SMCI Server, BIOS Power Profile Testing

AMD EPYC 9655P 96-Core testing with a Supermicro Super Server H13SSL-N v1.01 (3.0 BIOS) of the various power profile options. Each BIOS power profile option was tested with both acpi-cpufreq schedutil and acpi-cpufreq performance governor configurations within Linux. A wide variety of workloads tested. Benchmarks by Michael Larabel for a future article.

HTML result view exported from: https://openbenchmarking.org/result/2411285-NE-SMCISERVE81&rdt&grr.

PETSc

Test: Streams

NWChem

Input: C240 Buckyball

MariaDB

Test: oltp_read_write - Threads: 128

MariaDB

Test: oltp_update_index - Threads: 128

Xcompact3d Incompact3d

Input: X3D-benchmarking input.i3d

MariaDB

Test: oltp_read_only - Threads: 128

High Performance Conjugate Gradient

X Y Z: 144 144 144 - RT: 60

Palabos

Grid Size: 500

XNNPACK

Model: QS8MobileNetV2

XNNPACK

Model: FP16MobileNetV3Small

XNNPACK

Model: FP16MobileNetV3Large

XNNPACK

Model: FP16MobileNetV2

XNNPACK

Model: FP16MobileNetV1

XNNPACK

Model: FP32MobileNetV3Small

XNNPACK

Model: FP32MobileNetV3Large

XNNPACK

Model: FP32MobileNetV2

XNNPACK

Model: FP32MobileNetV1

LiteRT

Model: NASNet Mobile

TensorFlow

Device: CPU - Batch Size: 512 - Model: ResNet-50

LuxCoreRender

Scene: LuxCore Benchmark - Acceleration: CPU

LiteRT

Model: Quantized COCO SSD MobileNet v1

LuxCoreRender

Scene: Danish Mood - Acceleration: CPU

LiteRT

Model: Inception ResNet V2

LiteRT

Model: Mobilenet Quant

Epoch

Epoch3D Deck: Cone

Quicksilver

Input: CORAL2 P2

Palabos

Grid Size: 400

LiteRT

Model: DeepLab V3

LuxCoreRender

Scene: Orange Juice - Acceleration: CPU

Timed Linux Kernel Compilation

Build: allmodconfig

Apache IoTDB

Device Count: 800 - Batch Size Per Write: 100 - Sensor Count: 800 - Client Number: 400

Apache IoTDB

Device Count: 800 - Batch Size Per Write: 100 - Sensor Count: 800 - Client Number: 400

DuckDB

Benchmark: IMDB

SVT-AV1

Encoder Mode: Preset 3 - Input: Bosphorus 4K

High Performance Conjugate Gradient

X Y Z: 104 104 104 - RT: 60

Timed LLVM Compilation

Build System: Unix Makefiles

DuckDB

Benchmark: TPC-H Parquet

Apache IoTDB

Device Count: 800 - Batch Size Per Write: 100 - Sensor Count: 800 - Client Number: 100

Apache IoTDB

Device Count: 800 - Batch Size Per Write: 100 - Sensor Count: 800 - Client Number: 100

RocksDB

Test: Read While Writing

LAMMPS Molecular Dynamics Simulator

Model: 20k Atoms

Timed Gem5 Compilation

Time To Compile

PostgreSQL

Scaling Factor: 100 - Clients: 1000 - Mode: Read Only - Average Latency

PostgreSQL

Scaling Factor: 100 - Clients: 1000 - Mode: Read Only

PostgreSQL

Scaling Factor: 100 - Clients: 1000 - Mode: Read Write - Average Latency

PostgreSQL

Scaling Factor: 100 - Clients: 1000 - Mode: Read Write

Apache IoTDB

Device Count: 800 - Batch Size Per Write: 100 - Sensor Count: 500 - Client Number: 400

Apache IoTDB

Device Count: 800 - Batch Size Per Write: 100 - Sensor Count: 500 - Client Number: 400

Graph500

Scale: 26

Graph500

Scale: 26

Graph500

Scale: 26

Graph500

Scale: 26

PostgreSQL

Scaling Factor: 100 - Clients: 800 - Mode: Read Write - Average Latency

PostgreSQL

Scaling Factor: 100 - Clients: 800 - Mode: Read Write

PostgreSQL

Scaling Factor: 100 - Clients: 800 - Mode: Read Only - Average Latency

PostgreSQL

Scaling Factor: 100 - Clients: 800 - Mode: Read Only

Apache IoTDB

Device Count: 800 - Batch Size Per Write: 100 - Sensor Count: 500 - Client Number: 100

Apache IoTDB

Device Count: 800 - Batch Size Per Write: 100 - Sensor Count: 500 - Client Number: 100

PyTorch

Device: CPU - Batch Size: 512 - Model: ResNet-152

Apache Cassandra

Test: Writes

Blender

Blend File: Barbershop - Compute: CPU-Only

Timed Node.js Compilation

Time To Compile