S41.xx Observation Thread

groundhog
groundhog
Joined: 11 Feb 05
Posts: 11
Credit: 69,280
RAC: 0

Great work, akosf, many

Great work, akosf, many thanks to you.

My PIII 866 is a lot faster with S41.06 than an AthlonX2 4200+ with standard application on the same WU (me ca 7500 sec, the other ca 13500 sec.

My Athlon XP2800+ gains about 600 sec/WU compared to the latest D40 (now about 3000 sec/wu instead of 3600 sec/WU).

We drain einstein off WU's ;-).

greetings groundhog


MecRipper
MecRipper
Joined: 20 Feb 05
Posts: 5
Credit: 3,063,368
RAC: 0

My result with "Pentium 4 -

My result with "Pentium 4 - 630" Hyperthreading Enable :

S41.06 = 10,221.00 sec
S40 = 8,944.43 sec

for me the best version is S40 ;)
I wait new version of "U41" :)

bye and good work Akosf!! ;)

ps.
sorry for my english... i am italian ;)

Akos Fekete
Akos Fekete
Joined: 13 Nov 05
Posts: 561
Credit: 4,527,270
RAC: 0

RE: My result with "Pentium

Message 29980 in response to message 29979

Quote:

My result with "Pentium 4 - 630" Hyperthreading Enable :

S41.06 = 10,221.00 sec
S40 = 8,944.43 sec

It isn't good news.
I'm wondering on the Pentium4 processors they how dislike the mixture of FPU/SSE code.
I will try to find a solution to optimize for these cores.
(I will bring home a P4 laptop or something else...)

ExtraTerrestrial Apes
ExtraTerrestria...
Joined: 10 Nov 04
Posts: 770
Credit: 332,098,916
RAC: 625,727

RE: I'm wondering on the

Message 29981 in response to message 29980

Quote:
I'm wondering on the Pentium4 processors they how dislike the mixture of FPU/SSE code.

P4 doesn't like FPU code and I guess it doesn't like work in general ;)

MrS

Scanning for our furry friends since Jan 2002

anders n
anders n
Joined: 29 Aug 05
Posts: 123
Credit: 1,656,300
RAC: 0

I tried S41.07 and got this

I tried S41.07 and got this result

http://einsteinathome.org/task/27831199

Anders n

Akos Fekete
Akos Fekete
Joined: 13 Nov 05
Posts: 561
Credit: 4,527,270
RAC: 0

RE: I tried S41.07 and got

Message 29983 in response to message 29982

Quote:

I tried S41.07 and got this result

http://einsteinathome.org/task/27831199

Could you download it again?

anders n
anders n
Joined: 29 Aug 05
Posts: 123
Credit: 1,656,300
RAC: 0

RE: RE: I tried S41.07

Message 29984 in response to message 29983

Quote:
Quote:

I tried S41.07 and got this result

http://einsteinathome.org/task/27831199

Could you download it again?

Ok it is running now.

Anders n

[AF>France] paulatreides
[AF>France] pau...
Joined: 22 Apr 05
Posts: 4
Credit: 36,297
RAC: 0

@aksof 1°)Are you about to

@aksof

1°)Are you about to deplete your ideas of optimisation of the code.
Have you still some idea in your mind ?

2°)It could be interesting (for me of course :-) and i suppose for others)
to resume here (or in de pdf) all your techniques of optimisation for each client.(add eventually links on the web that inspire you)
You could start by copying part of the program (if you are autorised to...) that can be optimised... then perharps remember us main caracteristic of processor...
and finally dévelop procedures involved.

I appreciate very much your work ! :-)

Perhaps you will find i am very demanding. So consider part 2°) as only a query :-)

Akos Fekete
Akos Fekete
Joined: 13 Nov 05
Posts: 561
Credit: 4,527,270
RAC: 0

RE: 1?)Are you about to

Message 29986 in response to message 29985

Quote:
1?)Are you about to deplete your ideas of optimisation of the code.
Have you still some idea in your mind?

Of course. I have always newer and newer ideas, but I have also lots of other ideas from other areas of the life. And I would like to try out them also. :)

Quote:
2?)It could be interesting (for me of course :-) and i suppose for others) to resume here (or in de pdf) all your techniques of optimisation for each client. (add eventually links on the web that inspire you)

Well, I always try to find out new things, but of course I use lots of documentations and i have lot of experience. They are also important. I don't do a list because it is always changing and grows.

Quote:
You could start by copying part of the program (if you are autorised to...) that can be optimised... then perharps remember us main caracteristic of processor... and finally d鶥lop procedures involved.

I'm sorry, but I don't understand it exactly. It isn't to difficult to decode a program and exchange some parts. I use three softwares for it. Borland Turbo Debugger, Flat Assembler and Volkov Commander. As far as i know these softwares are free to everybody, so everybody can do omptimize any kind of executables.

J Langley
J Langley
Joined: 30 Dec 05
Posts: 50
Credit: 58,338
RAC: 0

Have you thought about doing

Message 29987 in response to message 29986

Have you thought about doing a version of E@H for GPUs? (Is that even possible with the current generation, or will we have to wait for DirectX 10 before this becomes realistic?)

Comment viewing options

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