On 6/21/07, Ingo Juergensmann <[EMAIL PROTECTED]> wrote:
On Thu, Jun 21, 2007 at 10:14:10PM -0700, Brian Morris wrote:
> watching the g3 closely with top while the quadra compiles. It looks > like compiling a routine the mac takes about 20 seconds, three seconds > on the g3 compiling and 2 seconds of the network (lights flash on my > hub-switch). So I figure simple arithmetic this is about the best one can > do. that is even if compile took 0 seconds that could only be another > 15% improvement (or with my g4 which is 3x the g3 then 10%). Uhm? If the m68k mac is needing 20s and the distcc setup needs 5 sec in total, then this is 1/4 of the time the m68k needed, or other way round 400% faster.
That's not what I meant ! Please consider this restatement: without distcc, the mac takes 75 sec/page to compile c++. with distcc, the complete compile takes 20 sec/page. of this 20 seconds the g3 spends 3 seconds compiling and the network transfer takes 2 seconds. thus the overhead for the mac is 15 seconds. thus given a g4 that was 3x faster than the g3, it could only possibly reduce this time to 18seconds (another 10%) the 3.5 times improvement of the compile time is great, but it is not even using more than 25% of the 250mhz g3. I wonder what the mac is doing during the 15 seconds, for a one page routine. statistically it is a very solid number, the example from the text i compiled about a hundred routines and the number was not varying at all as Iwatched it go for a while. the overhead is in the mac somewhere. not in the network per se (could be latency in the mac accessing the net though). I have plan to get a comm slot e/net card and then my q630 could take two cards (although the lcIII card in there now is need by q605), when my budget allows. then I could test using two network cards to increase the network speed. but yes that is not the bottleneck (or limiting factor). Brian -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
