16 Apr
2011
16 Apr
'11
2:19 p.m.
On 4/13/11 12:31 PM, Richard wrote:
Instead of increasing
single core speeds, CPUs are incorporating more processor cores. In
order to automagically get speedups now your code has to be
parallelizable by the compiler tool chain. Most code isn't structured
this way.
Autoparallelizing compilers have been around for a couple of decades
now. Of course one needs to analyze their feedback and occasionally
restructure loops and such to make them parallelizable, but such
compilers often give excellent feedback about it.
Explicitly threading code isn't the only way to take advantage of
multiple processors.
-Dave
--
Dave McGuire
Port Charlotte, FL