Puget Systems, one of the leading custom PC builder companies, was conducting a test on AMD and Intel PC server CPU-based systems that utilized NVIDIA GeForce GPUs from the last generation and the current generation and included the enterprise NVIDIA RTX 6000 Ada generation graphics card.
P2P functionality missing from newer NVIDIA GPUs as a resolution for NVIDIA to force customers to spend more
During the tests, the information revealed that the GeForce RTX 4090 graphics card had a faulty Peer-To-Peer (P2P) functionality. In turn, the company contacted NVIDIA, who explained that the new graphics card generation no longer offers P2P support.
When Puget Systems attempted to process P2P-based workloads in its benchmarking tests, it appeared "corrupted" or would fail upon access. The testing was sparked because a user on the NVIDIA forums had P2P failures on their system using dual GeForce RTX 4090 graphics cards. It was not until several months had passed that an NVIDIA representative responded to the user.
Hi all. Apologies for the delay. Feedback from Engineering is that Peer to Peer is not supported on 4090. The applications/driver should not report this configuration as peer to peer capable. The reporting is being fixed and future drivers will report the following instead.
— NVIDIA representative
P2P allowed for data to be able to transfer from one NVIDIA-supported graphics card's memory to another NVIDIA GPU. The functionality allows the system memory to be focused on other tasks as the P2P process uses the graphics cards' integrated memory, accelerating access and transmissions to the memory. CUDA-based programs benefit from this feature, especially those utilizing NVLink.
NVLink is a high-speed GPU interconnect that offers a "significantly faster alternative for multi-GPU systems than PCIe-based connections." P2P utilizes both PCIe and NVLink for its processes. The last NVIDIA generation that used P2P in its hardware was the Turing GPUs (RTX 20 series). The company has never answered why the functionality was dropped, but it is possible that it was due to requiring customers to purchase more expensive graphics cards.
NVIDIA has a significant focus on gaming with its graphics card products. However, over the last five years, NVIDIA has seen its cards used for other content outside the gaming realm. Other options that consumers were using the NVIDIA GPUs for were cryptocurrency mining and content creation, among other tasks requiring the power a GPU brought. During the Ampere generation, GeForce RTX 3090 graphics cards were used over expensive NVIDIA Quadro cards so that businesses could save overhead costs. When NVIDIA discovered this, the blower-based GPU saw fewer shipments to sellers and eventually evaporated.
Now, businesses are encountering an issue with GeForce RTX 4090 and RTX 6000 series graphics cards. The former offers the same architecture (Ada Lovelace) and the same silicon (AD102) as the latter, but for a cost difference of almost $4,000. The most considerable difference is that it does not offer the P2P found in the RTX 6000 series GPUs. This expensive undertaking for enterprises removes from their overall costs when before they were able to use that overhead for other tasks and processes. Due to the RTX A6000 GPU supporting P2P, businesses are not out of luck. The RTX A6000 graphics card costs $4,650, which is still less than the newer offerings from NVIDIA.
The post NVIDIA GeForce GPUs Do Not Incorporate Peer-To-Peer Support, Redirecting Consumers To Purchase Expensive GPUs by Jason R. Wilson appeared first on Wccftech.
Refference- https://wccftech.com
0 Comments