Linux port of Alex v8 code

Message boards : Number crunching : Linux port of Alex v8 code
Message board moderation

To post messages, you must log in.

1 · 2 · 3 · 4 . . . 5 · Next

AuthorMessage
Profile Crunch3r
Volunteer tester
Avatar

Send message
Joined: 15 Apr 99
Posts: 1546
Credit: 3,438,823
RAC: 0
Germany
Message 747017 - Posted: 2 May 2008, 16:51:58 UTC - in response to Message 746544.  
Last modified: 2 May 2008, 16:52:59 UTC

Very nice, my AMD Athlon X2 BE-2350 went from 11438s (2.4V) to 10362s (V8 SSE3) for a 63.98Cr WU.
I guess that's as fast as it gets...

Now we only need the Linux port :)


Linux Port ? I don't know but i made one on the fly...
AK V8 SSE3 on linux*
* running in vmware
first result should be done in about 20 mins or so...

Join BOINC United now!
ID: 747017 · Report as offensive
Profile jason_gee
Volunteer developer
Volunteer tester
Avatar

Send message
Joined: 24 Nov 06
Posts: 7489
Credit: 91,093,184
RAC: 0
Australia
Message 747018 - Posted: 2 May 2008, 16:53:48 UTC - in response to Message 747017.  

Very nice, my AMD Athlon X2 BE-2350 went from 11438s (2.4V) to 10362s (V8 SSE3) for a 63.98Cr WU.
I guess that's as fast as it gets...

Now we only need the Linux port :)


Linux Port ? I don't know but i made one on the fly...
AK V8 SSE3 on linux*
* running in vmware
first result should be done in about 20 mins or so...

Cool, I was hoping someone would pick it up quick ;D good 'o carry on!
"Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions.
ID: 747018 · Report as offensive
Profile Crunch3r
Volunteer tester
Avatar

Send message
Joined: 15 Apr 99
Posts: 1546
Credit: 3,438,823
RAC: 0
Germany
Message 747032 - Posted: 2 May 2008, 17:26:58 UTC - in response to Message 747018.  

First result is done...



Join BOINC United now!
ID: 747032 · Report as offensive
Profile zoom3+1=4
Volunteer tester
Avatar

Send message
Joined: 30 Nov 03
Posts: 65709
Credit: 55,293,173
RAC: 49
United States
Message 747047 - Posted: 2 May 2008, 17:47:20 UTC - in response to Message 747032.  
Last modified: 2 May 2008, 17:49:33 UTC

First result is done...


How quick did the WU crunch, Crunch3r? As in how much time did the WU take to be crunched?
The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's
ID: 747047 · Report as offensive
Profile Crunch3r
Volunteer tester
Avatar

Send message
Joined: 15 Apr 99
Posts: 1546
Credit: 3,438,823
RAC: 0
Germany
Message 747048 - Posted: 2 May 2008, 17:51:03 UTC - in response to Message 747047.  
Last modified: 2 May 2008, 17:53:00 UTC

First result is done...


How quick did the WU crunch, Crunch3r? As in how much time did the WU take to be crunched?

CPU time (sec)  claimed credit
602.87  	20.05   	 pending
3,249.19 	63.98 	         pending
2,432.77 	54.24 	         pending


Link to all results so far...

Join BOINC United now!
ID: 747048 · Report as offensive
Profile jason_gee
Volunteer developer
Volunteer tester
Avatar

Send message
Joined: 24 Nov 06
Posts: 7489
Credit: 91,093,184
RAC: 0
Australia
Message 747050 - Posted: 2 May 2008, 17:54:54 UTC
Last modified: 2 May 2008, 17:55:06 UTC

Build #42? ... The answer to the question of life the universe and everything?
"Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions.
ID: 747050 · Report as offensive
Profile Andy Lee Robinson
Avatar

Send message
Joined: 8 Dec 05
Posts: 630
Credit: 59,973,836
RAC: 0
Hungary
Message 747051 - Posted: 2 May 2008, 17:55:30 UTC - in response to Message 747032.  

First result is done...


Cool!
I struggled for days trying to get it to compile statically with Intel's latest IDE/IPP.

Well done!
ID: 747051 · Report as offensive
Profile Crunch3r
Volunteer tester
Avatar

Send message
Joined: 15 Apr 99
Posts: 1546
Credit: 3,438,823
RAC: 0
Germany
Message 747054 - Posted: 2 May 2008, 18:02:43 UTC - in response to Message 747051.  

First result is done...


Cool!
I struggled for days trying to get it to compile statically with Intel's latest IDE/IPP.

Well done!


why ? except for some typos like #include "analysefuncs.h" where it should have been "analyseFuncs.h" and some other stuff that i ripped off like the ipp_static_test.h" or so... all seems to be fine...

(of course the cpuid stuff is gone as well, i made a library out of the 2.2b/2.4 cpuid function and linked that in ...

Anyhow, jason got the binary and some of you folks at lunatics that run linux natively( not in vmware) should do some tests if it's as fast as the windows binary...

