Performance

vSMP Foundation provides the following key performance benefits:

Memory Bandwidth

The vSMP architecture enables the aggregation of memory-bandwidth across boards, as opposed to traditional SMP architecture where memory-bandwidth decreases as the machine scales. This enables solutions based on vSMP Foundation to show close-to-linear memory-bandwidth scaling. Solutions based on vSMP Foundation deliver the world's highest memory-bandwidth for four-sockets and larger x86 systems.

CPU Speed

vSMP Foundation leverages the latest and greatest CPU technology available in the market at any point in time. The CPU technology refresh cycle, driven by Moore's-law, creates a faster and better CPU generation every 12-18 months. The fastest CPUs, which are also the first to market for a new generation, are the CPUs targeted at the volume server segment, which use single- and dual-processors. vSMP Foundation provides the benefits of deploying cutting-edge CPUs at the very date those CPU models are launched, without the need to wait for a lengthy integration and productization cycle.

Reducing or Eliminating the Need for Swap

vSMP Foundation enables the creation of large memory solutions, which enables an application requiring large amounts memory to run within RAM, and reduce the need to use a hard-drive for swap or scratch space. Application runtime can be dramatically reduced by running simulations with in-core-solvers or by using memory instead of swap for large-memory footprint models.

Choice of Parallelization Paradigms

With the advent of multi-core processors, computer systems these days are using more than one core, and application developers are looking for ways to harness the added CPU power to perform more calculations in lower runtime. Some parallelization paradigms are easier to implement than others. vSMP Foundation creates true SMP machines, which enable application developers to scale and achieve shorter runtime with any of the parallelization paradigms, including threaded (OpenMP, TBB, or Explicit pthreads) and distributed (MPI, PVM, etc) codes, thereby driving performance and scalability into more applications.

---

Measured Results

	STREAM (OMP) STREAM (OMP) demonstares the memory-bandwidth of solutions based on vSMP Foundation. The memory-bandwidth scales with each system boards added. Click here for a larger image.
	SPEC CPU2000 SPEC CPU2000 rate is running a single process per core. By comparing the rate (which is 1 / runtime) of 8 cores to 128 cores, the system scaling (or efficiency) is determined. The theoretical system scaling from 8 to 128 cores is 16, which defined as 100% efficiency. vSMP Foundation based solutions deliver 76% efficiency. Click here for a larger image.

 

Click here to request more information or access our white-paper: The Versatile SMP™ (vSMP) Architecture and Solutions Based on vSMP Foundation™