Dynamic Register Allocation on AMD's RDNA 4 GPU Architecture
https://chipsandcheese.com/p/dynamic-register-allocation-on-amds

Dynamic Register Allocation on AMD's RDNA 4 GPU Architecture
https://chipsandcheese.com/p/dynamic-register-allocation-on-amds
Today, we’re super excited to announce our latest product addition: Continuous Profiling for GPUs! Check out the use cases and sign up for early access on the announcement post!
https://www.polarsignals.com/blog/posts/2025/04/01/introducing-continuous-gpu-profiling
In honor of Trans Day of Visibility, here’s something really cool for you to know…
A computer scientist trans woman by the name of Sophie Mary Wilson was a co-creator of the ARM architecture.
More about Sophie:
https://en.wikipedia.org/wiki/Sophie_Wilson
Great video on the history of ARM, which is computer science geeky stuff (sorry about YT):
https://www.youtube.com/watch?v=nIwdhPOVOUk
Looking forward to learning more about this libre-licensed RISC-V SoC with Kazan GPU and VPU.
https://www.crowdsupply.com/libre-risc-v/m-class
I'm really curious how these types of chips are prototyped. I know we can simulate a few hundred thousand logical operations with an FPGA, but is that even close to simulating a powerful chip of this size?
CoffeeLoader il malware che evade le difese sfruttando la GPU
https://gomoot.com/coffeeloader-il-malware-che-evade-le-difese-sfruttando-la-gpu/
I made this #FluidX3D #CFD simulation run on a frankenstein zoo of AMD +
Nvidia +
Intel #GPUs!
https://www.youtube.com/watch?v=_8Ed8ET9gBU
The ultimate SLI abomination setup:
- 1x Nvidia A100 40GB
- 1x Nvidia Tesla P100 16GB
- 2x Nvidia A2 15GB
- 3x AMD Instinct MI50
- 1x Intel Arc A770 16GB
I split the 2.5B cells in 9 domains of 15GB - A100 takes 2 domains, the other GPUs 1 domain each. The GPUs communicate over PCIe via #OpenCL.
Huge thanks to Tobias Ribizel from TUM for the hardware!
If you're having screen corruption or GPU crashes on an AMD Framework laptop, it seems there's now a solution.
https://community.frame.work/t/amdgpu-instability-6-13-4-firmware-20250219/65312/35?u=tripplehelix
Removing the flag `amdgpu.sg_display=0` from your kernel parameters seems to solve it. I added it to mine originally due to having this issue. Seems to have flipped.
Does anyone know what Vulkan dynamic rendering + multiview into a cubemap looks like? I can't find any examples anywhere and what I have (just setting VkRenderingInfo::viewMask) doesn't seem to be working
Boosts for visibility appreciated!
#gamedev #graphics #3dgraphics #vulkan #gpu #rendering
I got access to @LRZ_DE's new coma-cluster for #OpenCL benchmarking and experimentation
I've added a ton of new #FluidX3D #CFD #GPU/#CPU benchmarks:
https://github.com/ProjectPhysX/FluidX3D?tab=readme-ov-file#single-gpucpu-benchmarks
Notable hardware configurations include:
- 4x H100 NVL 94GB
- 2x Nvidia L40S 48GB
- 2x Nvidia A2 15GB datacenter toaster
- 2x Intel Arc A770 16GB
- AMD+Nvidia SLI abomination consisting of 3x Instinct MI50 32GB + 1x A100 40GB
- AMD Radeon 8060S (chonky Ryzen AI Max+ 395 iGPU with quad-channel RAM) thanks to @cheese
NEWS FOR #HAIKU: #NVIDIA GPU support coming soon!
Developer @X512 has successfully ported Nvidia kernel drivers to Haiku. The driver will support Turing+ GPUs and already includes Vulkan integration via Mesa's NVK.
Initial tests are working and show potential for future uses, including AI acceleration with llama.cpp.
A major step forward for the Haiku ecosystem and hardware compatibility!
Looking Ahead at Intel's Xe3 GPU Architecture
https://chipsandcheese.com/p/looking-ahead-at-intels-xe3-gpu-architecture
"#Nvidia's Vera Rubin CPU, #GPU roadmap charts course for hot-hot-hot 600 kW racks"
"Now that's what we call dense floating-point compute"
https://www.theregister.com/2025/03/19/nvidia_charts_course_for_600kw/
#ai #datacenter #ResourceConsumption out of control
Asahi Lina: I no longer feel safe working on Linux GPU drivers or Linux graphics
Akira ransomware's encryption cracked!
Security researcher Yohanes Nugroho decrypted files without paying, using cloud GPU power.
He shared the decryption tool on GitHub. A flaw in Akira’s encryption allowed timestamps to be correlated. More info: https://cyberinsider.com/akira-ransomware-encryption-cracked-using-cloud-gpu-power/ #ransomware #cybersecurity #datarecovery #GPU #newz
For all wondering what I've been doing at #Intel this entire time: I wrote and optimized big parts of the #GPU kernels for #XeSS Super Resolution and Frame Generation. XeSS 2 SDK is finally out now!
https://github.com/intel/xess/releases/tag/v2.0.1
If you want to tweak your GPU on Linux, the Linux GPU Configuration Tool (LACT) is about as easy as it gets.