Alain Galvan

@[email protected]
337 Followers
145 Following
38 Posts
Driver Engineer @ AMD! Prev. @ https://Marmoset.co! Love writing about 🔺 Computer Graphics and 🎨 Art. 👋 My views only. DMs open!
Sitehttps://alain.xyz
Githubhttps://github.com/alaingalvan
LinkedInhttps://linkedin.com/in/alaingalvan
YouTubehttps://www.youtube.com/c/AlainGalvan
Alain GalvanSep 30, 2024
Managed to get a library I wrote compiled and running on Windows 11 ARM64. Microsoft discourages installing Visual Studio 2022 + Build Tools with a dismissible popup that says ARM64 isn't supported, but otherwise Visual Studio, C++ apps, and Rust work fine.
Show threadAlain GalvanSep 8, 2024
@mehmetoguzderin So I can run D3D12 on Windows Subsystem for Linux, but I haven't tried running Docker with it. Do you have any suggestions for GPU virtualization tests? I could try Pytorch w/ DirectML: https://github.com/microsoft/DirectML/tree/master?tab=readme-ov-file#pytorch-with-DirectML
GitHub - microsoft/DirectML: DirectML is a high-performance, hardware-accelerated DirectX 12 library for machine learning. DirectML provides GPU acceleration for common machine learning tasks across a broad range of supported hardware and drivers, including all DirectX 12-capable GPUs from vendors such as AMD, Intel, NVIDIA, and Qualcomm.

DirectML is a high-performance, hardware-accelerated DirectX 12 library for machine learning. DirectML provides GPU acceleration for common machine learning tasks across a broad range of supported ...

GitHub
Show threadAlain GalvanSep 8, 2024
@kanashimia The interactive parts (like in the BVH post) were made using React components rendering Canvas (was debating using canvas or just SVGs but thought canvas might perform better on mobile) rendered with MDX: https://mdxjs.com/
Markdown for the component era | MDX

MDX lets you use JSX in your markdown content. You can import components, such as interactive charts or alerts, and embed them within your content. This makes writing long-form content with components a blast.

MDX
Alain GalvanSep 1, 2024
I recently bought one of the new Qualcomm Snapdragon X Elite machines, the Samsung Galaxy Book4 Edge. The battery is similar to other arm devices like your phone, but is running Windows and can run most apps natively or with x86 emulation. Love the battery life and low heat.
Alain GalvanAug 4, 2024
I was in High Performance Graphics 2024 in Denver, CO last weekend and the experience was inspiring and enlightening. Seeing distinguished speakers share impressive papers and posters such as H-PLOC, DGF, and neural denoisers with such a depth of knowledge and understanding in this domain made the trip worth it. https://gpuopen.com/advanced-rendering-research/
Advanced Rendering Research Group

Find out more about the work of our Advanced Rendering Research Group, including projects, publications, and more.

AMD GPUOpen
Show threadAlain GalvanJul 27, 2024
@AlexSneezeKing Hey Alex! Good to see you again after i3D!
Alain GalvanJul 24, 2024
Christopher BattyOct 19, 2023
For computer graphics academics and researchers, I threw this list of folks together. Apologies if I've forgotten anyone or included anyone who prefers to be forgotten! Let me know if there are any errors or corrections. Hoping this will help folks migrating from the birdsite. 🙂
https://christopherbatty.github.io/Mastodon-Graphics-Academics/
Graphics Academics on Mastodon

A list of Graphics Academics on Mastodon

Alain GalvanJul 11, 2024
I'll be in Denver, CO for High Performance Graphics (HPG) this year, if you're headed there feel free to say hi! 👋
Alain GalvanJul 1, 2024
In an attempt to speed up and automate publishing to Facebook/X/Mastodon/Threads/Bluesky/whatever, I've been working on making my site compliant with the activitystream 2.0 spec. Indieweb is a big inspiration: https://aaronparecki.com/2012/09/09/6/indieweb
Demo post (for @indiewebcampUK) -- demonstrating my posting UI. ...

Demo post (for @indiewebcampUK) -- demonstrating my posting UI. After the text gets too long it truncates automatically and adds a citation link. Hashtags are preserved. #indieweb

Aaron Parecki
Alain GalvanMay 4, 2024

I was experimenting with Phi-3, the small language model open sourced by Microsoft that uses ONNX and DirectML as the runtime for the model, and found it useful for performing 'image editor' like operations on text.

Like an image being rescaled, adjusted with curves or hue-saturation-lightness, or sharpened with edge detection, our sentences can also be adjusted by these models making them easier to understand, reword, and thus communicate our intent more precisely.

https://onnxruntime.ai/blogs/accelerating-phi-3

ONNX Runtime | Blogs/accelerating-phi-3

Cross-platform accelerated machine learning. Built-in optimizations speed up training and inferencing with your existing technology stack.