oblivus logo

Oblivus

H100x4

| 320GB VRAMs | H100_PCIE_80GB_x4

Explore Oblivus's H100 cloud instance specifications and benchmarks. Compare hardware configurations and performance metrics to optimize your AI and ML workloads.

LLM Benchmark Comparison

Hardware Specifications

GPU ConfigurationValue
GPU TypeH100
GPU InterconnectPCIE
GPU Model NameNVIDIA H100 PCIe
Driver Version535.104.12
GPU VRAM320
Power Limit (W)350.00
GPU Temperature (°C)33
GPU Clock Speed (MHz)345
Memory Clock Speed (MHz)1593
PstateP0
CPU ConfigurationValue
Model NameAMD EPYC 9554 64-Core Processor
Vendor IDAuthenticAMD
CPUs124
CPU Clock Speed6190.69
Threads Per Core1
Cores Per Socket62
Sockets2
MemoryValue
Total708Gb

Disks Specifications

StorageValue
Total3274.40GB

Available Disks

PropertyValue
Disk 1
Modelvda
Size100Gb
TypeHDD
Mount PointUnmounted
Disk 2
Modelvdb
Size3.1Tb
TypeHDD
Mount Point/mnt

Software Specifications

SoftwareValue
OSUbuntu
OS Version22.04.3 LTS (Jammy Jellyfish)
Cuda Driver12.2
Docker Version24.0.6
Python VersionPython 3.10.12

Benchmarks

BenchmarkValue
ffmpeg74
Coremark (Itterations per sec)36701.737
llama2Inference (Tokens per sec)60.86
Tensorflow Mnist Training1.660

Nvidia-smi output

Nvidia-smi topo -m outpu

Launch instance

CloudOblivus
GPU TypeH100
Shadeform Instance TypeH100x4
Cloud Instance TypeH100 PCIE 80GB x4
Spin Up Time5-10 mins
Hourly Price$7.92

By clicking launch you agree to our Terms of Service.

Feedback

© Shadeform, Inc