Transcript
Multi-GPU Solutions
NVIDIA® Quadro® 5000 Compute Pack The NVIDIA 5000 Compute Pack consists of four NVIDIA Quadro 5000 GPU boards. The NVIDIA Fermi-based Quadro 5000 GPU combines CUDA high performance computing, OpenGL hardware acceleration and advanced visualization capabilities. Featuring massive memory, 2.5GB per GPU, ECC memory integrity and fast double-precision, the new Fermi-based Quadro GPU solutions deliver “best of class” features. NVIDIA Quadro solutions are well known for high performance standards, stability and enterprise level support.
VCGPU5K-PCK
Quadro 5000/GPU GPU memory
2.5GB DDR5
Memory Interface
320-bit
CUDA Cores
352
Max Power Consumption Remote GPU Monitoring and Power Management
150W Yes
Target Markets: • Clustered GPU compute servers • Personal supercomputing systems • Remote rendering and video encoding • Clustered image processing systems • Large-scale visualization
5 Reasons why Quadro is the right choice: Quadro 5000 Performance Graphics
Compute
Software Power
Tesla
GeForce
SPECviewperf (Geomean)
13x (faster)
1x
1x
OpenGL 4.1 Hardware Acceleration
Yes
No
No (Optimized for OpenGL 3.2)
DirectX Acceleration
Yes
No
Yes
Quad Buffered OpenGL Stereo
Yes
No
No
ECC Memory
Yes
Yes
No
Fast Double Precision
Yes
Yes
No
Dual Copy Engines
Yes
Yes
No
ISV Workstation Certification
Yes
No
No
ISV CUDA Certification
Yes
Yes
No
Energy Star Compliant
Yes
No
No
For more information visit: www.pny.com/quadro
PNY Technologies, Inc. 100 Jefferson Road, Parsippany, NJ 07054 | Tel 408 567 5500 | Fax 408 855 0680 Features and specifications subject to change without notice. The PNY logo are registered trademarks of PNY Technologies, Inc. All other trademarks are the property of their respective owners. © 2012 PNY Technologies, Inc. All rights reserved.
Multi-GPU Solutions
NVIDIA Quadro Multi-GPU Computing Solutions for workstations and large scale host installations NVIDIA Quadro 5000 Specifications Memory*
2.5 GB GDDR5
Memory Speed
1.5 GHz
Memory Interface
320-bit
Memory Bandwidth
120 GB/sec
CUDA Parallel Processing cores
352
ECC (Error Correcting Code)
Yes
High Performance Double Precision
Yes
Double Precision floating point performance (peak)
359.04 Megaflops
Single Precision floating point performance (peak)
718.08 Megaflops
Maximum Power Consumption
150 W
Physical Dimensions
4.376˝ (H) x 9.50˝ (L) Dual Slot DisplayPort (2) Dual link DVI –I 3-Pin Stereo Connector PCIe x16 Gen2
Connectors System Interface Thermal Solution Display support
Active Fansink
Warranty
1 2 2 2560x1600 CUDA, OpenGL 4.1, DirectX 11, and Shader Model 5.0 CUDA C/C++/Fortran, OpenCL, DirectCompute Toolkits. NVIDIA Parallel Nsight™ for Visual Studio 3 Year Warranty
NVIDIA Quadro 5000 Compute Pack
PNY part# VCGPU5K-PCK
Dual-Link DVI-I DisplayPort Active Display Support Maximum Display Resolution @ 60Hz APIs Software Development Tools
* Note: With ECC on, a portion of the dedicated memory is used for ECC bits, so the available user memory is reduced by 12.5%. (e.g. 3 GB total memory yields 2.625 GB of user available memory.)
NVIDIA Quadro Multi-GPU ADVANTAGE P 352 CUDA cores per Fermi-based Quadro GPU processor
P NVIDIA Parallel DataCache technology
P ECC protection
P Engineered for Massive data set uses
P Full Double Precision performance
P Long life availability and world-class support
P Full CUDA and OpenGL programmability
P 3 year warranty standard
P OpenGL quad buffered stereo support
P Quadro ISV and CUDA ISV certifications
P OpenGL, DirectX 11, Shader Model 5.0, CUDA, OpenCL, and DirectCompute support
P Enterprise level support and technical engineering
P NVIDIA GigaThread™ engine
PNY Technologies, Inc. 100 Jefferson Road, Parsippany, NJ 07054 | Tel 408 567 5500 | Fax 408 855 0680 Features and specifications subject to change without notice. The PNY logo are registered trademarks of PNY Technologies, Inc. All other trademarks are the property of their respective owners. © 2012 PNY Technologies, Inc. All rights reserved.
