT-size matrices: https://github.com/apache/incubator-mxnet/issues/17980 . And it's open source. The same tea…
-
Hello.
Are there any benchmark results evaluating the performance of AMX-enabled processors for DNN operators, such as convolution and matrix multiplication?
-
### 🐛 Describe the bug
We have optimized pytorch memory alloction performance for pytorch Windows,
please take reference: https://github.com/pytorch/pytorch/issues/102534.
But oneDNN still use Win…
-
Dear Zhi,
Thank you for sharing this amazing work! I have installed the repo following your instructions and was able to reproduce you showed in the YT video with the provided experimental data.
…
-
Save Completed!
Save Completed!
Save Completed!
Start parsing settings...
Start parsing settings...