Scribd
Upload
46 views2 pages

DS Tesla M Class Aug11

TESLA(tm) M-CLASS GPU CoMPUTinG ModULES are the world's fastest parallel computing processors for high performance computing (HPc) Based on the cUDA architecture codenamed "fermi," these GPUs feature up to 665 gigaflops of double precision performance, 1 teraflop of single precision performance, Ecc memory error protection, and L1 and L2 caches. Tesla GPUs speed up HPc applications by 10x.

Uploaded by

Mihai Lazăr
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
46 views2 pages

DS Tesla M Class Aug11

TESLA(tm) M-CLASS GPU CoMPUTinG ModULES are the world's fastest parallel computing processors for high performance computing (HPc) Based on the cUDA architecture codenamed "fermi," these GPUs feature up to 665 gigaflops of double precision performance, 1 teraflop of single precision performance, Ecc memory error protection, and L1 and L2 caches. Tesla GPUs speed up HPc applications by 10x.

Uploaded by

Mihai Lazăr
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

TESLA M-CLASS GPU CoMPUTinG ModULES ACCELErATinG SCiEnCE

Based on the cUDA architecture codenamed Fermi, the Tesla M-class GPU computing Modules are the worlds fastest parallel computing processors for high performance computing (HPc). Tesla GPUs high performance makes them ideal for seismic processing, biochemistry simulations, weather and climate modeling, signal processing, computational finance, cAE, cFDand data analytics.
Accelerate your science with NVIDIA Tesla 20-series GPUs. A companion processor to the cPU in the server, Tesla GPUs speed up HPc applications by 10x. Based on the Fermi architecture, these GPUs feature up to 665 gigaflops of double precision performance, 1 teraflop of single precision performance, Ecc memory error protection, and L1 and L2 caches. The Tesla M-class GPU modules are integrated into GPU-cPU servers from oEMs. This gives data center IT staff much greater choice in how they deploy GPUs, with a wide variety of rackmount and blade systems and with remote monitoring and management capabilities enabling large data center, scale-out deployments.

oIL & GAS

TEcHNIcAL SPEcIFIcATIoNS
Tesla M2090 Peak double precision oating point performance Peak single precision oating point performance CUDA cores Memory size (GDDR5) Memory bandwidth (ECC off) 665 Gigaops 1331 Gigaops 1030 Gigaops 512 6 GigaBytes 177 GBytes/sec 448 6 GigaBytes 150 GBytes/sec
* Note: With Ecc on, 12.5% of the GPU memory is used for Ecc bits. So, for example, 3 GB total memory yields 2.625 GB of user available memory with Ecc on. MoLEcULAr DyNAMIcS

Tesla M2070 / M2075 515 Gigaops Tesla M2050 515 Gigaops 1030 Gigaops 448 3 GigaBytes 148 GBytes/sec
NVIDIA TESLA | DATASHEET | AUG 11

TESLA M-cLASS GPU coMPUTING MoDULE


Features
HUNDrEDS oF cUDA corES

Benefits
Delivers up to 665 Gigaflops of double-precision peak performance in each GPU, enabling servers from leading oEMs to deliver more than a teraflop of double-precision performance per 1 rU of space. Single precision peak performance is over one Teraflop per GPU. Meets a critical requirement for computing accuracy and reliability in datacenters and supercomputing centers. Internal register files, L1/L2 caches, shared memory, and external DrAM all are Ecc protected. Maximizes performance and reduces data transfers by keeping larger data sets in local memory that is attached directly to the GPU. Integrates the GPU subsystem with the host systems monitoring and management capabilities such as IPMI or oEM-proprietary tools. IT staff can thus manage the GPU processors in the computing system using widely used cluster/grid management solutions. Accelerates algorithms such as physics solvers, ray-tracing, and sparse matrix multiplication where data addresses are not known beforehand. Maximizes the throughput by faster context switching that is 10X faster than previous architecture, concurrent kernel execution, and improved thread block scheduling. Turbocharges system performance by transferring data over the PcIe bus while the computing cores are crunching other data. choose c, c++, opencL, Directcompute, or Fortran to express application parallelism and take advantage of the innovative Fermi architecture.

Ecc MEMory Error ProTEcTIoN

UP To 6GB oF GDDr5 MEMory PEr GPU

SySTEM MoNITorING FEATUrES

L1 AND L2 cAcHES AS PArT oF THE NVIDIA PArALLEL DATAcAcHE

NVIDIA GIGATHrEAD ENGINE

ASyNcHroNoUS TrANSFEr WITH DUAL DMA ENGINES

FLEXIBLE ProGrAMMING ENVIroNMENT WITH BroAD SUPPorT oF ProGrAMMING LANGUAGES AND APIS

SoFTWArE AND DrIVErS


> > Software applications page: http://www.nvidia.com/object/vertical_solutions.html Tesla M-class GPU computing modules are supported for both Linux and Windows 64-bit Drivers NVIDIA recommends that users get drivers for M-series products from their System oEM to ensure that driver is qualified by the oEM on their system. Latest drivers can be downloaded from http://www.nvidia.com/drivers Learn more about Tesla data center management tools at http://www.nvidia.com/object/software-for-tesla-products.html > Software development tools are available at http://www.nvidia.com/object/tesla_software.html

To learn more about NVIDIA Tesla, go to www.nvidia.com/tesla


2011 NVIDIA corporation. All rights reserved. NVIDIA, the NVIDIA logo, NVIDIA Tesla, cUDA, GigaThread, Parallel Datacache, and Parallel NSight are trademarks and/or registered trademarks of NVIDIA corporation. All company and product names are trademarks or registered trademarks of the respective owners with which they are associated. Features, pricing, availability, and specifications are all subject to change without notice.

You might also like