Image

NVIDIA GeForce RTX 5080 Core Count Same as the RTX 4080 Super, RTX 5090 Slightly Higher than the RTX 4090

6/05/2024

NVIDIA GeForce RTX 5080 Core Count Same as the RTX 4080 Super, RTX 5090 Slightly Higher than the RTX 4090

Let’s dive into the details of NVIDIA’s next-gen GeForce gaming GPUs, specifically the RTX 5080 and RTX 5090. According to recent rumors, these GPUs will be part of the Ada Lovelace lineup and share similarities with existing offerings.

  1. Core Counts:

    • The RTX 5080 is expected to feature up to 10,752 shader cores, with a few stream multiprocessors (SMs) likely disabled for better yields.
    • The RTX 5090, on the other hand, will likely have a partially disabled die, with 1-2 Graphics Processing Clusters (GPCs) fused off. This results in an overall shader count of either 22,528 or 20,480 FP32 shaders.

  2. Memory Configuration:

    • The RTX 5080 will use a 256-bit memory bus, paired with GDDR7 memory running at 28 Gbps. This yields a peak bandwidth of 896 GB/s, a 25% increase over the RTX 4080 Super.
    • The RTX 5090 is rumored to feature a 448-bit bus, up from the RTX 4090’s 384-bit bus. With 28 Gbps GDDR7 memory, this results in an impressive external bandwidth of 1,568 GB/s, a 56% uplift compared to the RTX 4090.

  3. Ray Tracing and Performance:

    • The RTX 5080 is expected to perform better than the RTX 4090 in ray-traced gaming, while offering similar performance in rasterization.
    • The performance deficit between the RTX 5080 and RTX 5090 should be similar to the gap between the RTX 4080 and RTX 4090.

In summary, NVIDIA’s upcoming GPUs promise exciting improvements in core counts and memory bandwidth, setting the stage for impressive gaming performance. Keep an eye out for official announcements to confirm these details! 🎮🔥