Bull GPGPU computing solutions

Hybrid cluster solutions and services that fully leverage the performance of accelerators

Demand for computing power is growing steadily, as scientists and engineers seek to tackle increasingly complex problems. The emergence of multi-core CPUs has allowed to keep pace with their demands, but energy consumption, space, and cooling have become major inhibitors to computing systems expansion. Hence the success of acceleration technologies such as GPGPUs (General-Purpose Graphics Processing Units), which offer both breakthrough performance and outstanding space and energy efficiency.

Bull expertise to maximize performance acceleration

However depending on the type of application and its degree of optimization, GPGPUs can accelerate processing by a factor of 1 to 100! Expertise is therefore essential to get the most out of GPUs. Bull engineers have had hands-on experience of porting and optimizing applications for hybrid clusters associating GPGPUs and CPUs through several cooperative projects with customers, such as a hybrid cluster with 46080 GPGPU cores, designed by Bull for GENCI. Bull experts deliver end-to-end services, from system design to advanced user training, through application optimization for GPGPUs.

Seamless integration in the HPC environment

Bull’s HPC software environment integrates all the tools needed to operate a hybrid system comprising Intel® Xeon®-based servers and NVIDIA® Tesla™ GPGPUs. The task management environment supports the allocation of applications to the relevant computing resource. An application can naturally use both computing resources so as to fully capitalize on the rich potential offered by the hybrid cluster.

NVIDIA® Tesla™ S1070 1U Computing System

With almost 1,000 cores in 1U the NVIDIA Tesla S1070 computing system delivers up to 4 teraflops of peak performance with a typical energy footprint of only 700 watts. It includes 16 GB of ultra-fast memory for maximum performance with larger data sets. Its massively-parallel, many-core architecture offers the ability to execute thousands of concurrent threads per GPGPU. The Tesla connects to a host that runs the OS and part of the applications, while compute-intensive parts are run on the GPGPU.

A perfect match for the NovaScale R422 (twin servers) series
The innovative packaging of the Bull NovaScale R422 series allows fitting 2 servers, i.e. 4 CPUs, into a 1U chassis. With its two PCIe x16 Gen2 slots, the drawer containing 2 servers is an ideal match for a Tesla S1070 system equipped with 2 PCIe connections, teaming 4 Intel® Xeon® CPUs with 4 GPGPUs. The Tesla S1070 connects to the twin servers through 2 interface cards installed in the 2 PCIe 16x Gen2 slots, and cabled to the Tesla.

More information on NVIDIA Tesla S1070
More information on NovaScale R422 E1

NVIDIA Tesla C1060 Computing Processor

The NVIDIA® Tesla™ C1060 computing processor, with 240 processor cores and a standard C compiler that simplifies application development, scales to solve the most important computing challenges more quickly and accurately. With the massively parallel architecture of the GPGPU, scientists and engineers can get a quantum jump in performance.

The CUDA C programming environment simplifies many-core programming and enhances performance by offloading computationally-intensive activities from the CPU to the GPGPU. It enables developers to utilize NVIDIA GPGPUs to solve the most complex computation-intensive challenges drug research, oil and gas exploration, and computational finance.

The powerful NovaScale R425 server (in tower or 4U rack form factor) is ideally suited to host an NVIDIA® Tesla™ C1060 computing processor, to create a high performance computing node or workstation.

More information on NVIDIA Tesla C1060
More information on NovaScale R425

Specifications: NVIDIA Tesla S1070 Computing System

Rack mount 1U drawer accommodating 4 Tesla C1060 processors # of Streaming Processor Cores: 960 (240 per processor) Frequency of processor cores: 1.296 GHz Single Precision floating point performance (peak) per processor 933 GFlops Double Precision floating point performance (peak) per processor 78 GFlops Floating Point Precision: IEEE 754 single & double Total Dedicated Memory 16 GB (organized as 4.0 GB per GPGPU) Ultra-fast memory access with 408 GB/sec total bandwidth (102 GB/sec peak bandwidth per GPU) 4x 512-bit GDDR3 memory interface (512-bit interface per GPGPU) Typical Power Consumption: 700 W Dimensions: 44 x 444 x 723 mm (HxWxD) System Interface: PCIe x16 or x8 Available with one or two PCIe connections per system Software Development Tools: C language compiler, debugger, profiler, and emulation mode for debugging Standard numerical libraries for FFT (Fast Fourier Transform), BLAS (Basic Linear Algebra Subroutines), and CuDPP (CUDA Data Parallel Primitives) Host requirements: Host system with PCI Express x8 or x16, PCIe Gen2 for best results Linux (64-bit and 32-bit) - Red Hat Enterprise Linux 4 and 5

Specifications: NVIDIA Tesla C1060 Computing Processor

Form Factor: 10.5" x 4.376", Dual Slot # of Tesla GPUs: 1 # of Streaming Processor Cores: 240 Frequency of processor cores: 1.296 GHz Single Precision floating point performance: 933 GFlops Double Precision floating point performance: 78 GFlops Floating Point Precision: IEEE 754 single & double precision floating point Total Dedicated Memory: 4 GB Memory speed: 800 MHz Memory Interface: 512-bit GDDR3 Memory Bandwidth: 102 GB/sec Max Power Consumption: <200 W peak, 160 W typical System Interface PCI Express x16 Generation 2 Auxiliary Power Connectors: Two 6-pin or one 8-pin Thermal Solution: Active fan sink Programming environment: C(CUDA)