SETI benchmark for XP3200

[Assimilator1]

THG have done a review of the XP3200 including a benchmark using SETI

Btw in the review the XP3200 doesn't fare well against the P4 3GHz despite its 200MHz FSB

 

[CyGoR]

LOL, even the 3.06 @ 533Mhz FSB owns the new "flagship" of AMD..
They need to hurry with the 64bit cpu to stay close to Intel..
And how about that ridiculous rating of 3200+
I'd rather buy a 2.53Ghz PIV... The new Barton is even $50 more expensive than Intel's flagship..
How weird...

 

[Wiz]

so what's it cost for a mobo/cpu/ram minimum to get those 1:29 WU's?

 

[MoFunk]

OK I am going to have to read through that. My 1700XP with a 215 FSB is doing 2:10 average! I would have to thing that in the real world the barton would do better then the 1700 even with the lower FSB.

 

[lane42]

I wonder what my 2700xp @ 2.5 Gig would do on an 8RDA Board. 2:22 on a KT400

 

[gistech1978]

Originally posted by: lane42
I wonder what my 2700xp @ 2.5 Gig would do on an 8RDA Board. 2:22 on a KT400

well with my 8rda/2100+ @ 2000/.417 AR.... i wanna say its around 2:30...however i will probably be editing this thread when i get home.

oh lets say you can expect maybe a 10-15 minute decrease...but then again im basing this on my ability to pull things out of my...


EDIT: i was right...2:30.

get an 8rda(+)...i love mine!

 

[Soggysocks]

Yes

Bump

 

[lobadobadingdong]

That doesn't quite make sense unless HT is helping a lot with Seti somehow. 2800+ bests the 2.8Ghz P4 by 9min. the 3000+and 3200+ use the same PR rating system but the 3.0&3.06Ghz P4's stomp on the 3000&3200+'s, there has to be something there, either HT or an advantage of motherboard platforms.

 

[mechBgon]

I think THG is witholding a few details there. I showed jhites how to run the Ars Technica standard WU on SETI@Home on his i875 system with a 3.0C and tweaked memory timings.

• For a single instance of the client, his system completed the standard 0.417 WU in 2:10.
• For two instances, his system completed in just a hair under 3:00 for both WUs, or 1:30 per WU effectively. Obviously the i875's flavor of Hyperthreading is working well here... but THG makes no mention of this little trick at all, AFAICS.

Here's jhites' thread for those interested: thread

I made a "SETIbench" test that includes a batch file to run the client with the -stop_after_process switch, the Ars Technica 0.417 WU, a link to where the client can be downloaded from Berkeley, and a link to the AnandTech S@H FAQ. I think this may be hosted at AnandTech soon, so the fiends in the CPU/Overclocking Forum can put their systems through the wringer For those interested in seeing the SETIbench, it can be found here (Zip format).

 

[Assimilator1]

Yeah its true those top end P4 SETI WU times are odd.

Btw doesn't it just suck that the XP3200 only has a 34MHz speed increase over the XP3000?

.I bet the price increase is way over the clock speed increase!

I've posted a querey thread about it over at THGs DC forum

 

[Intelligence3]

Okay, so I'm waiting for the 2.6c (2.6 GHz @ 800MHz FSB) processors next week for my new system. But now that I see how fast the 3.00c processes WUs, I almost want to spend the extra couple hundred $$ for the speed. Not for gaming performance, for SETI times.

SETI, it's a sickness, I tell ya. :evil:

 

[Assimilator1]

lol ,bear in mind there's something odd about the P4 3+GHz WU times ,I seem to remember the same WU times were shown for the dual P4 vs Opteron tests!