Matthew Nicely

Add support for cmath

3

We should add support for `cmath` https://en.cppreference.com/w/cpp/header/cmath

``` $ nvcc -arch=sm_70 -c test.cu test.cu(2): error: calling a constexpr __host__ function("sqrt") from a __global__ function("test") is not allowed. The experimental flag '--expt-relaxed-constexpr' can be used to allow this....

Would it be possible to add batching functionality? It seems the code is highly optimized to process one image at a time, but not, say 1M images. Let me know...

Would it be possible to convert _cudaMallocPitch_ calls to _cudaMalloc_? I understand why cudaMallocPitch was chosen, but those limitations are not as noticeable today with larger cache sizes. The main...

Quantify difference between compares

1

Is it possible to quantify improvements between compares? I know with `--benchmark-compare-fail=:%` I can see which test performed worse, but I would also like to know how much better my...

nvCOMP v2.0

3

Would you be interested in integrating the latest version of nvCOMP (i.e., v2.0)?

Question concerning cuTENSOR integration

2

Are there plans to integrate the cuTENSOR branch into main? If yes, do you have a timeline? If no, are there any blockers? + @v0i0 @springer13

Fix complex type bug with correlate

7

Found a bug with correlate 1D + complex data types.

Elementwise Kernel optimizations.

3

Hi @evanmayer, do you have a way to run your code without hardware? I've found a few places to make improvements. If no, I've added ideas below if you ever...

[INFO] cuDNN with fusions

1

Hi, I came across this repo while wiring up Whisper support to my Bazaar. I just wanted to say amazing repo! Diving through the code I noticed there was a...