S39 Observation Thread

Rojer
Rojer
Joined: 2 Apr 05
Posts: 23
Credit: 49,400,244
RAC: 0

RE: RE: I've replaced

Message 26249 in response to message 26247

Quote:
Quote:


I've replaced s38 with s39 in PII,and err rised.That's the reason?

2006-3-22 20:41:06|Einstein@Home|Aborting result r1_1346.0__1716_S4R2a_2: exceeded disk limit: 107477607.000000 > 100000000.000000

It appears that by bad luck your disk usage exceeded the limit you set in your preferences at this point. You may wish to go to the general preferences link on your account page and adjust the disk limits, assuming your system has room for more. You appear to be allowing 100 Megabytes.


That's my seting.
Leave at least
(Values smaller than 0.001 are ignored) 0.05 GB disk space free

My disk have 200M space to use.I find some strang thing,the 'stderr.txt' which in the directory \\BOINC\\slots\\0 have a quikly growth for about 1M/5second,till it used up all my space.I post some of info below.

----------------------------------------------
2006-03-22 23:23:58.4111 [normal]: Optimised by akosf (S-39) --> 'projects/einstein.phys.uwm.edu/albert_4.37_windows_intelx86.exe'
2006-03-22 23:23:58.4212 [normal]: Started search at lalDebugLevel = 0
2006-03-22 23:24:14.6345 [normal]: Checkpoint-file 'Fstat.out.ckp' not found.
2006-03-22 23:24:14.6345 [normal]: No usable checkpoint found, starting from beginning.

***UNHANDLED EXCEPTION****
Reason: Illegal Instruction (0xc000001d) at address 0x0040A84E

1: 03/22/06 23:24:14
1: e:\\einsteinathome\\cfs\\windows_build\\albert4.37\\cfslaldemod.c(836) +72 bytes (TestLALDemod)
1: e:\\einsteinathome\\cfs\\windows_build\\albert4.37\\computefstatistic.c(946) +28 bytes (boincmain)
1: e:\\einsteinathome\\cfs\\windows_build\\albert4.37\\computefstatistic.c(3521) +19 bytes (worker)
1: e:\\einsteinathome\\cfs\\windows_build\\albert4.37\\boinc\\api\\graphics_impl.c(74) +0 bytes (foobar)
1: e:\\einsteinathome\\cfs\\windows_build\\albert4.37\\computefstatistic.c(3521) +19 bytes (worker)
1: e:\\einsteinathome\\cfs\\windows_build\\albert4.37\\boinc\\api\\graphics_impl.c(74) +0 bytes (foobar)
1: e:\\einsteinathome\\cfs\\windows_build\\albert4.37\\computefstatistic.c(3521) +19 bytes (worker)
1: e:\\einsteinathome\\cfs\\windows_build\\albert4.37\\boinc\\api\\graphics_impl.c(74) +0 bytes (foobar)
.......

Wish you can understand my English:)

Crunchers For More Power
Crunchers For M...
Joined: 3 Aug 05
Posts: 69
Credit: 1,071,273
RAC: 0

S39 work's fine on my P3-S

S39 work's fine on my P3-S Dualsystem. Thanks to akosf.

S39 -> 2h fo "standard" WUs.

(C37 -> 3,5h and original albert -> 8h)

*wow*

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

Hi

Message 26251 in response to message 26249

Hi Rojer!

Quote:
***UNHANDLED EXCEPTION****
Reason: Illegal Instruction (0xc000001d) at address 0x0040A84E


This CPU doesn't support SSE.
Try to run C37.

josep
josep
Joined: 9 Mar 05
Posts: 63
Credit: 1,156,542
RAC: 0

RE: [pre] CPU S-39/S-38

Message 26252 in response to message 26241

Quote:

[pre]
CPU S-39/S-38 S-39/dist
P4 .70 .30
Banias .80 .22
PIII 1.05 .27
PII 1.07 .28
[/pre]

I have seen a similar improvement in a P4 Northwood 2.4Ghz (no HT capable)

[pre] (S39 time) / (S38 time) = 0.73[/pre]

that's 37% faster.

