Virtualization for NVIDIA Tesla GPUs
vSMP Foundation supports the latest GPU products from NVIDIA. All attached GPUs will be recognized by vSMP Foundation and are available to all applications running on the system. For example, if 8 systems exist, each with 2 GPUs, then all 16 NVIDIA GPUs are available to the application.
Virtualization for Intel Xeon Phi
vSMP Foundation for Intel Xeon Phi create a large core count system using cores made available by Intel Xeon Phi co-processor. By virtualizing all the cores from the Intel Xeon reside in the host with the Intel Xeon Phi coprocessor cores, applications will run transparently, and no code modifications are necessary. All scaleable applications will benefit from using the additional cores that are available on the Intel Xeon Phi coprocessor. Accelerated-computing allows for the easy use of specific CPU acceleration technologies.
By aggregating all of the cores available on a system that combines the standard Intel Xeon cores with about 60 cores from the Intel Xeon Phi coprocessor, a large scale, large core count shared-memory is created.
Using the Versatile SMP™ (vSMP) architecture, this solution aggregates all of the Intel Xeon cpus in a system and any number of Intel Xeon Phi coprocessors. A single system is created with one copy of Linux installed. vSMP Foundation for Intel Xeon Phi supports the latest generation of Intel Xeon processors (56xx, 65xx, 75xx series,E7-x8xx, E5-x6xx) with any of the Intel Xeon Phi coprocessors.
- Run all applications immediately on systems containing the Intel Xeon Phi coprocessor.
- Create a single, transparent view of all cores
- All cores, from the Intel Xeon CPUs and Intel Xeon Phi coprocessor are available
- All host and Intel Xeon Phi memory is available to all applications.
- A large core count SMP is at your desktop
- Large memory footprint: Host + Intel Xeon Phi memory all directly accessible
- No code modification is necessary.
- Single system utilizing all host cores and Intel Xeon Phi coprocessor cores
- Scalable applications will show performance improvement
- Significant number of cores available, transparently