|
[Sponsors] |
March 9, 2011, 03:21 |
GPU computing
|
#1 |
Member
lalupp
Join Date: Jul 2010
Location: India
Posts: 47
Rep Power: 16 |
I would like to apply GPU computing to my existing parallel CFD code (incompressible) which is written in C. How do I start. I heard that there are two platforms exists, CUDA and AMD FUSION which is preferred in terms of simplicity and speed
Where do I start ,What are the books, thesis available online Thanks in advance |
|
March 9, 2011, 04:42 |
|
#2 |
Senior Member
Raashid Baig
Join Date: Mar 2010
Location: Bangalore, India
Posts: 135
Rep Power: 16 |
Hi,
Do you have access to both Tesla machines and AMD Fusion machines ? Tesla machine cost upwards of 10,000 US Dollars here in India. What I have heard from my friends who have worked with CUDA is that you will end up rewriting almost the entire parallelization part of your solver. So it will be a long and non-trivial process. |
|
March 11, 2011, 02:12 |
|
#3 |
Member
lalupp
Join Date: Jul 2010
Location: India
Posts: 47
Rep Power: 16 |
Thanks
To start with Can I use lower end CUDA enabled graphics cards like QUARDO/GTX for scientific computing ? |
|
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
New Nvidia gpu aimed at gpgpu | bmeagle | OpenFOAM | 0 | November 9, 2006 10:41 |
CFD algorithm design with reconfigurable computing | Beatríz Navarro | CFX | 4 | June 27, 2006 05:39 |
CFD algorithm design with reconfigurable computing | Beatríz Navarro | Main CFD Forum | 1 | June 25, 2006 15:25 |
CFD algorithm design with reconfigurable computing | Beatríz Navarro | Fidelity CFD | 0 | June 24, 2006 19:23 |
CFD algorithm design with reconfigurable computing | Beatríz Navarro | FLUENT | 0 | June 24, 2006 19:22 |