homemade-jam's Profile
Reputation: 10
Tradesman
- Group:
- Contributor w/DIC++
- Active Posts:
- 1,298 (0.69 per day)
- Joined:
- 17-March 08
- Profile Views:
- 8,700
- Last Active:
Nov 19 2011 03:43 PM- Currently:
- Offline
Previous Fields
- Country:
- GB
- OS Preference:
- Linux
- Favorite Browser:
- FireFox
- Favorite Processor:
- Intel
- Favorite Gaming Platform:
- XBox
- Your Car:
- Volkswagen
- Dream Kudos:
- 150
Latest Visitors
-
Michael26 
06 Mar 2013 - 15:13 -
darek9576 
15 Jul 2012 - 20:31 -
strawhat89 
28 Jun 2012 - 23:29 -
gabehabe 
20 Mar 2012 - 09:40 -
brep 
24 Feb 2012 - 19:31
Posts I've Made
-
In Topic: start parallel programming
Posted 19 Nov 2011
OpenCL is probably better start with: it's platform independant across CPUs GPUs etc. With CUDA you're tied to nVidia's GPU. Both CUDA and OpenCL are easier to get into: CUDA has slightly better performance though the spread is tightening. Otherwise, both fairly similar and have the same learning curve though once you know one you can pick up the other fairly quickly. I prefer OpenCL myself since it is not owned by a single company.
For multithreading, boost has a very easy to use library. -
In Topic: Segmentation fault
Posted 19 Nov 2011
I don't think you can use the standard libraries in OpenCL. ie. How can you printf in a kernel? Just make it do something simple like i++ and store it in a target array. Or copy one of the kernels from the sdk.
My Information
- Member Title:
- Gabe's Nemesis
- Age:
- 22 years old
- Birthday:
- January 17, 1991
- Gender:
-
- Location:
- UK
- Interests:
- Computer Science, C++ Programming, Integrated systems, Public Key Encryption, Quantum Algorithms & Cryptosystems
- Years Programming:
- 5
Contact Information
- E-mail:
- Click here to e-mail me
- Website URL:
-
http://
Friends
|
|


Find Topics
Find Posts
View Reputation Given


|
Comments
gabehabe
22 Dec 2011 - 18:54