#Rusticl + #radeonsi works with #hashcat! The dictionary mode at least.

Trying to use brute force fails on the self test.

Still better than clover, which makes the GPU hang up so badly I can't even turn off the computer.

This is awesome cause this #APU (#Ryzen 2400G) is not supported by #rocm. Rusticl is the best #opencl I can get right now.

🧵10/9
On PVC (4x GPU Max 1100), P2P transfer does not work either. Both the implicit and explicit #OpenCL buffer migration variants cut performance in half compared to when sending buffers over the CPU. ❌

@ProjectPhysX making a massive scaling experiment on LUMI using only #opencl (no MPI) is still on our bucket list. Let us know when/if the code base might be more ready for that.

So deceiving that #beaglebone AI 64 does not support #opencl ! Or does it? There is a libOpenCL.so but did not manage to compile anything to use it...🫤
Will have to look now into #DSP with #c66x but already so few documentation...

@karolherbst congratulations this is an amazing achievement :blobcatcheer: . I remmber hearing about this on @BrodieOnLinux video and getting so hyped for it :ablobcatattention: .
P.s link to the video: https://www.youtube.com/watch?v=Xbb_ytPRipM
#mesa #opencl #rust #rustlang #dev #developer

@karolherbst Well, this gave me the motivation to hack my codegen so it doesn't ask for 64-bit atomics unless actually needed (and available). And... rusticl works! At least for some programs (the important visualisation ones), alhough I seem to get NaNs for others (the boring numerical ones). Thanks for your work; now I can finally use #OpenCL again on my AMD laptop.

chipStar 1.0 released! It's a tool for compiling and running CUDA/HIP applications on SPIR-V-supported OpenCL or LevelZero platforms. v1.0 can already run various HPC applications correctly. See: https://github.com/CHIP-SPV/chipStar/releases/tag/v1.0
#opencl #levelzero #spirv #cuda #hip

Another Place #rustICL could be super useful is for the media server users of the world.

RN migrating to #Jellyfin from #plex based on an old discussion with @geerlingguy.

JellyFin seems quite good, works mostly OOTB, with the only issue being the #OpenCL/#Cuda requirement for #HDR #tonemaping.

Since I'm running JF as a docker container on a TrueNAS SCALE box, there's no OOTB OpenCL/CUDA support.

The machine in question is running a 12100 and an rx5600xt, with VAAPI working fine

For those who have waited eagerly for the recording of our #oneapi meetup with @karolherbst - here it is! https://youtu.be/KUze0JbPSy8
#sycl #opencl #rustlang - if you are interested in joining our oneAPI meetup - feel free to subscribe here - https://www.meetup.com/oneapi-community-us/

Next time we will be meeting with Stephano Cetola who will be talking about RISC-V, onAPI, and other things.

The #Haskell multidimensional #data #parallel #array library #Accelerate supports #GPU only via #CUDA (no #OpenCL backend that I could find), and it seems abandoned (doesn't work with GHC newer than 8.10.7, docs mention LLVM 9).

The Haskell multidimensional data parallel array library #Repa doesn't support GPU at all as far as I can tell.

I prefer Accelerate's architecture, but Repa seems more maintained.

Repa seems to allow #unsafe operations:

```
$ ghci -package repa
ghci> :m + Data.Array.Repa
ghci> toList . extract (Z :. 10000000) (Z :. 100) . computeUnboxedS . fromFunction (Z :. (1 :: Int)) $ const (0 :: Int)
[Segmentation fault
$
```

This crash is an out of bounds read, trying to access index 10000000 of an array with one element.

So I started coding a repa-sized wrapper library (I checked Hackage and couldn't find any existing one), that statically ensures things are safe at compilation time. For example, trying to reshape a 3,3 matrix to 5,5 now gives a type error (can't match 9 with 25) instead of a runtime error that the sizes don't match.

I wrote up some notes on two projects inspired by #FAuSt #dsp block diagram algebra progamming language.

One from 2009 (typesafe #Haskell embedding of FAuSt semantics) that I abandoned shortly afterwards.

One vague idea (FAuSt + #GridFlow + #OpenCL = some kind of video processing thing) that I haven't started work on implementing (and might never), just in the what if thinking stage.

https://mathr.co.uk/web/faust.html

My ChEESE CoE webinar talk on #OpenCL #GPU programming for #HPC applications is now uploaded to my YouTube channel and provided with timestamps. Enjoy! 🖖😎🧀
👉 https://youtu.be/w4HEwdpdTns

AMD GPU Issues in Ubuntu 22.04.2 LTS: "amdgpu bonaire not supported" & "snd hda intel hsw/bdw hd audio hdmi/dp requires binding with gfx driver" #boot #amdgraphics #amdgpu #ubuntudrivers #opencl

https://askubuntu.com/q/1475108/612

#Mesa's Rusticl Lands Experimental FP16 #OpenCL Support

https://www.phoronix.com/news/Rusticl-OpenCL-FP16

Original tweet : https://twitter.com/phoronix/status/1673020161081880577

PoCL 4.0 released! Various new features to support more CUDA, HIP and SYCL app inputs via chipStar and DPC++. Also MacOS/CPU support is back and has now a maintainer, and a lot more: https://github.com/pocl/pocl/releases/tag/v4.0
#opencl #HeterogeneousComputing #opensource

New #HPC toy has arrived! AMD Instinct #MI210 #GPU with delicious 64GB VRAM 🖖😋 perfect for #OpenCL workloads 🖖🧐🔥

In 2 weeks I'll be hosting a webinar on #OpenCL #GPU programming at @cheese_coe #HPC Center of Excellence! ✨🧀✨
Register here, it's free: bit.ly/3C3cxey
Looking forward to sharing my knowledge with you! 🖖😎

I'll be speaking during the OneAPI dev summit tomorrow, specifically the panel discussion on accelerated computing.

Partially as a meme/Sanity check the challenge of this afternoon:

Can I got from no installs to a GPU accelerated simulation of OpenFOAM using an Arc a770 on Linux?

Plan is #OpenFoam 2212, #petsc 3.19, #OneAPI 2023.1 and #mesa 23 (in case I need to fall back to #OpenCL)

I am considering writing a test suite for "rusticl" (#rustlang #OpenCL implementation) and I am wondering how I should name my crate... 🤔 😜
@phoronix

Many of you have waited for this eagerly: PoCL 4.0-rc1 is now ready for testing! https://github.com/pocl/pocl/releases/tag/v4.0-RC1 The usual LLVM upgrade, an experimental Level Zero driver, CPU subgroups, prog. scope variables, Unified
Shared Memory, etc. etc! Please test with your favourite platform and OpenCL application!
#opencl @thekhronosgroup