a100-clpeak
2 x AMD EPYC 7742 64-Core testing with a NVIDIA DGXA100 v555.06901.0004 (0.34 BIOS) and ASPEED 40GB on Ubuntu 20.04 via the Phoronix Test Suite.


2 x AMD EPYC 7742 64-Core - ASPEED 40GB - NVIDIA: 

	Processor: 2 x AMD EPYC 7742 64-Core @ 2.25GHz (128 Cores / 256 Threads), Motherboard: NVIDIA DGXA100 v555.06901.0004 (0.34 BIOS), Chipset: AMD Starship/Matisse, Memory: 16 x 64 GB DDR4-3200MT/s 36ASF8G72PZ-3G2B2, Disk: 4 x 3841GB SAMSUNG MZWLJ3T8HBLS-00007 + 2 x 1920GB SAMSUNG MZ1LB1T9HALS-00007, Graphics: ASPEED 40GB, Network: 2 x Intel 82599ES 10-Gigabit SFI/SFP+ + 3 x Mellanox MT28908 + Intel I210

	OS: Ubuntu 20.04, Kernel: 5.4.0-80-generic (x86_64), Display Server: X Server, Display Driver: NVIDIA, OpenCL: OpenCL 1.2 CUDA 11.2.109, Compiler: GCC 9.3.0 + CUDA 11.2, File-System: ext4, Screen Resolution: 800x600


clpeak
OpenCL Test: Kernel Latency
us < Lower Is Better
2 x AMD EPYC 7742 64-Core - ASPEED 40GB - NVIDIA . 11.26 |=====================


clpeak
OpenCL Test: Integer Compute INT
GIOPS > Higher Is Better
2 x AMD EPYC 7742 64-Core - ASPEED 40GB - NVIDIA . 16758.84 |==================


clpeak
OpenCL Test: Single-Precision Float
GFLOPS > Higher Is Better
2 x AMD EPYC 7742 64-Core - ASPEED 40GB - NVIDIA . 16078.58 |==================


clpeak
OpenCL Test: Double-Precision Double
GFLOPS > Higher Is Better
2 x AMD EPYC 7742 64-Core - ASPEED 40GB - NVIDIA . 8912.24 |===================


clpeak
OpenCL Test: Global Memory Bandwidth
GBPS > Higher Is Better
2 x AMD EPYC 7742 64-Core - ASPEED 40GB - NVIDIA . 1287.13 |===================


clpeak
OpenCL Test: Transfer Bandwidth enqueueReadBuffer
GBPS > Higher Is Better
2 x AMD EPYC 7742 64-Core - ASPEED 40GB - NVIDIA . 9.81 |======================


clpeak
OpenCL Test: Transfer Bandwidth enqueueWriteBuffer
GBPS > Higher Is Better
2 x AMD EPYC 7742 64-Core - ASPEED 40GB - NVIDIA . 18.50 |=====================