New binary to test on beta

Message boards : Number crunching : New binary to test on beta
Message board moderation

To post messages, you must log in.

Previous · 1 · 2 · 3 · 4 · 5 . . . 7 · Next

AuthorMessage
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 13164
Credit: 1,160,866,277
RAC: 1,873
United States
Message 1871047 - Posted: 4 Jun 2017, 3:50:03 UTC - in response to Message 1871044.  

Since I never ran the 8.05 or 8.04 one, I am comparing it to the the AVX r3330 one since that is the only one I have ever run.
Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)
ID: 1871047 · Report as offensive
Profile Zalster Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 27 May 99
Posts: 5517
Credit: 528,817,460
RAC: 242
United States
Message 1871048 - Posted: 4 Jun 2017, 3:51:28 UTC - in response to Message 1871047.  

There is 1 8.05 that you did run Keith, took 23 min 20 sec, lol
ID: 1871048 · Report as offensive
Grant (SSSF)
Volunteer tester

Send message
Joined: 19 Aug 99
Posts: 13736
Credit: 208,696,464
RAC: 304
Australia
Message 1871049 - Posted: 4 Jun 2017, 4:02:09 UTC - in response to Message 1871047.  
Last modified: 4 Jun 2017, 4:10:45 UTC

Since I never ran the 8.05 or 8.04 one, I am comparing it to the the AVX r3330 one since that is the only one I have ever run.

AVX isn't a good comparison, as it's only good for CPUs and OSs that support it.
This application will be for general release, so maximum compatibility.
If the new application is as good as Raistmer feels it is, it's run time will make the present stock ones look bad.

So if you think the 8.06 is slow, wait till you run a few 8.05s to completion as a comparison.
:-)

EDIT-
I'm looking at 3hrs 25 to 3hrs 40 for 8.05 runtimes. With AVX they're more along the lines of 2hrs 15 for similar WUs.
If the new 8.06 knocks those run times to 3hrs or less, it will be a huge boost to the computing power of all the hosts running stock, which I expect is most of them.
Anonymous platform would probably be less than 5%?
Grant
Darwin NT
ID: 1871049 · Report as offensive
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 13164
Credit: 1,160,866,277
RAC: 1,873
United States
Message 1871050 - Posted: 4 Jun 2017, 4:05:27 UTC - in response to Message 1871049.  

Yes, of course I didn't let the 8.05's run to completion, but at their 50% mark it looked like the completion times were going to be in the 2.5- 3.0 hr mark. The AVX app gets the Arecibo standard AR 0.44 done in 2.15-2.30 hours.
Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)
ID: 1871050 · Report as offensive
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 13164
Credit: 1,160,866,277
RAC: 1,873
United States
Message 1871051 - Posted: 4 Jun 2017, 4:11:47 UTC - in response to Message 1871048.  

There is 1 8.05 that you did run Keith, took 23 min 20 sec, lol

But that was an overflow task, so not valid.
Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)
ID: 1871051 · Report as offensive
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 13164
Credit: 1,160,866,277
RAC: 1,873
United States
Message 1871053 - Posted: 4 Jun 2017, 4:57:21 UTC

$hit! I just realized I hadn't made this an apples-apples comparison. I run the AVX app with real core affinity. I just switched the 8.06 (alt) app over to same affinity as my normal AVX r3330 one. At least the next 8 8.06 (alt) tasks will get processed with the same resources as my normal AVX ones.
Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)
ID: 1871053 · Report as offensive
Profile Darrell
Volunteer tester
Avatar

Send message
Joined: 14 Mar 03
Posts: 267
Credit: 1,418,681
RAC: 0
United States
Message 1871054 - Posted: 4 Jun 2017, 5:16:20 UTC
Last modified: 4 Jun 2017, 5:22:50 UTC

New 8.06 app after 25 minutes:

https://1drv.ms/i/s!ArIvftV8roEagVxHXJ1olJuoFrBn

8.04 app took anywhere from 13.6K to 15.2K secs to complete.

note: this is with the process running at priority level 6 - below normal
ID: 1871054 · Report as offensive
Profile RueiKe Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 14 Feb 16
Posts: 492
Credit: 378,512,430
RAC: 785
Taiwan
Message 1871062 - Posted: 4 Jun 2017, 8:40:53 UTC

I just added beta project to my R7-1700 and FX-8370 systems in addition to my i7-69050X system. I will suspend SETI to get some beta results out quickly.
GitHub: Ricks-Lab
Instagram: ricks_labs
ID: 1871062 · Report as offensive
Profile Raistmer
Volunteer developer
Volunteer tester
Avatar

