boksajak Profile Banner
Jakub Boksansky Profile
Jakub Boksansky

@boksajak

Followers
3K
Following
6K
Media
161
Statuses
2K

Tracing rays one by one. Caffeinist and Ray Tracing Engineer at AMD. Former Computer Graphics Researcher @CVUTFEL, ex-Nvidia. Personal account.

Munich, Germany
Joined January 2015
Don't wanna be here? Send us removal request.
@boksajak
Jakub Boksansky
6 years
Merry raytracing Christmas everybody!
1
83
475
@boksajak
Jakub Boksansky
19 days
I am happy to announce publication of my chapter titled "Real-Time Neural Network Implementation for GPUs" in GPU Zen 4 book, available now. Chapter comes with full MIT-licensed source code, available on Github here: https://t.co/mmoVTA6LlG The book is available on Amazon.
3
24
205
@boksajak
Jakub Boksansky
3 months
Great work! Check out the ODIN port of my Dx12NN MLP imolementation by Lukasz:
@_woookie_
Łukasz Bogaczyński
3 months
After reading through @boksajak's great https://t.co/HFsYre4yPk, I implemented it in Odin. It's super slower as you might expect, but also super simple, so it's good base if you want to learn the basics (or SIMD/multithread further). https://t.co/zNWyqxeAMk
0
2
21
@GPUOpen
AMD GPUOpen
4 months
Capsaicin is not just for tacos, it's also our developer rendering framework! v1.3 is out now: 🌶️ Added heat with GI-1.2, which includes multiple-bounce indirect lighting. 🌶️ HDR support with spicier tone mapping. And much, much more:
Tweet card summary image
gpuopen.com
AMD GI-1.2, our real-time Global Illumination solution, is available now as part of the AMD Capsaicin Framework v1.3.
3
27
103
@GPUOpen
AMD GPUOpen
5 months
Get up to speed with our latest advanced graphics research: 📝 GATE (Geometry-Aware Trained Encoding) for neural rendering. 📝 Real-time procedural GPU tree generation. 📝 Neural Visibility Cache for real-time light sampling. 📥 Download the papers: https://t.co/y0XU5hbQXV
1
33
138
@haradatakahiro
takahiro harada
7 months
#pytorch started running on windows (!) on #AMD #GPUs. ## Set up - python -m pip install --index-url https://t.co/ydkIKlKXgU torch ## Then >>> import torch >>> print( https://t.co/Bl5V7mfV40_available()) True >>> print(torch.cuda.get_device_name(0)) AMD #Radeon RX 9070 XT
1
3
9
@boksajak
Jakub Boksansky
1 year
Are you guys enabling HLSL 2021 when compiling your shaders (e.g. for templates)? Is there anything preventing you from enabling HLSL 2021 by default?
3
1
12
@ShinjiOgaki
Shinji Ogaki
1 year
Unidirectional Path Tracing Light Sampling MIS🙂 The next step is to support more complex light paths and lenses.
@ShinjiOgaki
Shinji Ogaki
1 year
Fixed minor mistakes. The next step is of course MIS😀
1
8
83
@I3DCONF
I3D Symposium
1 year
The I3D 2025 Call for Papers has been published! Submission deadline: 10 of January, 2025 Get more information in our website: https://t.co/ZXmlE1WJCa #I3D2025
Tweet card summary image
i3dsymposium.org
ACM SIGGRAPH Symposium on Interactive 3D Graphics and Games
0
8
13
@boksajak
Jakub Boksansky
1 year
I put all of my Ray Tracing Gems chapters in one place where you can download them, extracted from books into separate PDFs: https://t.co/AOGFePRgoG There is: 1. "Ray Traced Shadows" 2. "Reference Path Tracer" 3. "Refraction Ray Cones" 4. "Many Lights With Grid-based Reservoirs"
4
76
315
@boksajak
Jakub Boksansky
1 year
My latest blog is now out! It discusses two GPU-friendly methods for sampling of Gaussian distribution, including source code and performance evaluations. Get it here:  https://t.co/gb0TCqAljE
Tweet card summary image
gpuopen.com
Sampling the normal distribution is not as straightforward as it seems! Explore two GPU-friendly methods including implementation and performance evaluation.
3
24
147
@GPUOpen
AMD GPUOpen
1 year
📉 Sampling the normal (Gaussian) distribution is not as easy as it seems! 📈 Our latest blog by @boksajak introduces two GPU-friendly methods, including implementation, source code, and a perf evaluation. 📖 Have a read on GPUOpen: https://t.co/Mc7QmuFI1X #gamedev #raytracing
Tweet card summary image
gpuopen.com
Sampling the normal distribution is not as straightforward as it seems! Explore two GPU-friendly methods including implementation and performance evaluation.
0
17
48
@GPUOpen
AMD GPUOpen
1 year
🎇 DMMs (displaced micro-meshes) deliver excellent geometry compression but don't represent animated geometry well 🎇 Our paper explores producing high-quality animated DMMs, showing how to preserve their low memory footprint, BVH build, and update costs: https://t.co/PQqjnFPPkA
0
43
150
@krupitskas
Nikita Krupitskas
1 year
Added support for RTX BLAS + TLAS generation and runtime Raster/Pathtracer switch, currently showing barycentrics. Next step add bindless textures and actually show "gbuffer" from the pathtracer view. Ray Tracing Gems 2, ch 14 is a REAL GEM😎@boksajak @acmarrs
2
3
63
@boksajak
Jakub Boksansky
2 years
What do ray tracing people do on vacation? Shoot rays towards the film 📸
1
0
36
@antonschrein
Anton Schreiner
2 years
My team is hiring
0
3
5
@theGioM
Guillaume H.
2 years
Some Chinese dudes imitating AI videos lol this is next level
470
14K
138K
@StefanPijnacker
Stefan Pijnacker
2 years
Hey there! I started a new article series on how to write compute shaders with DirectX 12. The first chapter discusses the theory of compute which is mostly applicable for any Graphics API or engine. https://t.co/0HQixLz8Yw
Tweet card summary image
stefanpijnacker.nl
We will discuss the general theory of parallel processing and compute shaders. We will talk about what compute shaders are, why they can be so powerful and some things to keep in mind while working...
4
29
156
@boksajak
Jakub Boksansky
2 years
Siggraph, here we come. See you all tomorrow!
0
0
24
@boksajak
Jakub Boksansky
2 years
My deep learning crash course is now available as a HTML web version on @GPUOpen. Enjoy!
@GPUOpen
AMD GPUOpen
2 years
If you're looking to learn #deeplearning in a #graphics dev context, @boksajak has written an excellent crash course which we're excited to share on GPUOpen 🎉 You'll learn about #neuralnetworks, practical graphics applications, + links to sample code! https://t.co/HQ09yUA7zD
1
6
63