More MAC G5folding performance data

Here is a thread I pulled from Team MAC OS X:

I tested my Dual 2 Ghz G5 against the new PIV 2.8 Ghz that Stanford uses for determining the points.

p732_BBA5_sd_h2o (gromacs) PIV 2.8 Ghz 5.89 hours • Dual 2 Ghz G5 (one processor) 5.57 hours

p563_BBA5_ext 12200 (gromacs) PIV 2.8 Ghz 5.02 hours • Dual 2 Ghz G5 (one processor) 4.88 hours

p733_villin_md (gromacs) PIV 2.8 ghz 3.93 hours • Dual 2 Ghz (one processor) 3.67 hours

It appears a Dual 2 Ghz G5 is more than twice as fast as a PIV 2.8 Ghz.

On a processor by processor comparision, the 2 Ghz G5 chip is faster than the PIV 2.8 Ghz chip.

I'm waiting for a Tinker to appear to see how it does on that.

I highlighted the controversial area for you all to reply.

Hooah

Well Macs have hardware bit-rotation making them # crunch better

Umm it says (1 processor)...what is that pertaining to?

Well I wonder what would happen if you overclocked that P4 or put it up against an AMD. The thing that might confuse people is that the only reason it is twice as fast is because it can use two chips to the P4's one chip. lol . I wouldn't mind seeing a database of real times for all chips and such for F@H.

is that with ht on, running one or 2 clients?

He is comparing his data to Stanfords Xeons. he is only runnning fah on one cpu His saying a dually g5 is 2x faster than a 2.8 p4 is the same as saying a dually amd is 2x faster than a single amd (or any dually) is not so. Just look at hyperthreading,we don't see 2x results, maybe 33-35% increase. The OS controls all plus bus speed, etc, etc.

I'm still going to lay my money down on a dually g5( hey fast) what are your dually g5 costs?

Hooah!

right, but is a p4 w/ ht on @ 2.8ghz better than a single mac g5 @ 2 ghz

Yes, like HT vs single AMD. HT ~33% increase over the single cpu, so you should see an increase over a single g5.

for it to be twice as fast wouldnt it have to do the work in half the time then listed?

i think i get where the twice as fast is coming from.
since his one processor is just as fast as a single 2.8, then if he used both processors he would be twice as fast as that 2.8
can he not do two instances and then check where his times are at against a HT enabled P4?