Send message
Joined: 16 Jun 01
Posts: 6325
Credit: 106,370,077
RAC: 121
Russia
Message 1871066 - Posted: 4 Jun 2017, 9:19:24 UTC

Windows/x86 running on an AMD x86_64 or Intel EM64T CPU 8.05 10 Feb 2017, 20:57:18 UTC 300 GigaFLOPS
Windows/x86 running on an AMD x86_64 or Intel EM64T CPU 8.06 (alt) 1 Jun 2017, 17:48:19 UTC 34 GigaFLOPS

Both numbers increased that means testers actively switch ON - thanks to all.

34/300=0.113 > 27/299=0.090 so good so far.
SETI apps news
We're not gonna fight them. We're gonna transcend them.
ID: 1871066 · Report as offensive
Grant (SSSF)
Volunteer tester

Send message
Joined: 19 Aug 99
Posts: 13736
Credit: 208,696,464
RAC: 304
Australia
Message 1871205 - Posted: 5 Jun 2017, 6:09:46 UTC - in response to Message 1871139.  
Last modified: 5 Jun 2017, 6:13:34 UTC

Looking good. And much faster so far on my CPU.

So far Ive seen roughly a good 20-30min or so speed up over the 8.04 & 8.05 applications, that's been with Arecibo work. I'm waiting to get some GBT work to see how it goes with them.

EDIT- 2 GBT WUs queued up to be run with 8.06 in the next few hours.

So far-
SETI@home v8 8.04 windows_intelx86     18.95 GFLOPS
SETI@home v8 8.05 windows_x86_64       14.03 GFLOPS
SETI@home v8 8.06 windows_x86_64 (alt) 16.82 GFLOPS

Grant
Darwin NT
ID: 1871205 · Report as offensive
Grant (SSSF)
Volunteer tester

Send message
Joined: 19 Aug 99
Posts: 13736
Credit: 208,696,464
RAC: 304
Australia
Message 1871212 - Posted: 5 Jun 2017, 7:50:22 UTC

It would appear Beta has run out of work.
Grant
Darwin NT
ID: 1871212 · Report as offensive
Profile Wiggo
Avatar

Send message
Joined: 24 Jan 00
Posts: 34744
Credit: 261,360,520
RAC: 489
Australia
Message 1871215 - Posted: 5 Jun 2017, 8:32:10 UTC

If those numbers Grant are for you E6600 then they're not bad at all, but if they're for your i7 2600 then it's much better to stick with the AVX app.

Cheers.
ID: 1871215 · Report as offensive
Grant (SSSF)
Volunteer tester

Send message
Joined: 19 Aug 99
Posts: 13736
Credit: 208,696,464
RAC: 304
Australia
Message 1871216 - Posted: 5 Jun 2017, 9:09:56 UTC - in response to Message 1871215.  
Last modified: 5 Jun 2017, 9:11:37 UTC

If those numbers Grant are for you E6600 then they're not bad at all, but if they're for your i7 2600 then it's much better to stick with the AVX app.

The i7 and definitely staying with AVX.
A lot of hardware can't run the AVX application, so for the 95% or so who run stock, it will give a huge boost to their output (and those of us using Lunatics on older hardware).
It's still looking like up to 30min off of Arecibo work, and possibly as much as 45-60min off of some GBT work (though only done a couple of GBT WUs so far).
Grant
Darwin NT
ID: 1871216 · Report as offensive
Profile RueiKe Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 14 Feb 16
Posts: 492
Credit: 378,512,430
RAC: 785
Taiwan
Message 1871226 - Posted: 5 Jun 2017, 10:42:23 UTC

My results so far:
i7-6950X: https://setiweb.ssl.berkeley.edu/beta/host_app_versions.php?hostid=81607
SETI@home v8 8.04 windows_intelx86 19.02 GFLOPS
SETI@home v8 8.05 windows_x86_64 17.65 GFLOPS
SETI@home v8 8.06 windows_x86_64 (alt): 17.00 GFLOPS
R7-1700: https://setiweb.ssl.berkeley.edu/beta/show_host_detail.php?hostid=82652
SETI@home v8 8.04 windows_intelx86: 11.91 GFLOPS
SETI@home v8 8.05 windows_x86_64 13.90 GFLOPS
SETI@home v8 8.06 windows_x86_64 (alt): 12.80 GFLOPS
FX-8370: https://setiweb.ssl.berkeley.edu/beta/show_host_detail.php?hostid=82653
SETI@home v8 8.04 windows_intelx86: no tasks
SETI@home v8 8.05 windows_x86_64 11.08 GFLOPS
SETI@home v8 8.06 windows_x86_64 (alt): 12.04 GFLOPS
GitHub: Ricks-Lab
Instagram: ricks_labs
ID: 1871226 · Report as offensive
Profile Tom M
Volunteer tester

