gfx1151
PulseAugur coverage of gfx1151 — every cluster mentioning gfx1151 across labs, papers, and developer communities, ranked by signal.
2 day(s) with sentiment data
-
Muse Glimmer 30B benchmarked on AMD Strix Halo platform
The author conducted a week-long performance benchmark of the Muse Glimmer 30B model running on AMD's Strix Halo platform, specifically the Ryzen AI Max+ 395 chip. The tests, performed on Linux with ROCm, explored vario…
-
llama.cpp optimizes RDNA3.5 GPU performance with new tiling strategy
The llama.cpp project has released an update, b11618, which optimizes performance for RDNA3.5 GPUs. This update introduces 64-row MMQ tiles for specific quant types, aiming to improve processing speeds on hardware like …
-
ROCm 7.10 update resolves stability issues for specific AMD GPUs
Users on specific AMD GPU architectures (gfx1100, gfx950, gfx1151) are advised to upgrade to ROCm 7.10. This update resolves issues such as dynamic VRAM not producing NaN errors when offloading to RAM, and allows for th…
-
Stable Diffusion Forge WebUI runs on AMD Strix Halo via ROCm 7.2
A user has created a Docker image that enables Stable Diffusion Forge WebUI to run on AMD's Strix Halo (gfx1151) hardware. This setup utilizes the official ROCm 7.2 PyTorch build, offering an alternative to community-de…
-
Ideogram 4 LoRA training achieved on AMD Strix Halo with ROCm
A user successfully trained an Ideogram 4 face LoRA on an AMD Strix Halo APU using ROCm and the AI-Toolkit. The process involved several AMD-specific challenges, including the incompatibility of bitsandbytes, issues wit…
-
MoonMath AI open-sources AMD MI300X attention kernel outperforming AITER v3 · 3 sources tracked
MoonMath AI has released an open-source HIP attention kernel for AMD's MI300X GPU, which reportedly outperforms AMD's own AITER v3. The kernel achieves speedups of up to 1.26x by optimizing memory placement and using on…
-
Ideogram 4 LoRA training detailed for AMD hardware and style examples
Users are sharing their experiences and results training LoRAs for the Ideogram 4 model, a diffusion model praised for its open-source capabilities. One user detailed the process of training a face LoRA on an AMD Strix …