Sergey 'Jin' Bostandzhyan

I think I finally got somewhere, just not as I expected. After various debugging attempts I stumbled upon a similar issue https://github.com/ROCm/ROCm/issues/2418 related to `hipMemcpy` which among other things suggested...

@Artisto1 you are dealing with cutting edge stuff which is in development, don't expect a Windows-like installer and an icon to launch the app; as much progress as Linux has...

> But the example never stops without results. Same here, I tried the yolov8 one: `cargo run --features hip --example yolo-v8 --release -- candle-examples/examples/yolo-v8/assets/bike.jpg` it downloaded the model and then...

> The implementation take the first GPU available and some GPU doesn't seems to work as expected. I have the same behavior on my computer, by default HIP select the...

I have zero experience with GPU programming, so maybe someone could chime in. I attached with gdb to the Yolo example, here's `thread apply all bt`: ``` (gdb) thread apply...

For those who are not suffering from the ROCm related CPU-hog bug, this project looks like a very interesting alternative: https://github.com/vosen/ZLUDA If it does what it says it does, we...

@vberthet I had to patch build.rs, `"-parallel-jobs=15"` did not work for me (Rawhide 40), I had to remove it in order to compile. ``` clang: error: unknown argument: '-parallel-jobs=15' thread...

I am currently on RawHide which provides ROCm 6.0.0, although judging from the comments in the issue this problem was present in 5.7 as well. I could perhaps try to...

I finally got past the ROCm hanging memcpy issue, turned out it was enough to `export HSA_ENABLE_SDMA=0`, took me a while to find this option though. So, once that worked...

OK, so... after a longer time I finally got back to this and I am not sure what changed - I guess I fixed my installation without realizing it -...