I'm still unable to connect to lunatics.kwsn.net ... otherwise i would have posted there myself :(








Join BOINC United now!
ID: 747054 · Report as offensive
Profile Crunch3r
Volunteer tester
Avatar

Send message
Joined: 15 Apr 99
Posts: 1546
Credit: 3,438,823
RAC: 0
Germany
Message 747055 - Posted: 2 May 2008, 18:03:34 UTC - in response to Message 747050.  
Last modified: 2 May 2008, 18:05:48 UTC

Build #42? ... The answer to the question of life the universe and everything?


LOL ... well that's what was in the analysefuncts.cpp ... i didn't change that ...

Join BOINC United now!
ID: 747055 · Report as offensive
Profile zoom3+1=4
Volunteer tester
Avatar

Send message
Joined: 30 Nov 03
Posts: 65709
Credit: 55,293,173
RAC: 49
United States
Message 747061 - Posted: 2 May 2008, 18:13:37 UTC - in response to Message 747048.  

First result is done...


How quick did the WU crunch, Crunch3r? As in how much time did the WU take to be crunched?

CPU time (sec)  claimed credit
602.87  	20.05   	 pending
2,432.77 	54.24 	         pending
3,249.19 	63.98 	         pending


Link to all results so far...

Let's see that's around:

10.04 minutes for the 20.05
40.54 minutes for the 54.24
54.15 minutes for the 63.98

Not bad, Doesn't sound too bad at all.
The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's
ID: 747061 · Report as offensive
Profile Crunch3r
Volunteer tester
Avatar

Send message
Joined: 15 Apr 99
Posts: 1546
Credit: 3,438,823
RAC: 0
Germany
Message 747065 - Posted: 2 May 2008, 18:19:23 UTC - in response to Message 747061.  
Last modified: 2 May 2008, 18:19:46 UTC


Let's see that's around:

10.04 minutes for the 20.05
40.54 minutes for the 54.24
54.15 minutes for the 63.98

Not bad, Doesn't sound too bad at all.


i have no clue if that's good or not...

i'm running that test app in vmware using 2 cpus of my V8 rig which is a dual quad core xeon 5365 @ 3 ghz
(on the windows part it's cruching 6 riesel sieve wus at once)

Join BOINC United now!
ID: 747065 · Report as offensive
Profile jason_gee
Volunteer developer
Volunteer tester
Avatar

Send message
Joined: 24 Nov 06
Posts: 7489
Credit: 91,093,184
RAC: 0
Australia
Message 747068 - Posted: 2 May 2008, 18:30:00 UTC
Last modified: 2 May 2008, 18:37:44 UTC

I've posted in Lunatics development areas for anyone running Linux NATIVE to test.

As it happens my Ubuntu VM is running stock at the moment, took me a a while to figure that out, now I need to figure out how to install opt app in Linux :S
[Calling all Linux Dudes .... Impending Emergency Eject :S]

Jason
"Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions.
ID: 747068 · Report as offensive
Profile Gecko
Volunteer tester
Avatar

Send message
Joined: 17 Nov 99
Posts: 454
Credit: 6,946,910
RAC: 47
United States
Message 747079 - Posted: 2 May 2008, 18:58:31 UTC - in response to Message 747048.  
Last modified: 2 May 2008, 19:01:41 UTC

First result is done...


How quick did the WU crunch, Crunch3r? As in how much time did the WU take to be crunched?

CPU time (sec)  claimed credit
602.87  	20.05   	 pending
3,249.19 	63.98 	         pending
2,432.77 	54.24 	         pending


Link to all results so far...


Wooaahhh!....CRASH!!! (as I fall out of chair) : 0
So, you're running 2 instances w/ RS on the other 6?
Those times are freakishly fast compared to times on: [#10 Bad to the Bone]
54.24
63.98

The difference is so massive there's got to be another factor in this.
Reported time for VLAR is even @25% quicker than #1 MAC? [63.98]
Running in VMware quicker than running native? (how does this make sense?)

Ideas to explain these times???
ID: 747079 · Report as offensive
Profile Crunch3r
Volunteer tester
Avatar

Send message
Joined: 15 Apr 99
Posts: 1546
Credit: 3,438,823
RAC: 0
Germany
Message 747082 - Posted: 2 May 2008, 19:05:24 UTC - in response to Message 747079.  
Last modified: 2 May 2008, 19:20:51 UTC


Wooaahhh!....CRASH!!! (as I fall out of chair) : 0
So, you're running 2 instances w/ RS on the other 6?


exactly...

EDIT
here's a picture


Those times are freakishly fast compared to times on: [#10 Bad to the Bone]
54.24
63.98

The difference is so massive there's got to be another factor in this.
Reported time for VLAR is even @25% quicker than #1 MAC? [63.98]
Running in VMware quicker than running native? (how does this make sense?)
Ideas to explain these times???


actually from what i've seen in the past, running in vmware is slower than running it natively.

