Flopper.io Flopper.io

FP64

precision
Also known as: double precision

Definition

Double-precision 64-bit floating point, required for scientific and engineering workloads where rounding error compounds across many operations. Datacenter GPUs vary enormously in FP64 throughput, and recent AI-focused designs have cut it sharply because AI training does not need it.

GPUs featuring FP64

NVIDIA Rubin

Rubin · 2026 · 288 GB

NVIDIA Vera Rubin

Rubin · 2026 · 576 GB

NVIDIA B200

Blackwell · 2025 · 180 GB

AMD Instinct MI350X

CDNA 4 · 2025 · 288 GB

AMD Instinct MI355X

CDNA 4 · 2025 · 288 GB

AMD Instinct MI325X

CDNA 3 · 2025 · 256 GB
Browse All GPUs →

Related Terms

FP32

Single-precision 32-bit floating point, the long-standing default for general-pu...

TF32

TensorFloat-32 - A 19-bit precision format introduced with NVIDIA Ampere archite...

TFLOPS

Trillion Floating-Point Operations Per Second - A measure of computational perfo...

More precision Terms

TF32

TensorFloat-32 - A 19-bit precision format introduced with NVIDIA Ampere archite...

FP16

16-bit floating-point precision format (half precision). Commonly used in AI tra...

BF16

Brain Float 16 - A 16-bit floating-point format developed by Google with the sam...

Explore More Terms

Browse the full glossary or dive into architecture guides

Back to Glossary Architecture Guides