Send message
Joined: 28 Nov 02
Posts: 5124
Credit: 276,046,078
RAC: 462
Message 1871248 - Posted: 5 Jun 2017, 14:18:09 UTC

I have just re-installed Seti Beta on my elderly Xeon that does not support AVX. The Lunatics version of the cpu task runs at about 3.5 hours. The reason I am using it rather than the stock task is the stock Seti task was running about 5.5 hours.

I wait with baited breath :)

Tom
A proud member of the OFA (Old Farts Association).
ID: 1871248 · Report as offensive
Profile Tom M
Volunteer tester

Send message
Joined: 28 Nov 02
Posts: 5124
Credit: 276,046,078
RAC: 462
Message 1871261 - Posted: 5 Jun 2017, 15:12:36 UTC - in response to Message 1871249.  

I have just re-installed Seti Beta on my elderly Xeon that does not support AVX. The Lunatics version of the cpu task runs at about 3.5 hours. The reason I am using it rather than the stock task is the stock Seti task was running about 5.5 hours.

I wait with baited breath :)

Tom


Hmmmm, Not getting anything from the Beta server...... slowly turning blue (in face)....
A proud member of the OFA (Old Farts Association).
ID: 1871261 · Report as offensive
EdwardPF
Volunteer tester

Send message
Joined: 26 Jul 99
Posts: 389
Credit: 236,772,605
RAC: 374
United States
Message 1871263 - Posted: 5 Jun 2017, 15:14:51 UTC - in response to Message 1871261.  

I think, maybe, they got a better response from us than they expected.

Ed F
ID: 1871263 · Report as offensive
Profile Tom M
Volunteer tester

Send message
Joined: 28 Nov 02
Posts: 5124
Credit: 276,046,078
RAC: 462
Message 1871330 - Posted: 5 Jun 2017, 21:23:08 UTC - in response to Message 1871019.  

Least detailed test result would be:

Windows/x86 running on an AMD x86_64 or Intel EM64T CPU 8.05 10 Feb 2017, 20:57:18 UTC 299 GigaFLOPS
Windows/x86 running on an AMD x86_64 or Intel EM64T CPU 8.06 (alt) 1 Jun 2017, 17:48:19 UTC 27 GigaFLOPS

If/when second number will exceed first - testing will be completed.


I'm missing something. I have never had a cpu with 3 digit GigaFlops like above. So what does it mean?

Tom
A proud member of the OFA (Old Farts Association).
ID: 1871330 · Report as offensive
Profile Brent Norman Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester

Send message
Joined: 1 Dec 99
Posts: 2786
Credit: 685,657,289
RAC: 835
Canada
Message 1871335 - Posted: 5 Jun 2017, 21:30:35 UTC - in response to Message 1871330.  
Last modified: 5 Jun 2017, 21:33:08 UTC

Large CPU flops means they have been rescheduling and the number is invalid for any comparison.
I can get my CPU apr over 1000GFlops :)

EDIT: Oh, I thought he meant from host page.
ID: 1871335 · Report as offensive
Profile Tom M
Volunteer tester

Send message
Joined: 28 Nov 02
Posts: 5124
Credit: 276,046,078
RAC: 462
Message 1871344 - Posted: 5 Jun 2017, 22:04:59 UTC - in response to Message 1871333.  
Last modified: 5 Jun 2017, 22:06:16 UTC

That's not a single CPU. It's total GigaFLOPS for those two latest apps, from the Beta Applications page:

https://setiweb.ssl.berkeley.edu/beta/apps.php

What we're hoping for, is that the latest app (now in testing) will pass the older app in total GigaFLOPS.


So that is the sum total of all testing rig Gflops for that app? Seems like if 3,000 testers suddenly piled on, the results would be completely misleading. Shouldn't it be an average/rig? Or is the point to say when the 2nd passes the first, it has been thoroughly tested?

Mumble, mumble, mumble...

Tom
A proud member of the OFA (Old Farts Association).
ID: 1871344 · Report as offensive
Previous · 1 · 2 · 3 · 4 · 5 . . . 7 · Next

Message boards : Number crunching : New binary to test on beta


 
©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.