Join BOINC United now!
ID: 747082 · Report as offensive
Profile Gecko
Volunteer tester
Avatar

Send message
Joined: 17 Nov 99
Posts: 454
Credit: 6,946,910
RAC: 47
United States
Message 747090 - Posted: 2 May 2008, 19:31:33 UTC - in response to Message 747082.  
Last modified: 2 May 2008, 19:32:13 UTC


Wooaahhh!....CRASH!!! (as I fall out of chair) : 0
So, you're running 2 instances w/ RS on the other 6?


exactly...

EDIT
here's a picture


Very nice & I don't just mean the set-up ; >)

actually from what i've seen in the past, running in vmware is slower than running it natively.


Well, that is what one should expect so at least my sanity is still intact :P
Doesn't the speed surprise you?
Will it do that running the aps on all 8 cores? (still scratching head...)
ID: 747090 · Report as offensive
Profile hiamps
Volunteer tester
Avatar

Send message
Joined: 23 May 99
Posts: 4292
Credit: 72,971,319
RAC: 0
United States
Message 747093 - Posted: 2 May 2008, 19:34:29 UTC - in response to Message 747017.  

Very nice, my AMD Athlon X2 BE-2350 went from 11438s (2.4V) to 10362s (V8 SSE3) for a 63.98Cr WU.
I guess that's as fast as it gets...

Now we only need the Linux port :)


Linux Port ? I don't know but i made one on the fly...
AK V8 SSE3 on linux*
* running in vmware
first result should be done in about 20 mins or so...


I don't know why....But I am not suprised...Now if you could do a sse2 on the fly...Just joking, but right on...
Official Abuser of Boinc Buttons...
And no good credit hound!
ID: 747093 · Report as offensive
Profile Crunch3r
Volunteer tester
Avatar

Send message
Joined: 15 Apr 99
Posts: 1546
Credit: 3,438,823
RAC: 0
Germany
Message 747094 - Posted: 2 May 2008, 19:37:28 UTC - in response to Message 747090.  
Last modified: 2 May 2008, 20:37:05 UTC


Very nice & I don't just mean the set-up ; >)

got no clue what your talking about :P

actually from what i've seen in the past, running in vmware is slower than running it natively.

Well, that is what one should expect so at least my sanity is still intact :P
Doesn't the speed surprise you?


no... cuz i don't have a reference to windows... i haven't been running sah for months because the credit is poor.


Will it do that running the aps on all 8 cores? (still scratching head...)


I don't know... i could reboot the machine an try running the app on fedora core 8 (x64) to see if it behaves the same way there while running 8 instances of sah..


EDIT
of course it had to happen ... first wu got validated and i got pared up with a host running boinc 4.45 ... instead og getting 20 credits i got 8 :(

Join BOINC United now!
ID: 747094 · Report as offensive
Profile Gecko
Volunteer tester
Avatar

Send message
Joined: 17 Nov 99
Posts: 454
Credit: 6,946,910
RAC: 47
United States
Message 747136 - Posted: 2 May 2008, 20:39:10 UTC - in response to Message 747094.  
Last modified: 2 May 2008, 20:41:37 UTC


I don't know... i could reboot the machine an try running the app on fedora core 8 (x64) to see if it behaves the same way there while running 8 instances of sah..


Your 54.36=2415 secs is substantially quicker than Francois running JDWale ap + Vista on this WU: 54.33=3452 secs
(best compo I could find w/ his rig)

....except he is running a Hapertown X9775 3.2GHz OC'd to 4.0Ghz.

The JDWhale ap is within a few % of the new release ap for -doze.
F' is running 1Ghz more clock and 50% more cache of the next-gen cpu.

The linux port on your rig is somehow still spanking it....ALOT.
ID: 747136 · Report as offensive
Profile JDWhale
Volunteer tester
Avatar

Send message
Joined: 6 Apr 99
Posts: 921
Credit: 21,935,817
RAC: 3
United States
Message 747139 - Posted: 2 May 2008, 20:47:07 UTC - in response to Message 747136.  


The linux port on your rig is somehow still spanking it....ALOT.


No doubt. But the compare is not complete without running 8 cores loaded native Linux. You might check out UL1's X5482 rig, though still slower on a per/core basis the clock is much closer to Crunch3rs rig.
ID: 747139 · Report as offensive
Profile [KWSN]John Galt 007
Volunteer tester
Avatar

Send message
Joined: 9 Nov 99
Posts: 2444
Credit: 25,086,197
RAC: 0
United States
Message 747141 - Posted: 2 May 2008, 20:53:36 UTC

Dang...and I just blew my Linux box away and reinstalled XP last night...I would have been able to test a Q6600@2.997gHz (best OC for that board as of now) running 4 SETI...

Well, I did get a Q9300 this afternoon. Might have to try it at stock speed before giving it over to XP.


Clk2HlpSetiCty:::PayIt4ward

ID: 747141 · Report as offensive
1 · 2 · 3 · 4 . . . 5 · Next

Message boards : Number crunching : Linux port of Alex v8 code


 
©2024 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.