And

[pre] (S39 time) / (C37 time) = 0.55 (80% faster)[/pre]

That's really, really good. I remember, in the same machine, S37a (SSE2) being 9% slower than C37

For my Athlon XP 2600+

[pre] (S39 time) / (S38 time) = 0.88 (14% faster)[/pre]

these results are also good, but not as impressive as P4 Northwood ones:


Rudy
Rudy
Joined: 12 Dec 05
Posts: 33
Credit: 3,733,942
RAC: 1,954

Just noticed the S39, but

Just noticed the S39, but cannot download it.

Could someone who already has S39 be so kind and mirror it somewhere?

Honza
Honza
Joined: 10 Nov 04
Posts: 136
Credit: 3,332,354
RAC: 0

@ Rudy: just for a while, you

@ Rudy: just for a while, you can download from http://www.boinc.cz/temp/alb_S39.zip
Hope it helps...

Rudy
Rudy
Joined: 12 Dec 05
Posts: 33
Credit: 3,733,942
RAC: 1,954

Thanks, got it.

Thanks, got it.

_heinz
_heinz
Joined: 4 Jan 06
Posts: 79
Credit: 130,476
RAC: 0

With the original Software

With the original Software albert437 I need with a Pentium4 2,6GHZ for 1 WU (8h) more than 10 hours.
With the new S39 I need still 1hour 30 Minutes for 1 WU. That´s a funtastic speed up.
thanks to the developer akosf for his excellent work.
seti_britta

archae86
archae86
Joined: 6 Dec 05
Posts: 3,157
Credit: 7,200,004,931
RAC: 877,698

RE: [pre] CPU S-39/S-38

Message 26257 in response to message 26241

Quote:
[pre]
CPU S-39/S-38 S-39/dist
PIII 1.05 .27
PII 1.07 .28
[/pre]My initial results from my PII and PIII suggest that S-39 actually slowed them down by 5 to 7% compared to S-38. I would be eager to hear from others running these older CPUs, which were so hugely helped by C-37 and S-38.

There are at least two and possibly three errors in the table as I posted it.

1. The oldest slowest machine which is reported by the BOINC systems as a Pentium II is reported by wcpuid and cpu-z as a Pentium III EB of the Coppermine flavor. (this makes it the same major family as the machine I show as Pentium III, though a different stepping, clock rate, motherboard, and chip set)
2. I flat got it wrong--the slowest machine (which I showed as PII but is in fact a 733 MHz Coppermine PIII) has an improvement on S-39, getting .85 of S-38 execution time, and .22 of dist. Still appreciably less S-39 speedup than my P4, but entirely respectable, and the best overall improvement to dist I've seen.
3. So that leaves the question of what is going on on my 933 MHz Coppermine PIII (the one labelled as PIII in my original table). It has finished three results and continues to look as posted. I shall switch it back to S-38 now and report later. Perhaps by bad luck the computational requirements of results within major datafile 1415.5 had a secular shift near the point where I updated the science ap.

If you wonder how I could get so basic a thing as the processor family wrong, the problem is that I relied on how it posts through BOINC-- to wit as:
GenuineIntel Pentium(r) II Processor

More later.
But given that one of my Pentium III machines shows appreciable improvement, and one other user has reported slight improvement--I'd not suggest users hold back on giving S-39 a try on Pentium III machines.

Since Pentium II did not have SSE, I _would not_ use S-39 on one of those.

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

Hi! I'm working on S39L that

Hi! I'm working on S39L that uses less L1 data cache (~10kB).
Good news to Pentium-III owners.
But now, I'm sharing my measured results with you.

Pentium4 mobile 1,8 GHz (Northwood): 30200 -> 7600 -> 397%
Athlon XP 1,53 GHz (ThoroughbredB): 21200 -> 5550 -> 382%
PentiumM 1,86 GHz (Dothan): 18500 -> 4500 -> 411%
Duron 2 GHz (Applebread): 15200 -> 4200 -> 362%

S39L needed 4100 sec on one of my Durons. :-)

Comment viewing options

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