It seems like the Intel Xe GPU development or so called project Odyssey has been almost been finalized. As per a leak, Intel recently held a high profile “Xe Unleashed” event internally where the GPU leadership presented their finalized Xe methodology that will power Aurora Exascale Supercomputer.
Aurora Exascale, the world’s first supercomputer is capable of sustained exascale computing and would be delivered to the Argonne National Laboratory in 2021. The US Department of Energy, Intel, and Cray have signed a contract under which the two companies and DOE’s Argonne National Laboratory will develop and build the Aurora supercomputer capable of a “quintillion” floating point computations per second.
- GPUs will be based on 7nm node which Intel is yet to come up with,
- Intel Xe 2 GPU will boast a total of 12288 stream processors, with each internal chip packing 6144,
- These cores will have a base clock of 1600MHz with a boost of 2718MHz,
- The GPU will be leveraging Intel’s 4D XPoint memory and that too 32 GBs of it,
- The total bandwidth will be around 8TB/s with a processing performance of 66.8 TFLOPs,
- The TDP is rated at 350W with a Direct3D 14_2, API capability.
The Aurora supercomputer is designed to chew through data analytics, HPC and AI workloads at an exaFLOP pace, marking a new high for the supercomputing realm and possibly handing the U.S. the leadership position on the Top500 supercomputing list.
The Aurora supercomputer will be infused with a whole host of Intel’s technologies that form the basis of its new focus on six pillars: process technology, architectures, memory, interconnects, security and software. It also leverages Intel’s new OneAPI software.