The Intel Xeon Phi x100 series of co-processors offer double-precision floating point performance approaching a teraFLOPS in a single add-in card. This performance is accessible through normal x86 instructions that leverage a high number of cores on the card and 16GB of high-speed shared memory that this Phi 7120P model comes equipped with. Further, each core features a 512 bit wide SIMD vector unit and 4 layer hardware threading. Code that has been optimized for standard Intel SSEx/AVX instructions should port readily to Phi.
From a system perspective, the card appears as an additional node on an internal network over the PCIe bus and is running an embedded Linux OS with an interface provided by openbox. This means you can log into the card as a separate node and have a normal Linux command environment available. Booting, resetting, re-configuring, user management, monitoring, etc is handled by a set of commands and kernel module which communicate with the card via a system daemon on the host.
With all of this factored in, the Xeon Phi can be an attractive alternative to the more well established NVIDIA Tesla card with its CUDA environment. The Phi provides a much more familiar programming "feel" and can take full advantage of Intel's advanced compiler suites.
Please note that this specific version does not have active cooling, so it can only be used in systems which are equipped to provide plenty of airflow to it - like the Puget Systems Peak workstation line.
From a system perspective, the card appears as an additional node on an internal network over the PCIe bus and is running an embedded Linux OS with an interface provided by openbox. This means you can log into the card as a separate node and have a normal Linux command environment available. Booting, resetting, re-configuring, user management, monitoring, etc is handled by a set of commands and kernel module which communicate with the card via a system daemon on the host.
With all of this factored in, the Xeon Phi can be an attractive alternative to the more well established NVIDIA Tesla card with its CUDA environment. The Phi provides a much more familiar programming "feel" and can take full advantage of Intel's advanced compiler suites.
Please note that this specific version does not have active cooling, so it can only be used in systems which are equipped to provide plenty of airflow to it - like the Puget Systems Peak workstation line.
Model: Intel SC7120P
Specifications
# of Cores: 61
Clock Speed: 1.238 GHz
Max Turbo Frequency: 1.333 GHz
Cache: 30.5 MB
Lithography: 22 nm
Max TDP: 300 W
PCI Express Revision: 2.0
Memory Specifications
Memory Size: 16 GB
# of Memory Channels: 16
Max Memory Bandwidth: 352 GB/s
ECC Memory Supported: Yes
Performance
Double Precision Floating Point (Peak): 1.2 Tflops
Single Precision Floating Point (Peak): 2.4 Tflops
