The NVIDIA AD102 GPU driving the next-generation GeForce RTX 40 series is rumored to remain in the PCIe Gen 4.0 protocol.

Home » Tech » The NVIDIA AD102 GPU driving the next-generation GeForce RTX 40 series is rumored to remain in the PCIe Gen 4.0 protocol.

April 26, 2022
No Comments

GeForce RTX 40 series graphics cards with NVIDIA’s Ada Lovelace GPU architecture are expected to remain traditional PCIe Gen 4.0 compliant, as reported by Kopite7kimi.

NVIDIA GeForce RTX 40 “AD102 GPU” graphics card will maintain PCIe Gen 4.0 compliance

NVIDIA plans to launch a GeForce RTX 40 series graphics card with the all-new Ada Lovelace GPU architecture later this year.

The specifications and specific configuration of the graphics card lineup have already been leaked, but the design of the card itself is more interesting.

PCIe Gen4

– kopite7kimi (@ kopite7kimi) April 24, 2022

So far, NVIDIA GeForce RTX 40 series graphics cards have been found to employ a new ATX 3.0 compliant 12PVHPWR 16-pin connector that can draw up to 600W of power through the new PCIe Gen 5 power connector interface. ..

This power connector has already been adopted on GeForce RTX 3090 Ti graphics cards and is now capable of drawing up to 450W of power through a triple 8-pin adapter.

However, there is another aspect to enabling full PCIe Gen 5.0 compliance, which is the interface connector itself.

NVIDIA GeForce RTX 40 “AD102 GPU” graphics card remains PCIe Gen 4.0 compliant 2

Today, modern graphics cards communicate with the CPU via the PCIe Gen 4.0 protocol.

The PCIe Gen 4.0 protocol enables a total bandwidth of 64GB / sec and bidirectional 32GB / sec.

However, the latest Intel and AMD platforms support the all-new PCIe Gen 5.0 interface protocol.

The new standard can deliver up to 128GB / sec in total and up to 64GB / sec in both directions.

This effectively doubles the bandwidth, but upcoming graphics cards, and at least high-end GeForce RTX 40 graphics cards based on the AD102 GPU, don’t seem to have a PCIe Gen 5.0 interface yet. is.

According to Kopite7kimi’s tweet, the next GeForce RTX 40 lineup will maintain the PCIe Gen 4.0 protocol.

This is NVIDIA’s bold move to stay out of the next-generation standard, even though the Hopper GPU will be one of the first to take advantage of the new protocol in the HPC segment.

This is because the server requires a lot of bandwidth and the Gen 5.0 protocol will help these environments.

For consumers, the PCIe Gen 5.0 interface is too wide bandwidth, and today’s GPUs haven’t fully mastered the PCIe Gen 4.0 interface yet.

In the Radeon RX 6500 series and RX 6400 series, switching to the Gen3 standard resulted in unnecessarily narrow graphics bandwidth, which resulted in poor performance against PCIe Gen4.0 compliant standards, but PCIe Gen4.0 is low. With lanes, you don’t have to worry about bottlenecks, so there are good signs in the entry-level lineup.

If the high-end lineup isn’t hungry for the Gen 4.0 standard, then the low-end lineup is far from meeting the maximum threshold.

For now, I can’t say for sure if NVIDIA will really maintain PCIe Gen 4.0 on the next RTX 40 series card, but it could change as we market new cards with the PCIe Gen 5.0 logo. hey.

Beyond support for PCIe Gen 5.0 and PCIe Gen 4.0, NVIDIA appears to be making major changes to the way CUDA cores are placed within the Ada Lovelace architecture.

The GeForce RTX 40-series GPUs could not just add CUDA cores from Ampere, but also include a number of new mixed-precision cores whose details have yet to be revealed.

The lineup is still a few months away, so it may change a lot, but I would like to keep you up to date with the latest information.

NVIDIA CUDA GPU (rumors) provisional version

ソース：wccftech – NVIDIA AD102 GPUs Powering Next-Gen GeForce RTX 40 Series Rumored To Stick With PCIe Gen 4.0 Protocol

Explanation:

It is said that the RTX 4000 series will stay at PCIe Gen 4.0.

Of course, PCIe is backwards compatible, so you can use it with Gen5.0 compatible products, so don’t worry.

That’s why nVidia stayed at PCIe Gen 4.0, but you probably decided that the Gen 4.0 bandwidth was sufficient.

Only the power supply is Gen 5.0.

If you have enough bandwidth to prevent performance, forcing a wideband standard will only increase costs and not benefit.

If nVidia makes such a decision after repeating the test with chitin, I think that the decision is correct.

Despite the addition of one rival called Intel, nVidia, which did not support Gen 5.0, which is (probably) only meaningful for marketing, feels like a champion.

nVidia RTX 3000 series GPU

RTX3090