|
February 2012 |
The NVIDIA Tesla K20X is a performance server compute GPU for applications such as scientific and engineering simulations. It provides over a teraflop of double precision performance allowing for detailed simulation which wouldn't be possible on a regular GPU.
| Key Features |
Applications |
> 2688 CUDA Cores
> 6GB GDDR5 Memory |
> CAE (structural analysis)
> CFD
> Financial computing
> Computational chemistry
> Computational physics
> Data analytics
> Satellite imaging
> Weather modelling
|
Specifications
| Specification |
Description |
|
Peak double precision floating point
performance
|
1.31 teraflops |
|
Peak single precision floating point
performance |
3.95 teraflops |
GPU chipset
|
GK110 |
| Number of CUDA cores |
2688 |
| Memory size (GDDR5) |
6 GB |
|
Memory bandwidth for board (ECC off)
|
250 GBytes/sec |
|
SMX
(streaming multiprocessor) design
|
Yes |
| Dynamic Parallelism |
Yes |
| Hyper-Q |
Yes |
Return to previous page
|
