5.25 is released for Windows. Let me know if this fixes any of the illegal instruction problems.

Eric

---

I picked up a WU just a couple of minutes after you posted this:

01/08/2007 19:13:10|SETI@home Beta Test|Scheduler RPC succeeded [server version 509]

(TZ UTC+1), but it was still tagged for the 5.24 I got this morning. Does anyone know roughly what time interval we ought to expect between 'release' and 'allocation/download'?

That WU was the first I've got in a long while, and it's been trying constantly but running on fumes since then. Eric, I think you're going to have to re-start the splitter if you want a meaningful test.

Edit - got another one at 19:30:30, but it still hasn't downloaded the new app - still shows 5.24

I have two finished and ready to report.

---

David Stites
Mount Vernon, WA USA

Hmm, resultid 2179810 on my Pentium-MMX Win95 system 5.24 exited immediately (~1 second) with:

exit code -2122873184 (0x817786a0)

I hope the flag fix will apply there too, but suspect it won't. Google doesn't show any hits for that exit code and I haven't had problems with other UPX packed executables on that system.

                                                                Joe

Edit: It also crashed very quickly in standalone tests, without giving useful information until I unpacked it. Then it gave a startup error messagebox:

The SETIATHOME_5.24_WINDOWS_INTELX86.EXE file is
linked to missing export ADVAPI32.DLL:CryptAcquireContextA.

Same story with 5.25, resultid 2183870 gave:

exit code -2122745628 (0x817978e4)

And an unpacked version standalone test gets the same startup error. Not unexpected, I'm sure that's coming from some of the BOINC API stuff. I haven't tried building against the latest versions of that, I sure wish they'd make a stable version once in awhile rather than advising projects to build apps with unstable code.

                                                                 Joe

5.25 is released for Windows. Let me know if this fixes any of the illegal instruction problems.

Eric, with v5.25 the illegal instruction error is gone, but now i get reproducable error 0xc0000005 (access violation). The "Choosing Optimal functions" is the last that shows onscreen.

Results:
2184640
2184652
2184666

edit: When running in standalone mode i get the same error.

---

_\|/_
U r s

Edit - got another one at 19:30:30, but it still hasn't downloaded the new app - still shows 5.24

If you aren't well into crunching one of these (or have just returned one), you should be able to do an "update project" followed by a "reset project." That should send you the same results, but with 5.25 this time.

Eric

---

Edit - got another one at 19:30:30, but it still hasn't downloaded the new app - still shows 5.24

If you aren't well into crunching one of these (or have just returned one), you should be able to do an "update project" followed by a "reset project." That should send you the same results, but with 5.25 this time.

Eric

Neat trick, and it worked perfectly: I interrupted it between WUs, and the server sent me a nice shiny new 5.25 and resent all my previous work - didn't lose a WU.

Host 12316 on parade and reporting for duty.

This is strange: one of my PIII ate 5.25 W.U. in a mater of seconds with the same errors Urs mentioned above. But I have two PII that appear to be crunching 5.25 quite happily. hum!!

---

Research indicated that an advapi32.dll for Win95 with the CryptAcquireContextA export is provided with Internet Explorer 4.01 or later, so I suppose most of the few remaining Win95 systems have it. I've avoided that virus attracter, so had the original version.

After some research I replaced it with version 4.71.0118.0 and my 200MHz Pentium-MMX system is now running 5.25. In another 25 hours or so it should have a result to report.

                                                                 Joe

i've completed one 5.25 result on my Laptop with WinXP SP2
the other result of this workunit is from 5.21
Validate state is now Initial

result
workunit

no errors and grafic ist ok

Matthias

---

Matthias

My first 5.24 linux x86_64 WU.... first error :-(

http://setiathome.berkeley.edu/beta/result.php?resultid=2173426

Validate error between the results from linux 5.24 and win XP 5.21
workunit 590267

one detail i've seen on my win xp laptop with seti 5.25
beginning time "to completition" ist 2 to 3 times of total used "CPU time" after competition

on this workunit i got too an validate error
workunit 590259

---

Matthias

enhanced 5.25 on Windows.

I have an estimate of 7 hours, but the WUs are finishing in less than one hour.

It is a P4 with HT and 1 GB memory: http://setiweb.ssl.berkeley.edu/beta/show_host_detail.php?hostid=2834

---

Supporting BOINC, a great concept !

