Intel Announces Knights Landing Xeon Phi Accelerator Availability
The first systems with Intel’s Knights Landing Xeon Phi accelerator have been sent to customers. As a result, the manufacturer has not fulfilled its promise to make the first systems available in the first quarter.
With the introduction of the new Xeon Phi accelerators, it is possible to boot from the CPU, the memory is integrated on the CPU for extra performance and it is possible to create a complete cluster with only Xeon Phi without other Xeon systems.
The Knights Landing Xeon Phi map has been tested with several partners since late last year, such as Cray. The new Xeon Phi card is produced at 14nm and contains a maximum of 72 cores and has 16GB multi-channel dram with which a throughput speed of 400GB/s should be achieved. In addition, Knights Landing supports 36 pci-e 3.0 lanes and the Omni-Path architecture.
Omni-Path is part of Intel’s Scalable System Framework for high performance computing. Intel now says that the accelerator is suitable for applications in the field of artificial intelligence, among other things. That emphasis seems to be prompted by Google, which has started its own artificial intelligence project around hardware.
| 7290 | 72 | 1,5 | 16GB, 7.2GT/s | Yes | 384GB/2400MHz | 245W | $6254 |
| 7250 | 68 | 1.4 | 16GB, 7.2GT/s | Yes | 384GB/2400MHz | 215W | $4876 |
| 7230 | 64 | 1.3 | 16GB, 7.2GT/s | Yes | 384GB/2400MHz | 215W | $3710 |
| 7210 | 64 | 1.3 | 16GB, 6.4GT/s | Yes | 384GB/2133MHz | 215W | $2438 |
The Xeon Phi 7290 will not be available until early September. Tdp with Omni-Path architecture is 15W higher. Surcharge Omni-Path architecture is $278