GNU/Linux S5R3 "power users" App 4.35 available

th3
th3
Joined: 24 Aug 06
Posts: 208
Credit: 2208434
RAC: 0

RE: Finished and validated

Message 79517 in response to message 79516

Quote:
Finished and validated allright and it SEEMS faster, but I'm too much of a maths newb to figure out how much faster the app really is... can't make much sense of those formulas.


I cant see any 4.35 in the results from your T2060. Something wrong isnt quite right there.

Annika
Annika
Joined: 8 Aug 06
Posts: 720
Credit: 494410
RAC: 0

That would be because the

That would be because the results were already branded with 4.27. The last two were crunched with 4.35, though.

Bikeman (Heinz-Bernd Eggenstein)
Bikeman (Heinz-...
Moderator
Joined: 28 Aug 06
Posts: 3522
Credit: 753177284
RAC: 1188432

RE: That would be because

Message 79519 in response to message 79518

Quote:
That would be because the results were already branded with 4.27. The last two were crunched with 4.35, though.

All the submitted results are definitely still 4.27 (visible from the stderr.txt output), maybe the 4.35 results are not "reported" yet?

CU
Bikeman

Annika
Annika
Joined: 8 Aug 06
Posts: 720
Credit: 494410
RAC: 0

My bad. It WAS 4.27. Somehow

My bad. It WAS 4.27. Somehow I must have messed up my app_info.xml *bangs head on desk*

Erik
Erik
Joined: 14 Feb 06
Posts: 2815
Credit: 2645600
RAC: 0

RE: My bad. It WAS 4.27.

Message 79521 in response to message 79520

Quote:
My bad. It WAS 4.27. Somehow I must have messed up my app_info.xml *bangs head on desk*


Don't take it too harsh, I did the same thing on the 4.27 app info.xml from 4.24... didn't realize it for about two weeks. Dur...

Bikeman (Heinz-Bernd Eggenstein)
Bikeman (Heinz-...
Moderator
Joined: 28 Aug 06
Posts: 3522
Credit: 753177284
RAC: 1188432

RE: My bad. It WAS 4.27.

Message 79522 in response to message 79520

Quote:
My bad. It WAS 4.27. Somehow I must have messed up my app_info.xml *bangs head on desk*

You will know that you have 4.35 installed when you look at the runtimes ;-).

CU
Bikeman

Edit PS: Here's another speed comparison:

Core ("1" :-) ) Duo 1.6 GHz Windows, non-SSE, vs. Dual Pentium III, 1.4 GHz, Linux, SSE optimized 4.35. guess who wins :-)

http://einsteinathome.org/workunit/37166012

th3
th3
Joined: 24 Aug 06
Posts: 208
Credit: 2208434
RAC: 0

RE: Edit PS: Here's another

Message 79523 in response to message 79522

Quote:

Edit PS: Here's another speed comparison:

Core ("1" :-) ) Duo 1.6 GHz Windows, non-SSE, vs. Dual Pentium III, 1.4 GHz, Linux, SSE optimized 4.35. guess who wins :-)

http://einsteinathome.org/workunit/37166012


E2140 is based on Core 2, surprising its THAT slow, even with Windows and default app. E2000's are actually quite fast for their clockspeed. Mine is at 2.8GHz (and 0,05v undervolted at the same time), with 4.35 it goes below 18,500 sec. Thats a lot of performance per dollar if you ask me.

Zxian
Zxian
Joined: 23 Oct 06
Posts: 40
Credit: 5121474
RAC: 0

RE: E2140 is based on Core

Message 79524 in response to message 79523

Quote:
E2140 is based on Core 2, surprising its THAT slow, even with Windows and default app. E2000's are actually quite fast for their clockspeed. Mine is at 2.8GHz (and 0,05v undervolted at the same time), with 4.35 it goes below 18,500 sec. Thats a lot of performance per dollar if you ask me.

My E2160 (1.8GHz vs 1.6GHz) Has runtimes in the ~37,000 second range with the SSE optimized app. The advantage that the PIII has is that it's only got a 10-stage pipeline vs something in the 20's for the Core 2 Duo architechture. You've also got to realize that there are two independent cores in the E2140, so with an optimized application, you'd be pushing out twice the WUs of the PIII.

rroonnaalldd
rroonnaalldd
Joined: 12 Dec 05
Posts: 116
Credit: 537221
RAC: 0

RE: RE: E2140 is based on

Message 79525 in response to message 79524

Quote:
Quote:
E2140 is based on Core 2, surprising its THAT slow, even with Windows and default app. E2000's are actually quite fast for their clockspeed. Mine is at 2.8GHz (and 0,05v undervolted at the same time), with 4.35 it goes below 18,500 sec. Thats a lot of performance per dollar if you ask me.

My E2160 (1.8GHz vs 1.6GHz) Has runtimes in the ~37,000 second range with the SSE optimized app. The advantage that the PIII has is that it's only got a 10-stage pipeline vs something in the 20's for the Core 2 Duo architechture. You've also got to realize that there are two independent cores in the E2140, so with an optimized application, you'd be pushing out twice the WUs of the PIII.

...why twice the WUs? Bikemans P3s ia a dual-sockel-system. Means also two independent cores.

Annika
Annika
Joined: 8 Aug 06
Posts: 720
Credit: 494410
RAC: 0

My last WU was crunched

My last WU was crunched partially (only a small part at the end, though) by 4.35 and it has about the shortest crunch time ever, so I'm really curious what this app will be able to do. Going by the estimates (which should be fairly accurate by now, since this box has had so many S5R3 WUs already- except of course it doesn't fully take into account shorter crunch times with the new app) my first WU done only with 4.35 should be ready in about two hours' time.

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.