Skip to main content

Xbox Series X: 12 Teraflops of GPU Performance Confirmed, More Details Revealed

(Image credit: Microsoft)

Microsoft has confirmed some of the Xbox Series X specs, including 12 teraflops of GPU performance. Xbox head Phil Spencer published a blog post listing information about a custom-built SSD, support for gaming at 120 frames per second (fps) and HDMI 2.1.

It was previously announced that the new Xbox will use AMD's Zen 2 processor technology and RDNA graphics, but now we're getting that 12 TFLOPS number direct from Microsoft itself. That's double an Xbox One X. 

Additionally, it will support hardware-accelerated DirectX ray tracing and variable rate shading (Sony previously announced ray tracing for the PS5).

Microsoft also touted a custom built, "next-generation SSD," as well as a quick resume feature to switch between multiple games in "suspended" states. 

The company claimed backwards compatibility for "four generations of gaming," including to the original Xbox. It also boasted of "Smart Delivery," which makes sure you get a version of the game specifically for the Xbox you're playing it on. Microsoft is suggesting you'll only need to buy each game once, should you want to buy a game on Xbox One and later on the Series X.

It's possible we'll learn more about the Xbox Series X at the GDC and E3 conferences in March and June, respectively, but what we do know is that the system will launch in time for the holidays later this year.

  • bwana
    10% the flops of a titan. So why don’t PCs have 10x the frame rate in games?
    Reply
  • prtskg
    bwana said:
    10% the flops of a titan. So why don’t PCs have 10x the frame rate in games?
    The article is talking about fp32 (single precision) compute while what you are talking about is tensor cores compute performance. This will give better comparison
    https://www.anandtech.com/show/13668/nvidia-unveils-rtx-titan-2500-top-turing
    Reply
  • bwana
    Thank you. So it looks like the Xbox is very close to the titan in single precision. But why is the 2080ti so low in tensor performance if its CUDA count is on par w the others?
    Reply
  • bit_user
    bwana said:
    why is the 2080ti so low in tensor performance if its CUDA count is on par w the others?
    Because Nvidia intentionally nerfed it. The same GPU delivers 2x the fp16-multiply/fp32-accumulate performance in the Titan RTX and equivalent Quadro RTX model. They just didn't want people buying gaming cards for AI training workloads, which is the main purpose of that feature. So, they cut the throughput of that particular instruction in half.

    However, if you compare the fp16-multiply/fp16-accumulate performance (not shown in that table), they're on par. That's for inference, which is used to accelerate things like global illumination ray tracing. So, they kept it at full performance.

    There's a wealth of information, buried in this page: https://en.wikipedia.org/wiki/List_of_Nvidia_graphics_processing_units
    Reply
  • Giroro
    bwana said:
    Thank you. So it looks like the Xbox is very close to the titan in single precision. But why is the 2080ti so low in tensor performance if its CUDA count is on par w the others?

    Nvidia likes to reduce certain features (especially Double Precision) in their gaming cards, so the customers who need it like professionals and data centers need to upgrade to the far more expensive Quadro line. Tensor performance isn't really that important to gaming right now compared to how important they are to Nvidia's big AI customers.
    Reply
  • bit_user
    Giroro said:
    Nvidia likes to reduce certain features (especially Double Precision) in their gaming cards, so the customers who need it like professionals and data centers need to upgrade to the far more expensive Quadro line.
    Fun fact: they haven't done that since Kepler. Since then, all of the consumer GPUs really don't have the hardware on die for more fp64 performance. In the case of Titan V, their only data center GPU to reach consumers since then, they kept fp64 performance at full speed.

    The Titan RTX is not properly built on a datacenter GPU - it's just an uncrippled version of the RTX 2080 Ti, which is a consumer GPU without more than token fp64.

    BTW, AMD's Radeon VII is built on a datacenter GPU, and AMD crippled its fp64 to 1/4th of the native capability. Even after that, it's still the fastest fp64 you can get below the $3000 Titan V.
    Reply
  • hannibal
    Interesting to see how big part of that 12 teraflops computational power is from raytrasing hardware...
    It is possible that this has less rasterisation power than 5700 has but still have more computational power!
    Reply
  • Giroro
    bit_user said:
    Fun fact: they haven't done that since Kepler. Since then, all of the consumer GPUs really don't have the hardware on die for more fp64 performance. In the case of Titan V, their only data center GPU to reach consumers since then, they kept fp64 performance at full speed.

    The Titan RTX is not properly built on a datacenter GPU - it's just an uncrippled version of the RTX 2080 Ti, which is a consumer GPU without more than token fp64.

    BTW, AMD's Radeon VII is built on a datacenter GPU, and AMD crippled its fp64 to 1/4th of the native capability. Even after that, it's still the fastest fp64 you can get below the $3000 Titan V.

    If it really is the case that the Tensor and RTX cores weren't left over from their datacenter GPUs... then I very much can't explain why they wasted so much of the TU102 die space and power consumption on them.
    Reply
  • LordVile
    admin said:
    Xbox Series X got some official specs, including 12 teraflops of GPU power, hardware accelerated ray tracing and smart delivery

    Xbox Series X: 12 Teraflops of GPU Performance Confirmed, More Details Revealed : Read more

    Eh the vega 56 has 10.5 TFLOPs and the 64 had a shade under 13 (the 2070 comes in at 7.5). Not really a measure of a GPUs power in gaming.
    Reply
  • bit_user
    Giroro said:
    If it really is the case that the Tensor and RTX cores weren't left over from their datacenter GPUs... then I very much can't explain why they wasted so much of the TU102 die space and power consumption on them.
    See for yourself, there's no Tesla card with a TU102:

    https://en.wikipedia.org/wiki/List_of_Nvidia_graphics_processing_units#Tesla
    Quadro RTX? Yes, it's on the 6000 and 8000 cards. While you can put them in servers, but they're mainly workstation-oriented cards.

    If servers were a big market for the TU102, there should be a Tesla model - like the Tesla P40, which featured the GP102 (of GTX 1080 Ti fame).
    Reply