Given that millions of scientists, engineers, researchers, and even animation designers around the world are seeing the need for bigger, faster, and more powerful computing solutions, there is an unmet demand for supercomputers at affordable prices. Global Processing Unit (GPU) powered supercomputing clusters, also called GPGPU, provide access to massive computing capacity while reducing energy consumption. In the recent past, GPUs have evolved to enable several applications perform at much higher speeds than is possible using multi-core systems.In partnership with several industry leaders, Tyrone offers a complete range of easy-to-deploy, pre-configured, GPU-optimized supercomputing solutions that combine Green features with unmatched performance.
Tyrone brings the power of supercomputing on to your desktop so you can experience cluster-level computing performance – up to 250 times faster than standard PCs and workstations – right at your desk. This supercomputer, meant for office / personal use, is:
+ Designed to be plugged into regular power sockets
+ Very quiet, and
+ Scalable to harness the power of thousands of processor cores
+ Compatible with Windows and Linux
Here is what you get with a Tyrone personal supercomputer powered by NVIDIA’s revolutionary CUDA architecture:
Peak Performance 18.5 teraflops of computing capability in Tower
Network/Interconnect Compatibility Ethernet
GPU clusters are the next level to supercomputing that goes beyond personal supercomputers in terms of speed and scale. Tyrone presents HPC solutions powered by Tesla S1070 computing systems. This 4 teraflop 1U system is powered by the world’s first 1 teraflop processor. Ideal for computational research technical projects that require dedicated resources, Tyrone’s high density GPU cluster offers incredible performance per watt and can help scale your application to multiple GPUs.
Fitted with 4 1U cards, our solution enables a performance of around 4 teraflops in a 2U form factor. Theoretically, 250 of such systems can be clustered to provide peak performance of a petaflop.
Peak Performance 14TF in 1U, 18.5TF in 2U/4U, 560TF per 42U rack, One PetaFLOP in 2 racks
Network Compatibility InfiniBand and Ethernet
Tyrone, in association with NVIDIA, offers Tesla Bio Workbench, a specialized HPC platform that aims to convert a scientist’s standard PC into a “computational laboratory” capable of running complex bioscience codes at unparalleled speed. Using NVIDIA’s massively parallel CUDA architecture, the Tesla Bio Workbench consists of the following:
A range of GPU-optimized bioscience applications for molecular dynamics and quantum chemistry- based research, including: AMBER, GROMACS, LAMMPS, NAMD, TeraChem, VMD, and bioinformatics applications like CUDASW++ (Smith-Waterman), GPU-HMMER, and MUMmerGPU.
vA community site for downloading applications, checking out the latest benchmarks, reading academic papers and tutorials, joining discussion forums with the application developers themselves, and more.The Tesla Bio Workbench can be deployed on GPU-based desktop personal supercomputers or in data center solutions.
Jacket HPC Family
Jacket is a software platform designed to leverage GPU resources to the maximum to obtain maximum performance without being hassled by low-level programming details. Jacket runs on NVIDIA GPUs on any type of system, including laptops, desktops, servers, and even on clusters and the Cloud. Currently front-ended by the MATLAB language, the Jacket platform consists of a language processing system and a runtime that automatically optimizes applications and algorithms for GPU computing.
Tyrone, in association with Jacket, provides solutions to simplify high performance computing and make GPU-based computing accessible to a wider audience in the industry.