GPU Computing

One of the major technologies transforming workflows at research institutions and universities and in industries ranging from finance to manufacturing is GPU computing—a hybrid computing model using GPUs and CPUs to boost application performance. GPU computing takes advantage of the massively parallel architecture of a GPU, which is composed of thousands of small processing cores, making it uniquely suited to accelerating parallel calculations. Originally developed for graphics processing and now integral to the high-end CAD workstation, the GPU is utilized as a coprocessor to perform compute-intensive tasks such as rendering and finite element analysis. By enabling applications to offload these tasks to the GPU, GPU computing allows users to do much more with a single machine.

The Next Generation of GPU Computing: The Power of Multi-GPU

The introduction of NVIDIA® Tesla® coprocessors—GPUs engineered specifically for parallel computing and programming—and multi-GPU technology ushered in the next generation in GPU computing. Offering better double precision floating point performance, faster PCI-E communication, and more on-board memory, Kepler-based Tesla GPUs turbocharge parallel processing and give users access to cluster-level performance in a compact, highly efficient package. Many design and engineering professionals now depend on dual- or multi-GPU systems, using Quadro® GPUs for design workflows and Tesla GPU accelerators for simulation. With a GPU workstation, professionals no longer have to compromise their time, creativity, or productivity due to unstable applications or inadequate system performance.

Personal GPU Computing: Thinkmate-Built Supermicro GPU Workstations

If you’re using one of the many industry-standard software programs compatible with CUDA, NVIDIA’s GPU computing specification, a GPU-optimized Supermicro workstation can help you leverage the full capabilities of your application and accelerate complex workflows. For a future-proof GPU computing solution, Thinkmate recommends Supermicro GPU SuperWorkstations—designed by Supermicro and built by Thinkmate to deliver superior performance, reliability and expandability. Featuring Supermicro’s innovative GPU server architecture, these personal supercomputing systems support dual Intel® Xeon® E5-2600 CPUs, up to 512 GB registered ECC memory, and up to 4 GPUs or Intel® Xeon Phi™ coprocessors.

Enterprise-level GPU Computing: Thinkmate-Configured Supermicro Supercomputing Servers

For cost-effective supercomputing in the data center, Supermicro GPU-optimized SuperServers are the ideal technology platform for running computationally-intensive applications in medicine, oil and gas exploration, astrophysics, and other demanding fields. Featuring industry-leading GPU system density and Supermicro’s pioneering non-blocking GPU system architecture, Supermicro GPU rack servers are available in a variety of form factors and offer users compelling benefits including:

  • Breakthrough performance. Get up to 1 TeraFLOPS of double precision performance per system with the fastest 1U GPU rackmount server in the industry.
  • Excellent manageability. Implement smart enterprise-class server management and GPU status monitoring via an onboard BMC supporting IPMI 2.0 and KVM/virtual media over a dedicated LAN.
  • Superior efficiency. Achieve greater energy efficiency and minimize TCO with Supermicro’s gold level power supplies, advanced cooling subsystems, and energy-saving motherboards.

Configure and Buy

Speak with an Expert Configurator at 1-800-371-1212