Monday, May 20th 2024
Intel's Next-Gen Falcon Shores GPU to Consume 1500 W, No Air-Cooled Variant Planned
Intel's upcoming Falcon Shores GPU is shaping up to be a powerhouse for AI and high-performance computing (HPC) workloads, but it will also be an extreme power hog. The processor, combining Gaudi and Ponte Vecchio successors into a single GPU, is expected to consume an astonishing 1500 W of power - more than even Nvidia's beefy B200 accelerator, which draws 1000 W. This immense power consumption will require advanced cooling solutions to ensure the Falcon Shores GPU operates efficiently and safely. Intel's partners may turn to liquid cooling or even full immersion liquid cooling, a technology Intel has been promoting for power-hungry data center hardware. The high power draw is the cost of the Falcon Shores GPU's formidable performance promises. Intel claims it will deliver 5x higher performance per watt and 5x more memory capacity and bandwidth compared to its Ponte Vecchio products.
Intel may need to develop proprietary hardware modules or a new Open Accelerator Module (OAM) spec to support such extreme power levels, as the current OAM 2.0 tops out around 1000 W. Slated for release in 2025, the Falcon Shores GPU will be Intel's GPU IP based on its next-gen Xe graphics architecture. It aims to be a major player in the AI accelerator market, backed by Intel's robust oneAPI software development ecosystem. While the 1500 W power consumption is sure to raise eyebrows, Intel is betting that the Falcon Shores GPU's supposedly impressive performance will make it an enticing option for AI and HPC customers willing to invest in robust cooling infrastructure. The ultra-high-end accelerator market is heating up, and the HPC accelerator market needs a Ponte Vecchio successor.
Sources:
ComputerBase.de, via Tom's Hardware
Intel may need to develop proprietary hardware modules or a new Open Accelerator Module (OAM) spec to support such extreme power levels, as the current OAM 2.0 tops out around 1000 W. Slated for release in 2025, the Falcon Shores GPU will be Intel's GPU IP based on its next-gen Xe graphics architecture. It aims to be a major player in the AI accelerator market, backed by Intel's robust oneAPI software development ecosystem. While the 1500 W power consumption is sure to raise eyebrows, Intel is betting that the Falcon Shores GPU's supposedly impressive performance will make it an enticing option for AI and HPC customers willing to invest in robust cooling infrastructure. The ultra-high-end accelerator market is heating up, and the HPC accelerator market needs a Ponte Vecchio successor.
34 Comments on Intel's Next-Gen Falcon Shores GPU to Consume 1500 W, No Air-Cooled Variant Planned
Was not expecting that from Intel but I don't keep up with professional GPUs at all. I guess there is indeed a tremendous market for such a beast now that the AI craze is upon us.
YES, power efficiency. Peak power draw may be insane, but performance in tasks might be even moreso.
Aurora (supercomputer) - Wikipedia
Intel Xe - Wikipedia
Maybe BPU (basic processing unit) and MPU (massive processing unit) could replace CPU and GPU respectively?
I'd propably call these things something like "compute accelerators", which seems to be the common denominator.
The architecture itself is still very similar to GPUs.
Same with NVIDIA data center offerings supporting MIG and/or GRID to remotely share virtual instances for 3D applications.
CPUs are still the master in the design. The OS runs on them in the address space managed by them. Being interconnected and cache-coherent changes nothing from software perspective (apart from increased performance and decreased latency of course). As in, it's not a completely new compute paradigm that would warrant changing the meaning of established terms.
With such TDP Intel's superchip ( GPU + x86 CPU cores ) doesn't have good future.