FERRAMENTAS LINUX: GPU Acceleration
Mostrando postagens com marcador GPU Acceleration. Mostrar todas as postagens
Mostrando postagens com marcador GPU Acceleration. Mostrar todas as postagens

quarta-feira, 11 de junho de 2025

DragonFlyBSD’s DRM Graphics Driver Update: Limited Progress for Legacy Hardware

 


DragonFlyBSD updates its Linux-ported DRM graphics driver to support aging AMD Picasso, Vega 20, and Intel Ice Lake GPUs—though still lagging behind Linux 6.16. Learn how this impacts BSD users, hardware compatibility, and open-source GPU innovation.

sexta-feira, 6 de junho de 2025

Rusticl OpenCL Driver Achieves Native FP16 Support in Mesa 25.2, Closing Feature Gap with Clover

 

Mesa

Mesa’s Rusticl OpenCL driver now supports native FP16 (cl_khr_fp16), matching Clover’s capabilities. Tested on Asahi (Apple Silicon), RadeonSI, and more—key for high-performance GPU computing. Learn how this impacts developers & hardware acceleration.

domingo, 1 de junho de 2025

AMD ROCm 7.0 & HIP 7.0: Major CUDA Alignment Coming in 2025

 

AMD

AMD’s ROCm 7.0 update brings HIP 7.0 closer to CUDA, simplifying GPU programming for AI & HPC. Learn about breaking changes, early preview builds, and how to prepare for the 2025 release.

quinta-feira, 29 de maio de 2025

Rusticl OpenCL Driver Achieves Major Milestones: SVM & Intel Subgroups Support

 

Mesa


Rusticl, Mesa's Rust-based OpenCL driver, now supports Intel Subgroups and Shared Virtual Memory (SVM)—boosting performance for AI, machine learning, and high-performance computing (HPC). Learn how these updates impact GPU acceleration & OpenCL 2.0+ development.