NVIDIA’s next-generation graphics cards may not boast support for PCI Express 5.0. The rumor comes from prominent leaker kopite7kimi, who claimed in a recent tweet that green team’s Lovelace-based AD102 GPU, which is expected to power the upper echelon of the GeForce RTX 40 Series, will be limited to PCIe 4.0. This is rather surprising, as NVIDIA’s new Hopper products feature support for the latest and greatest version of PCIe.
PCIe Gen4
— kopite7kimi (@kopite7kimi) April 24, 2022
NVIDIA AD102 “Ada” GPU for GeForce RTX 40 series might not feature PCIe Gen5 support (VideoCardz)
The “Ada Lovelace” architecture which is set to launch by the end of this year was rumored to feature PCIe Gen5 interface support, mainly because it already has the Intel ATX 3.0/PCIe Gen5 spec 16-pin connector unlocking up to 600W of power per cable.
Furthermore, Kopite mentioned that the structure of the single-precision (FP32) core configuration might not be as simple as Ampere’s. One should not rule out that the CUDA core count will not change again with the new generation of GeForce RTX series. A matrix of different integer and floating-point FP32 units might increase the number of cores, compute power, or introduce a new core type altogether. It is simply too early to tell.
Intel’s latest hardware already boasts support for PCIe 5.0, but AMD will soon follow with its next-generation AM5 platform, which is expected to launch with support for both PCIe Gen 5 and DDR5 memory later this year. PCIe 4.0 transfers data at 16 GT/s, while PCIe 5.0 doubles that rate at 32 GT/s.
Discussion (6 replies)
Join Discussion →Leaves them a feature to put on the ti range
Just makes sense because there is a lot more pcie 4 than 5. At least as I see it.
That doesn’t matter unless it’s not backwards compatible. Might save them $5 in components to not include pcie5 though
Still haven't seen many meaningful instances where PCI 3x16 hasn't been enough.
Sort of makes sense.
We're just barely getting to the limits of 3x16, there is plenty of bandwidth in 4.0. I remember seeing some tests when the 3090 came out on TechPowerUp and they showed you really had to push 4K over 120/144 FPS to get close to the limit of 3.0 and in reality how many people are going to do that while still rocking a mobo with 3.0?
I wouldn't be surprised to see data showing a 4090 doing 8K/120 FPS that exceeded 3.0 but definitely not 4.0.
What would be most interesting to me is to have the cards be 5.0 and then cram a bunch on 1 lane PCIE 5 slots to use for folding@home. 1 lane pcie 5 is, what the same as 4x pcie 3? That might be enough to not completely starve the cards for F@H