Home > Workload Solutions > High Performance Computing > White Papers > Dell Technologies Validated Design for Genomics with NVIDIA Clara Parabricks On AMD-powered Dell PowerEdge > System configuration
Performance benchmarking was performed in the Dell Technologies HPC and AI Innovation Lab using the system configurations that are listed in the table below.
Building block | Components | Quantity |
PowerEdge XE8545 computational server | Dual AMD EPYC 7543 32-Core Processor 512 GB RAM (sixteen 32 GB 3200 MT/sec DIMMs) Four NVIDIA A100-SXM4-40 GB NVIDIA ConnectX-6 HDR100 adapter | 1 |
PowerEdge R7525 computational server | Dual AMD EPYC 7543 32-Core Processor 512 GB RAM (sixteen 32 GB 3200 MT/sec DIMMs) Three NVIDIA A100-80 GB NVIDIA ConnectX-6 HDR100 adapter | 1 |
Network switch | NVIDIA QM8790 InfiniBand switch | 1 |
The BIOS configuration options that were used for the reference system are listed the table below.
BIOS option | Setting |
Logical Processor | Disabled |
Virtualization Technology | Disabled |
System Profile | Performance Profile |
NumaPerSocket | 4 |
The software versions that were used for the reference system are listed in the table below.
Component | Version |
Operating System | Red Hat Enterprise Linux release 8.3 (Ootpa) |
Kernel | 4.18.0-240.22.1.el8_3.x86_64 |
OFED | NVIDIA Mellanox 5.2-2.2.0.0 |
Bright Cluster Manager | 9.0 |
CUDA | 11.4.2 |
NVIDIA driver | 470.57.02 |
NVIDIA Clara Parabricks | 3.7.0-1 |