When I used to crunch Seti, I had a RAC of over 2000 with just my E6400. Now with Rosetta, I'm getting ~700 RAC. Someone in an earlier post said that lack of cache and a slow FPU were causes of this in Celerons, but with the latest core on the market, how could there be a ~60% difference for me?

If you were running an optimized app at seti, then that might explain the difference. It cuts run times substantially, and thereby increases "work done", and thereby RAC.

Try running Seti with the stock app to get a comparison.

If you were running an optimized app at seti, then that might explain the difference. It cuts run times substantially, and thereby increases "work done", and thereby RAC.

Try running Seti with the stock app to get a comparison.

I think is was 1800 beofore I started on an optimized app. Did I mention I'm overclocked quite a bit :P

i think a few months ago you would have got around 1000-1100 for an Allendale/Conroe @ 3.2GHz, but it has dropped quite a lot recently. Hopefully it'll get sorted out soon as there are lots of threads popping up on the subject.

Cross project parity isn't straight forward though because of things like extensions (which Rosetta can't use ATM) and different fpu/integer and cache requirements.

---

John, I posted a comparison at Einstein in this post. Which shows how my machines are seeing things. IMO Einstein and Rosetta are VERY close in "granted credit/hour", with the stock seti being minimally lower than both. Ofcourse, Einstein is using S5R2 now. I don't have any data using Simons 2.2b app. The 2.0 is close though. I suspect it(2.2b) to be a little more efficient. How much more, I can't say until I collect the data. I'll be doing that after both Einstein and Rosetta catch up to my seti total score (see sig below).

---

When I used to crunch Seti, I had a RAC of over 2000 with just my E6400. Now with Rosetta, I'm getting ~700 RAC. Someone in an earlier post said that lack of cache and a slow FPU were causes of this in Celerons, but with the latest core on the market, how could there be a ~60% difference for me?

heya,
i have noticed a vast difference in credits quoted to me within rosetta itself.
i've had cpu time of 10,700 with a claimed credit of 50.38 and a granted credit of 38.96. then i've had cpu time of aprox the same time a claimed credt of 43 and granted credit of 56. the only consistent thing with rosetta for me is the inconsistency of credit. i've kept the same settings the whole time except for a one day experiment with climatechange. i have a mac pro 4 core system with four gigs of ram and a setting of 100 percent cpu usage on idle. swap file is enabled blah blah blah but it's still inconsistent. arg. hopefully a more consistant system will be employed.
ken

Well, I found something interesting.

My E6400 has a RAC of 500 at the moment. That's 2 cores, so 250 each.

My proliant server has a RAC of 200. That's 4 P2 cores at 50 each.

because the P2s are at 500Mhz, each core has a RAC/Mhz ratio of 1:10
My E6400 is at 3000 Mhz, so it's ratio is 1:12

So the ratio for the P2s is better, even though it's FSB, RAM... Everything is slower.