From: RUBICON: a framework for designing efficient deep learning-based genomic basecallers
CPU | AMD EPYC 7742Â [91] |
 | @2.25GHz, 4-way SMT [92] |
Cache-Hierarchy | 32\(\times\)32 KiB L1-I/D, 512 KiB L2, 256 MiB L3 |
System memory | 4\(\times\)32GiB RDIMM DDR4 2666 MHz [93] PCIe 4.0 \(\times\)128 |
OS details | Ubuntu 21.04 Hirsute Hippo [94], GNU Compiler Collection (GCC) version 10.3.0 [95] |
GPU | AMD Radeon Instinctâ„¢ MI210Â [68] 6656 Stream Processors@1.7GHz 64GB HBM2 PCIe 4.0 \(\times\)16, ROCm version 5.1.1Â [96] NVIDIA A40Â [69] 10,752 CUDA Cores@1.2GHz, 48GiB DRAM NVIDIA System Management Interface (NVIDIA-SMI) version 510.47.03Â [97] NVIDIA CUDA Compiler Driver (NVCC) version 11.4Â [98] |
AMD-Xilinx Versal ACAP | Versal ACAP VC2802Â [64], 304\(\times\)AIE-ML@1GHz, 19MB local memory, Dual-Core Arm Cortex-A72 [99] |