ROCm
PyTorch 2.15 makes rewritten NCCL backend the default and adds FP8 on Apple Silicon
PyTorch 2.15 is out, and the rewritten NCCL backend now runs every nccl process group by default. The release also adds one-sided transports to torch.distributed, finer…
PyTorch adds FlyDSL backend for faster operators on AMD MI350 GPUs
PyTorch users on AMD MI350-series GPUs can now use FlyDSL kernels through existing operator APIs for GEMM, attention, RMSNorm and TopK. According to the PyTorch blog…
AMD Shows YOLO on Radeon GPUs and ROCm at Ultralytics YOLO Vision 2026
AMD, title sponsor of Ultralytics YOLO Vision 2026 (YV26), used the event to show how Ultralytics YOLO models can be trained and deployed on AMD GPUs with PyTorch and…