D41.xx Observation Thread

Akos Fekete
Akos Fekete
Joined: 13 Nov 05
Posts: 561
Credit: 4527270
RAC: 0
Topic 191145

D41 uses 3DNow! and MMX instructions to improve the calculation performance.

3DNow!/MMX compatibles: K6-2, K6-III and newer AMD processors, Ciryx3, Samuel-II, Ezra, Nehemiah, Winchip2

Honza
Honza
Joined: 10 Nov 04
Posts: 136
Credit: 3332354
RAC: 0

D41.xx Observation Thread

Completed 2 WUs (r_1362) in about 2660 sec using D41 on AMD X2, no erros so far
Will put D40 app to compare speed with D41; got 3 more WUs to go...

Honza
Honza
Joined: 10 Nov 04
Posts: 136
Credit: 3332354
RAC: 0

Another 2 WUs with previous

Another 2 WUs with previous app version D40 took 3470 sec each.
D41 looks like 0.77 time-to-complete compating D40.
I would call that a way "to improve the calculation performance" :-)

Good job, akosf.

Hope to see results from other users/machines soon...

Higgi
Higgi
Joined: 4 Jun 05
Posts: 19
Credit: 8168519
RAC: 0

honza wrote: Hope to see

Message 29734 in response to message 29733

honza wrote:

Hope to see results from other users/machines soon...


I started to test it on my Opteron 144@2,32GHz. I will psot the result here as soon as I have one ;o)

AnRM
AnRM
Joined: 9 Feb 05
Posts: 213
Credit: 4346941
RAC: 0

Loaded an Athlon 3000+ and so

Loaded an Athlon 3000+ and so far it looks like D41.12 will be a rocket! Will load the dual cores as well. Akosf, fantastic!
>First result indicates ~12% speedup over S40.12!! No validation yet....Cheers, Rog.
(I hope they raise the MDQ above 32/day soon!!)
>First result on a 1.8MHz Duron indicates ~15% speed increase over S40.12...:)

TauCeti
TauCeti
Joined: 1 Apr 05
Posts: 16
Credit: 1336558
RAC: 0

Well, on my Athlon 64 3000+

Well, on my Athlon 64 3000+ it looks like that D41.12 is about 15-20% faster then S40.12(!) Awesome! Thanks akosf!

Honza
Honza
Joined: 10 Nov 04
Posts: 136
Credit: 3332354
RAC: 0

Another WU with D41 finish so

Another WU with D41 finish so my previous evaulation was not a measurement error - you can expert about +23% boots over D40 on X2 machine.

Yeah, it's a rocket...I've even lost comparation to oficial tortoise, D41 is a real cruncher, not a crawler :-)

EDIT: Based on first WUs completed and comparing it with other resultsIDs, D41 is almost 3 x faster over Linux 4.55 app (AMD X2).

Another good news - first WU with D41 got validated.

Bad news - speed of 45 minutes per 'long' WU gives exactly 32WU/day/core which makes my machine to meet daily quota and get dry on Einstein.

AnRM
AnRM
Joined: 9 Feb 05
Posts: 213
Credit: 4346941
RAC: 0

RE: Another good news -

Message 29738 in response to message 29737

Quote:

Another good news - first WU with D41 got validated.


>Great news, Honza......thanks, Rog.

Higgi
Higgi
Joined: 4 Jun 05
Posts: 19
Credit: 8168519
RAC: 0

RE: honza wrote: Hope to

Message 29739 in response to message 29734

Quote:
honza wrote:

Hope to see results from other users/machines soon...


I started to test it on my Opteron 144@2,32GHz. I will psot the result here as soon as I have one ;o)


First result: 1432 cycles ("long" WU) - 43 minutes! Last instaled opti version, S40.12 - 51 minutes. Great job! I am waiting for validation. Bad news - I reached the daily quota, :(((.

arashi
arashi
Joined: 12 Jun 05
Posts: 37
Credit: 41761784
RAC: 0

it's time to try the "fake"

it's time to try the "fake" CPU feature from BoincStudio, but consider that BS is in an early beta state.
BoincStudio

anders n
anders n
Joined: 29 Aug 05
Posts: 123
Credit: 1656300
RAC: 0

It is running faster. But my

It is running faster. But my first result was invalid.

http://einsteinathome.org/task/27154431

Anders n

Comment viewing options

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