Difference between revisions of "Talk:Fibonacci primes in parallel"
Jump to navigation
Jump to search
DonStewart (talk | contribs) |
(compiler options tried but gave nothing) |
||
| Line 8: | Line 8: | ||
[[User:DonStewart|dons]] 20:50, 26 May 2008 (UTC) |
[[User:DonStewart|dons]] 20:50, 26 May 2008 (UTC) |
||
| + | |||
| + | I have tried your options and have found not performance increase on my configuration (GHC 6.8.2 on Windows). Most probable, your better times are due to better hardware. |
||
| + | |||
| + | Have you increased the number in the line |
||
| + | <code> |
||
| + | fibprimes = catMaybes $ parBuffer 10 rnf maybeFibprimes |
||
| + | </code> |
||
| + | to improve CPU utilization on the quad-core machine? |
||
| + | |||
| + | [[User:Fed|Fed]] 17:57, 21 June 2008 (UTC) |
||
Latest revision as of 17:57, 21 June 2008
I note you're compiling your code with:
ghc --make –threaded parfibs.hs
As this is performance sensitive, it makes sense to turn on the optimiser:
ghc -O2 -fvia-C -optc-O2 --make –threaded parfibs.hs
dons 20:50, 26 May 2008 (UTC)
I have tried your options and have found not performance increase on my configuration (GHC 6.8.2 on Windows). Most probable, your better times are due to better hardware.
Have you increased the number in the line
fibprimes = catMaybes $ parBuffer 10 rnf maybeFibprimes
to improve CPU utilization on the quad-core machine?
Fed 17:57, 21 June 2008 (UTC)