Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > comp.programming > #5151 > unrolled thread
| Started by | fir <profesor.fir@gmail.com> |
|---|---|
| First post | 2015-01-22 02:35 -0800 |
| Last post | 2015-01-22 02:35 -0800 |
| Articles | 1 — 1 participant |
Back to article view | Back to comp.programming
gpu computing fir <profesor.fir@gmail.com> - 2015-01-22 02:35 -0800
| From | fir <profesor.fir@gmail.com> |
|---|---|
| Date | 2015-01-22 02:35 -0800 |
| Subject | gpu computing |
| Message-ID | <91984130-e444-4ea3-b6fa-d42aeaeba3e1@googlegroups.com> |
have maybe someone hera has some experience with gpu general computing /code optimisation? how many faster loop throughput can be when implemented in cude/opencl? (i am suspecting from some reading that it could in the range of 10 times more throughtput than cpu core, but im not sure, i was only doing basic opencl kernel, got no much time to invest in real benchmarking or tests)
Back to top | Article view | comp.programming
csiph-web