Lightnews — Scholar-powered news

François Guthmann

@frguthmann.bsky.social

1K followers 300 following 28 posts

Your friendly neighborhood Guthmann | Computer graphics enthusiast | GPU Dev Tech at AMD. Opinions are my own. http://frguthmann.github.io

frguthmann.github.io

Posts Media Videos Starter Packs

Pinned

François Guthmann @frguthmann.bsky.social · Dec 4

The recordings from GPC24 are up, including my talk ! This contains much more information than the slides on their own. Enjoy !
www.youtube.com/watch?v=sHFb...
gpuopen.com/presentation...

Occupancy Explained Through the AMD RDNA™ Architecture

YouTube video by Graphics Programming Conference

www.youtube.com

23 76

François Guthmann @frguthmann.bsky.social · Jun 18

We have multiple open job opportunities in game engineering, come work with us !

Anton Schreiner @aschrein.bsky.social · Jun 18

Sharing this job opportunity at my team: development and productization of ML+Graphics technologies for games. EU remote. careers.amd.com/careers-home...

GPU Software Development Engineer in Warsaw, Poland | Advanced Micro Devices, Inc

AMD | Careers Home is hiring a GPU Software Development Engineer in Warsaw, Poland. Review all of the job details and apply today!

careers.amd.com

3 7

Reposted by François Guthmann

Marco Castorina @marcocastorina.bsky.social · Mar 14

As a reviewer, I got a few codes for a free copy of "Vulkan 3D Graphics Rendering Cookbook". PM me if you'd like one, first come, first served :)

2 2

François Guthmann @frguthmann.bsky.social · Mar 12

Democratize GPU profiling on the web, let's go !

Brandon Jones @toji.dev · Mar 11

I've written a brief guide for profiling WebGPU commands with XCode's Metal debugger. It's... not exactly easy, but it works! I've got several ideas for how to improve it in the future, too.

toji.dev/webgpu-profi...

Profiling WebGPU with XCode’s Metal Debugger

Brandon Jones - Graphics and XR on the web

toji.dev

François Guthmann @frguthmann.bsky.social · Feb 13

Seb wrote a fantastic article exploring how to make matrix multiplications go fast on AMD GPUs. This is a deep dive into both the algorithm and hardware, with great visualizations. He is also looking for work and I only have good things to say about him, so hire him!
seb-v.github.io/optimization...

Optimizing Matrix Multiplication on RDNA3: 50 TFlops and 60% Faster Than rocBLAS

Introduction

seb-v.github.io

1 7 30

Reposted by François Guthmann

Amélie Heinrich @beachedxo.bsky.social · Jan 8

Hey there bsky!

I'm looking for an internship in graphics/engine programming anywhere in the world (I'm willing to relocate). If you happen to know any open positions or if you work at a company that is taking interns, feel free to contact me!

Portfolio ⬇️

1 21 36

François Guthmann @frguthmann.bsky.social · Dec 21

Oh man this is so much better.

François Guthmann @frguthmann.bsky.social · Dec 21

Thanks! Want to grab a beer at some point after the holidays if you're still in Paris btw?

1 4

François Guthmann @frguthmann.bsky.social · Dec 16

Thank you very much 😁

François Guthmann @frguthmann.bsky.social · Dec 4

The recordings from GPC24 are up, including my talk ! This contains much more information than the slides on their own. Enjoy !
www.youtube.com/watch?v=sHFb...
gpuopen.com/presentation...

Occupancy Explained Through the AMD RDNA™ Architecture

YouTube video by Graphics Programming Conference

www.youtube.com

23 76

François Guthmann @frguthmann.bsky.social · Dec 3

Oh I didn't know it had Tint and Naga now, nice! Thanks, TIL!

François Guthmann @frguthmann.bsky.social · Dec 3

I'm not sure if there is something, I use PIX on windows to get the HLSL.
It's ok if you don't have an easy way to get the code, I won't be able to test on Mac anyway.

François Guthmann @frguthmann.bsky.social · Dec 3

😱😱😱

François Guthmann @frguthmann.bsky.social · Dec 2

I know the explanation is short but it would take quite a few tweets to fully go through it 😅

François Guthmann @frguthmann.bsky.social · Dec 2

V0 is worse than original for testing, V1 is the original code, V2 is a failed fix, V3 fixes the execution sync issue but lacks a read after write synchronization. It seems the writes make it to memory fast enough though!

The code is here if you want to take a look github.com/frguthmann/w...

GitHub - frguthmann/webgpu_racy_tests

Contribute to frguthmann/webgpu_racy_tests development by creating an account on GitHub.

github.com

1 2

François Guthmann @frguthmann.bsky.social · Dec 2

I was looking at some code synchronizing multiple threadgroups within the same dispatch and I had a hunch something was off. There are actually multiple causes for the original race(s) so I wrote increasingly less racy versions of it, to validate my hunch and finally fix the original code.

1 2

François Guthmann @frguthmann.bsky.social · Dec 2

Thanks! What did you test this on?

François Guthmann @frguthmann.bsky.social · Dec 2

Oh wow, interesting! I only inspected the generated HLSL and AMD ISA but I would have expected this to work on Apple devices as well, thanks for testing!
Would it be easy for you to send me the generated MSL code?

François Guthmann @frguthmann.bsky.social · Dec 2

Here's what I get:
V0: ❌, V1: ❌, V2: ❌, V3: ✅, V4: ✅

François Guthmann @frguthmann.bsky.social · Dec 2

Hey all, I wrote some WebGPU code to test some race conditions in a shader. If you're willing to try, let me know what you see. Red/flicker means there's a race.
I'm especially curious about V3 which could theoretically race but I was never able to catch it.
frguthmann.github.io/webgpu_racy_...

Racy WebGPU tests

frguthmann.github.io

5 6 16

Reposted by François Guthmann

Anton Schreiner @aschrein.bsky.social · Dec 1

My team is hiring careers.amd.com/careers-home...

Senior GPU/ML Modelling Engineer in Warsaw, Poland | Advanced Micro Devices, Inc

AMD | Careers Home is hiring a Senior GPU/ML Modelling Engineer in Warsaw, Poland. Review all of the job details and apply today!

careers.amd.com

5 6

François Guthmann @frguthmann.bsky.social · Nov 17

Yes, definitely! When rehearsing I always tried to phrase occupancy as *one* way to hide latency, but not sure how I did when presenting that day.

François Guthmann @frguthmann.bsky.social · Nov 16

A fun night 😁

François Guthmann @frguthmann.bsky.social · Nov 16

Sorry for the terrible photo quality, knees were weak and arms heavy :p.
Also, shout-out to the first row who noticed I was taking a picture and decided to have fun with it 😁.

François Guthmann @frguthmann.bsky.social · Nov 10

I would love a write up explaining the design choices, pitfalls avoided and so on afterwards!

1 2

François Guthmann @frguthmann.bsky.social · Oct 31

Nice, see you there!