Batmobile: 10-20x Faster CUDA Kernels for Equivariant Graph Neural Networks(elliotarledge.com)83 points by ipnon 9 months ago | 11 comments
Why libvirt supports only 14 PCIe hotplugged devices on x86-64(dottedmag.net)223 points by andreyvit 3 years ago | 53 comments
SIMD < SIMT < SMT: Parallelism in Nvidia GPUs (2011)(yosefk.com)127 points by shipp02 2 years ago | 34 comments