NVIDIA
GPUDatacenterReleased Nov 8, 2022

NVIDIA A800 SXM4 80GB

80 GB HBM2e · 5120-bit · 6,912 cores

The A800 SXM4 80GB is the export-compliant counterpart to NVIDIA's flagship Ampere data center accelerator, created in 2022 so that NVIDIA could continue supplying the Chinese market after United States export controls capped the permitted chip-to-chip interconnect rate. Silicon and memory are unchanged from the A100 it derives from; the sole difference is a deliberately reduced NVLink rate, which leaves single-GPU performance intact while slowing the all-to-all communication that very large multi-GPU training runs depend on. In the SXM module form factor it mounts on an HGX baseboard in groups of four or eight rather than plugging into a PCIe slot, giving it a higher power ceiling and faster memory than the PCIe card version of the same part.

General

Manufacturer
NVIDIA
Model
A800 SXM4 80GB
Segment
Datacenter
GPU chip
GA100
Architecture
Ampere
Release date
Nov 8, 2022

Memory

Memory size
80 GB
Memory type
HBM2e
Bus width
5120-bit
Bandwidth
2,039 GB/s
Memory clock
1,593 MHz

Clocks & silicon

GPU clock
—
Boost clock
1,410 MHz
Shader units
6,912 CUDA Cores
Tensor / AI cores
432
RT cores
—
Transistors
54.2 billion
Technology (nm)
7 nm
TSMC N7
TDP / TBP
400 W

Interface & APIs

Bus interface
SXM4 (PCIe 4.0 x16 host)
SLI / multi-GPU
NVLink 3.0 (400 GB/s)
DirectX
N/A (compute-only)
OpenGL
N/A (compute-only)
Vulkan
N/A (compute-only)
Other APIs
CUDA (compute capability 8.0), OpenCL 3.0

AI support

Precisions
FP8FP16BF16INT8INT4TF32FP64
AI accelerator
Third-generation Tensor Cores (432)
Peak AI TOPS
—
Notes
Third-generation Tensor Cores supporting FP64, TF32, BF16, FP16, INT8, INT4 and binary; FP8 arrived with Hopper. Peak throughput with structured sparsity: INT8 1,248 TOPS, FP16 and BFLOAT16 624 TFLOPS, TF32 312 TFLOPS; dense rates are half. FP64 is 9.7 TFLOPS, FP64 Tensor Core and FP32 19.5 TFLOPS each. Multi-Instance GPU partitions the board into up to seven isolated instances.
All NVIDIA GPUs →

Datacenter gpus from NVIDIA announced around the same time as the NVIDIA A800 SXM4 80GB. Add any of them to your compare list to see the specs side by side.

NVIDIAgpu

NVIDIA L40G

24 GB GDDR6 · 384-bit · 18,176 cores

Released
Jan 19, 2023
NVIDIAgpu

NVIDIA L40

48 GB GDDR6 ECC · 384-bit · 18,176 cores

Released
Sep 20, 2022
NVIDIAgpu

NVIDIA H100 CNX

80 GB HBM2e · 5120-bit · 14,592 cores

Released
Mar 22, 2022