All my 5.24 / 5.25 results so far have errored after < 90 seconds.

http://setiweb.ssl.berkeley.edu/beta/result.php?resultid=2173968 5.24 - exit code -1073741795 (0xc000001d)
http://setiweb.ssl.berkeley.edu/beta/result.php?resultid=2189736 5.25 - exit code -1073741819 (0xc0000005)
http://setiweb.ssl.berkeley.edu/beta/result.php?resultid=2189770 5.25 - exit code -1073741819 (0xc0000005)
http://setiweb.ssl.berkeley.edu/beta/result.php?resultid=2189862 5.25 - exit code -1073741819 (0xc0000005)

Win XP SP2
AMD Athlon XP 2200+
256MB RAM

BOINC ver. 5.10.7

I was monitoring the slot folder during the last 2 of these. The WU seemed to error at about the point when the state.sah would usually be written for the first time.

---

enhanced 5.25 on windows:

This box crashed all WUs after less than 1 minute (it's an AMD): http://setiweb.ssl.berkeley.edu/beta/show_host_detail.php?hostid=16345

The WUs: http://setiweb.ssl.berkeley.edu/beta/results.php?hostid=16345

---

Supporting BOINC, a great concept !

My host 12316 (Xeon 5320, Vista) seems to be coping well with 5.25, but my wingmen are having all sorts of problems with the same WUs.

These are all Windows 5.25 results from WUs which crunched OK here.
2184712 - 0xc0000005, PIII, W2000 (hi Urs!)
2184726 - 0xc0000005, PIII, W2000 (hi Urs again)
2172650 - 0xc000001d, Intel x86 F6 M8 S10, W2000
2187966 (-9) - 0xc0000005, Intel x86 F6 M8 S6, XP
2187968 (-9) - 0xc0000005, Athlon XP 3000+, W2000
2192390 - 0xc0000005, Athlon XP 2400+, XP
2191044 - 0xc0000005, Athlon XP 2500+, W2000 (hi Simon)
2191046 - 0xc0000005, Athlon XP 3200+, XP
2191082 - 0xc0000005, Athlon XP 2500+, W2000 (hi Simon again)
2191092 - 0xc0000005, Athlon XP 2500+, W2000 (and again)
2191094 - 0xc0000005, Athlon XP 2400+, XP

Would it be fair to conclude that it's the older CPUs which are having the problems again?

Edit - in light of the above, I've attached host 20205 to Beta - 400Mhz Celeron Mendocino MMX, running W98SE. It's just done 1% of the first WU, but no problems so far.

Windows, enhanced 5.25

Has run for 2 minutes, since 45 minutes is sitting there and doing nothing.

Found in stderr.txt in slot-directory:

*********************************

setiathome_enhanced 5.25 DevC++/MinGW

Work Unit Info:
...............
WU true angle range is : 1.490551
Optimal function choices:
-----------------------------------------------------
name
-----------------------------------------------------
v_BaseLineSmooth (no other)
v_vGetPowerSpectrum 0.00400 0.00000
sse1_ChirpData_ak 0.07516 0.00000
v_vpfTranspose8x4ntw 0.03602 0.00000


Unhandled Exception Detected...

- Unhandled Exception Record -
Reason: Access Violation (0xc0000005) at address 0x004B7AAE read attempt to address 0x805428E0

Engaging BOINC Windows Runtime Debugger...

***********************************************

[edit:] this host: http://setiweb.ssl.berkeley.edu/beta/show_host_detail.php?hostid=16348

---

Supporting BOINC, a great concept !

host 15539 AMD laptop.
I am also getting computation errors on Enhanced 5.25/5.24 after 60 - 70 seconds. Other two host running ok.

---

A computer program will always do what you tell it to do, but rarely what you want it to do.

Same with a second WU on the same host:

setiathome_enhanced 5.25 DevC++/MinGW

Work Unit Info:
...............
WU true angle range is : 1.491445
Optimal function choices:
-----------------------------------------------------
name
-----------------------------------------------------
v_BaseLineSmooth (no other)
v_vGetPowerSpectrum 0.00398 0.00000
sse1_ChirpData_ak 0.07554 0.00000
v_vpfTranspose8x4ntw 0.03622 0.00000


Unhandled Exception Detected...

- Unhandled Exception Record -
Reason: Access Violation (0xc0000005) at address 0x004B7AAE read attempt to address 0x805428E0

Engaging BOINC Windows Runtime Debugger...

---

Supporting BOINC, a great concept !

Since my previous post, my host 14006 has trashed another 6 WUs; all at around the sime time 73 - 75 seconds.

I've set Beta to "No new tasks" until this gets resolved, as it seems pointless to create any more Client error, Compute error results. I will monitor the boards to await 5.26 or whatever.

Regards, Keith