QuData

Rent A100X by the Hour by QuData Ai

Explore the specifications and benchmarks of A100X. Compare hardware configurations and performance metrics for extreme AI and HPC workloads with the enhanced Ampere architecture.

GPU Description

Explore the specifications and benchmarks of A100X. Compare hardware configurations and performance metrics for extreme AI and HPC workloads with the enhanced Ampere architecture.

GPU Characteristics

GPU typeA100X
GPU connectionPCIe
GPU model nameNVIDIA A100X
Driver version535.183.01
GPU memory80GB HBM2e
Power Limit (W)400.0
GPU temperature (°C)58
GPU Clock Speed (MHz)1695
Memory Clock Speed (MHz)1512
Memory bandwidth2039 GB/s

GPU Configuration

CUDA cores6912
Tensor core432 (3rd generation)
Support for FP16/INT8Yes
ArchitectureAmpere
VirtualizationNVIDIA MIG

Benchmarks

ffmpeg~8ms
Coremark (iterations per second)~95000-115000
Llama2 Inference (tokens per second)~240-300
Tensorflow Mixed Training~8.5-10.8
Stable Diffusion (images/min)~85-105
Mixed Precision Training~12.8-16.2x

Providers with A100X