510.226.7366 | sales@exxactcorp.com |

Workstation Graphics and HPC Solutions  

Exxact Corp - Professional Graphics, Consumer Graphics, GPU Computing, Visual Computing, Other Products

NVIDIA Tesla GPUs - Accelerate scientific and technical computing workloads
GPU computing is the pairing of a GPU (Graphics Processing Unit) with a traditional CPU to extract maximum performance from a minimal financial investment, software license redundancies, and physical/carbon footprint. While a GPU is not necessary for a server or workstation, performance is severely bottlenecked by the CPUs inherent architecture. CPUs are made up of a handful of complex cores intended maximize serial processing where cores wait for cores to complete and pass on tasks. Hundreds of GPU cores fill the gaps between CPU cores with parallel processing to offload compute intensive work. As a result a performance maximizing symbiosis forms between CPU and GPU.
NVIDIA Tesla GPUs - Maximize performance for supercomputing professionals
• Full double precision floating point performance: Maximum 1.31 TFlops
• Faster PCIe communication: The only NVIDIA product with two DMA engines for bi-directional PCIe communication
• Higher performance on technical applications with large data sets: Larger on-board memory
• Faster communication with InfiniBand using NVIDIA GPUDirect™: Special Linux patch, InfiniBand driver, and CUDA driver
• Higher performance CUDA driver for Windows OS: TCC driver reduces CUDA kernel overhead and enables Windows Remote Desktop and Windows Services
The revolutionary NVIDIA® Pascal™ architecture is purpose-built to be the engine of computers that learn, see, and simulate our world—a world with an infinite appetite for computing. From silicon to software, Pascal is crafted with innovation at every level. Pascal architecture is built on five technological breakthroughs, enabling a new computing platform that is disrupting conventional thinking from the desk-side data center.

16 Nanometer FinFET for unprecedented energy efficiency

With 150 billion transistors built on bleeding-edge 16 nanometer FinFET fabrication technology, Pascal GPU is the world's largest FinFET chip ever built. It is engineered to deliver the fastest performance and best energy efficiency for workloads with near-infinite computing needs.

Pascal architecture for an exponential performance leap

Pascal is the most powerful compute architecture ever built inside a GPU. It transforms a computer into a supercomputer that delivers unprecedented performance. Pascal delivers over 5 TeraFLOPS of double precision performance for HPC workloads. For deep learning, a Pascal-powered system offers over 12x leap in neural network training performance compared to the current-generation GPU architecture.

NVIDIA NVLink for maximum application scalability

Pascal is the first architecture to integrate the revolutionary NVIDIA NVLink™ high-speed bidirectional interconnect. This technology is designed to scale applications across multiple GPUs, delivering a 5x acceleration in interconnect bandwidth compared to today's best-in-class solution.

CoWoS with HBM2 for big data workloads

Pascal architecture unifies processor and data into a single package to deliver unprecedented compute efficiency. Using an innovative approach to memory design, CoWoS® (Chip-on-Wafer-on-Substrate) with HBM2 provides a 3x boost in memory bandwidth performance over NVIDIA Maxwell™ architecture.

New artifical intelligence (AI) algorithms

Pascal is designed from the ground up to offer the highest performance for deep learning and other computational workloads. It takes advantage of new half-precision instructions to deliver more than 21 TeraFLOPS of peak performance for deep learning.


NVIDIA Maxwell architecture uses innovative design to dramatically improve energy efficiency. This architecture provides substantial application performance improvements over prior architectures by featuring large dedicated shared memory, shared memory atomics, and more active thread blocks per SM.

More efficient multiprocessors: Improvements to control logic partitioning, workload balancing, clock-gating granularity, instruction scheduling, number of instructions issued per clock cycle, etc
Larger, dedicated shared memory: 64KB of dedicated shared memory per SM
Fast shared memory atomics: Native shared memory atomic operations for 32-bit integers

NVIDIA Kepler features SMX, Hyper-Q, and Dynamic Parallelism, Kepler has greatly expanded the possibilities of scientific computing applications and levels the hybrid computing playing field for more application developers and researchers.

SMX: Same power consumption, three times the power
Hyper-Q: 32x throughput meants the GPU is always working
Dynamic Parallelism: More detail and better performance only where necessary

  Number of GPUs Number of CUDA Cores Memory Size (GDDR5) Memory Bandwidth (ECC off) Architecture Features Peak Double precision Floating Point Peak Single Precision Floating Point System
NVIDIA Tesla P100 for PCIe 1x GP100 3584 12GB or 16GB 540 or 720 GBytes/sec Extreme performance, NVLink (SXM2 only), HBM2, Unified Memory and Compute Preemption, 16nm FinFET 4.7 teraflops 9.3 teraflops Servers and Workstations
NVIDIA Tesla P100 for SXM2 1x GP100 3584 16GB 720 GBytes/sec 5.3 teraflops 10.6 teraflops Servers
NVIDIA Tesla P40 1 x GP102 3840 24GB 346 GBytes/sec Real-Time Inference, 47 TOP/s of INT8 47 TOPs of INT8 performance 12 teraflops Servers
NVIDIA Tesla P4 1 x GP104 2560 8GB 192 GBytes/sec Real-Time Inference, 22 TOP/s of INT8 22 TOPs of INT8 performance 5.5 teraflops Servers
NVIDIA Tesla M60 2 x GM205 4096 16GB 320 GBytes/sec SMX, Dynamic Parallelism, Hyper-Q 0.3 teraflops 9.7 teraflops Servers
NVIDIA Tesla M40 1 x GM200 3072 12GB / 24GB 288 GBytes/sec SMM, Dynamic Parallelism 0.21 teraflops 7.0 teraflops Servers and Workstations
NVIDIA Tesla M4 1 x GM206 1024 4GB 88 GBytes/sec 0.07 teraflops 2.20 teraflops Servers and Workstation
NVIDIA Tesla K80 2 x GK210 4992 24GB 480 Gbytes/sec SMX, Dynamic
1.87 teraflops 5.60 teraflops Servers
NVIDIA Tesla K40 1 x GK110B 2800 12GB 288 Gbytes/sec 1.43 teraflops 4.29 teraflops Servers and Workstations
NVIDIA Tesla K20X 1 x GK110 2688 6GB 250 GBytes/sec 1.31 teraflops 3.95 teraflops Servers only
NVIDIA Tesla K20 1 x GK110 2496 5GB 208 GBytes/sec 1.17 teraflops 3.52 teraflops Servers and Workstations
NVIDIA Tesla K10 2 x GK104s 2x 1536 8GB 320 GBytes/sec SMX 0.19 teraflops 3.95 teraflops Servers
GPU Accelerated Applications
In the past year there has been a 60% growth in the number of CUDA supported applications that utilize the tremendous computing capability of NVIDIA Tesla GPUs. Exxact works closely with application developers to customize our GPU systems for maximizing our customers‘ productivity and workflow. See how Exxact GPU solutions are transforming computational research in fields such as:
Molecular Dynamics Government and Defense Computational Fluid Dynamics Numerical Analytics
> AMBER Certified MD Systems > NVIDIA GeoInt Accelerator    
Manufacturing: CAD & CAE Oil & Gas Bio-Informatics & Life Science Research & Higher Education