Skip to content
vsMars
Spec drill · CUDA / Stream cores

GeForce RTX 5090 vs Radeon RX 9070 XT

A side-by-side readout for cuda / stream cores.

NVIDIA · GeForce RTX 5090
21760
▲ Lead
AMD · Radeon RX 9070 XT
4096
VerdictGeForce RTX 5090 wins on cuda / stream cores.
Context

Understanding cuda / stream cores

CUDA cores are NVIDIA's general-purpose shader units. They handle raster rendering, compute, and general parallel workloads.

Cross-generation comparison

CUDA cores from different architectures (Ampere vs Ada vs Blackwell) are not equivalent — Blackwell cores deliver more work per clock thanks to architectural improvements. Within the same generation, CUDA count is a strong predictor of performance.

  • Tensor Cores — accelerate matrix math for DLSS, ML inference.
  • RT Cores — accelerate ray traversal.
  • CUDA Cores — everything else (raster, general compute).

A modern RTX card lists all three. Tensor and RT counts scale roughly with CUDA count within a generation.

This matchupGeForce RTX 5090's 21760 is roughly 431% higher than Radeon RX 9070 XT's 4096 (a 17664 gap). Whether that gap is noticeable depends on workload — small percentage gaps rarely change day-to-day experience, while gaps of 20% or more usually do.

Glossary

What is cuda cores?

NVIDIA's parallel shader processors. Each CUDA core executes one floating-point or integer operation per clock; total count is a rough indicator of raster throughput within a generation.

Read the full CUDA cores explainer →
Drill more

Other specs on this comparison

See the full Graphics Cards category for all products ranked by Mars Score.