-
# Summary
onMKL shows underwhelmed SGEMM performance for small matrices.
# Version
oneMKL v0.2
# Environment
* HW you use: NVIDIA V100
* Backend library version: CUDA/10.2
* OS name and ve…
-
## Summary
This is a request for a new code sample called fourier_correlation, which composes the Fourier correlation algorithm from oneMKL functions.
##Purpose
The samplesy show how to compose m…
-
**Read my next post**
Originally I got confused by the `ormqr()`, `ormrq()` and the LAPACK `ormql()` methods. I elaborated on the confusion and the problems with OneMKL in my next post. I will keep…
-
This was noticed on Intel devcloud, with Julia 1.7-rc1, and the `tb/onemkl` branch
```Julia
julia> x1 = oneArray(rand(576, 64, 2));
julia> x2 = oneArray(rand(576, 64, 36));
julia> function t…
-
Steps to reproduce:
```python
with np.errstate(invalid='ignore', divide='raise'):
for i in range(10000):
z = np.zeros(i)
try:
z / z
except(FloatingPoin…
evfro updated
3 years ago
-
I want to call [`mkl_get_dynamic`](https://www.intel.com/content/www/us/en/develop/documentation/onemkl-developer-reference-c/top/support-functions/threading-control/mkl-get-dynamic.html) and [`mkl_se…
-
oneMKL open source interfaces is using the context "get" API for CUBLAS backend. As noted below, it is being deprecated. What is the recommendation to replace this API?
https://github.com/intel/llv…
-
I've optimized several DL models so far using both Openvino and TensorRT.
And recently I'm trying to implement some non-supported, custom ops (like NonZero or GatherND) using OneAPI as it would be …
-
Hi. Is something wrong with the path of Intel MKL?? MKL path on Ubuntu 16 is here: __/home/turgut/intel/oneapi__
```
turgut@ubuntu:~/arrayfire/build$ cmake ..
-- Could NOT find OpenCL (missing: …
-
# Summary
When compiling oneMKL with both tests and cuRAND or cuBLAS enabled a compilation error occurs.
# Version
The current oneMKL develop head is used eg: 1ed12c7270a68b78c29178ccc582d0239a4b…