Separate names with a comma.
Discussion in 'MacBytes.com News Discussion' started by MacBytes, Jul 17, 2003.
Link: PowerMac G5 Dual 2GHz XBench Results
Posted on MacBytes.com
I'm happy with that, my PowerBook 12" gets 75.46 with Panther DP1, and that scores a lot lower than Jaguar in many areas (which Apple will hopefully fix).
Not the amazing gulf I expected, but then i guess when XBench is updated and Panther is released, there may well be a greater difference.
They're also only using 2 Gigs of RAM - if they use more I'm sure on some things you'd see a great improvement.
But it does say something if you have to spend a ton more money to get the RAM to get the speed they advertise....
I'd like to see the comparison on a dual 1.4 G4.
Same here, anyone?
That would be the site... IF they keep it ONLINE!!
I'm comparing with a dual 1Ghz G4,
the G4 actually kicks the dual 970 in GCD Recursion and AltiVec Basic.
G4 scores 125.99 in GCD and 126.08 in Altivec to compare with 135.68 for GCD and 52.41(!) for Altivec for the G5.
From these numbers the G4 is almost 5 times as fast per GHz as the G5. Looks like something isn't optimized here...
Incidentally in floating points the dual 2GHz G5 beats the dual 1GHz G4 and is approx 3.3 times as fast, or 1.6 times as fast per GHz.
Anyway, the poor AltiVec and GCD seems to account for the low CPU score.
It also shows different efficiency in the thread test. In Computation the G5 is only faster by 30% whereas in Lock Contention it is 145% faster.
It does well in the memory test (318 to 132) but inexplicably "System Fill" is only at 183 to 268 of the dual 1GHz G4. That's a whole lot slower. Odd again.
Graphics is better by 60% only, but I guess this depends a lot on the graphics card.
The disk-test only scores slightly higher with negative surprises, like Uncached Read for some reason being half that of the G4.
i've had ebough G5 benchmark tests. wait untill they come out and see how they do in real real world apps.