I discovered I had set the FSB jumpers incorrectly on an Athlon XP1600+. The net effect was that it was running 100Mhz fsb and an actual clock speed of 1050. It should have been 1400. In effect it wasnt going much better than the Duron 1100 that was in there before. Windows XP still chearfully reported the cpu to be an XP1600+. So I fixed the fsb.
Comparing the preformance of the Athlon XP1600+ (about 12,000 seconds a unit of seti) with the Sempron 2400+ (about 10,500 seconds a unit of seti). That doesn't sound right does it? The seti is about 15% faster, but the 'clock' numbers are 50% faster. So I thought maybe I've stuffed up the fsb of the Sempron as well?
Nope a quick check revealed a fsb of 333 and multiplier of 10, but hang on I thought the XP2400+ ran at an actual clock of 2000. 166 * 10 = 1666 or there abouts. That doesnt feel like much more than my XP1600+ at 1400. So whats going on there?
A quick search on the web Toms hardware says ".... value systems ... this seemed to be an opportunity to adjust the speed rating according to the clock speed of the competitor, namely Intel Celeron D." And futher that AthlonXP2000+ is the Sempron 2400+
Did I not know this already? I had read the artcilce before. Had I missed that point? Did AMD just magic 400Mhz from marketing?
2 comments:
so what speed have you got it running at now? I thought semprons were the same real mhz speed as well... is this the same for AMD64's? I doubt I'll be buying any anyway... (more likely to try buy pentium m)
sorry - what i meant was - was speed are you running your athlon-m at?
Post a Comment