Nvidia’s RTX 5080 Has Faster Ray Tracing Than the RTX 4090

nvidia rtx 4090
Credit: Nvidia


nvidia rtx 4090
Credit: Nvidia

Breakdown

  • GeForce RTX 4090 is currently Nvidia’s most powerful graphics card
  • The 4090 currently holds the internal company record as the fastest ray tracer
  • RTX 5080 will be able to outperform the RTX 4090 with even faster ray tracing

The GeForce RTX 4090 is currently Nvidia’s most powerful graphics card. However, leaked information indicates that the rumored RTX 5080 will have faster ray tracing, better performance, and a much lower price tag.

Ray tracing is the technology that allows graphics cards to recreate the behavior of light in the real world, resulting in unparalleled realism in renders and images. It is the core technology behind Nvidia’s RTX cards, which can perform this process in real-time.

As the cream of Nvidia’s current crop of GPUs, the RTX 4090 currently holds the internal company record as the fastest ray tracer. However, information leaked by X user AGF indicates that the RTX 5080 will be able to break that performance standard when it hits the market.

According to the leaker, the RTX 5080 will be developed based on the performance of the "Blackwell" GB203 GPU. This GPU has been observed on tests to be faster at ray tracing than the AD102 on which the RTX 4090 is based, although both GPUs have nearly equal performances in rasterizing.

In other words, the RTX 5080 will be able to perform faster ray tracing than the RTX 4090 when it comes out.

AGF also claimed that the RTX 5080 and the other Blackwell GPUs in the same lineup will use GDDR7 memory technology. Previous Samsung tests have shown that the GDDR7 can run up to 32GBps.

nvidia ceo jensen huang b200 blackwell gpu
expand image

However, Nvidia has not yet released any information on future gaming GPUs, so this leaked information is merely speculation.

Nvidia CEO Jensen Huang unveiled the B200 GPU, the successor to the highly successful H100 “Hopper” cards used for artificial intelligence training, during GTC 2024. The GB200 is capable of 20 petaflops in AI performance, five times the four petaflops observed from the H100.

This Article's Topics

Explore new topics and discover content that's right for you!

Nvidia