I hope your app is not more buggy as the stock.. ;-)

wuid=389527115

---

Raistmer wrote:

There is new version of package available here:
http://lunatics.kwsn.net/gpu-crunching/modified-seti-mb-cuda-opt-ap-package-for-full-gpu-utilize.msg12576.html#msg12576

It incorporates triplet overflow bug-fix provided by CUDA MB developer.
Should give invalid overflow on lesser amount of tasks now.
Other bugs still remain, debugging in progress, please, be patient.

Sutaru Tsureku wrote:

I hope your app is not more buggy as the stock.. ;-)
wuid=389527115

yours:
-----
SETI@Home Informational message -9 result_overflow
Spike count: 7
Pulse count: 24
Triplet count: 0
Gaussian count: 0
called boinc_finish

:P

---

Raistmer could you make the new version downloadable from a different site? I have been unable to register/never received a conformation email from lunatics.

---

I hope your app is not more buggy as the stock.. ;-)

wuid=389527115

1) You don't use current build. It's rev396, not 380.
2) Current build less buggy then current stock (6.06)
3) It also faster even in free single core run.

Raistmer could you make the new version downloadable from a different site? I have been unable to register/never received a conformation email from lunatics.

Ok, download it from here:

http://files.mail.ru/UG8OQM

Though I don't really follow my RAC that closely I thought this was a telling sign of how well CUDA can do.

The RAC on my AMD64 3800+ 4537486 RAC 836.86

Just passed my E2180 Dual Core 4489245 RAC 719.05

I know it's not a real benchmark but I thought it was interesting that my AMD single core pc would be able at some point to get a higher RAC than my Dual Core. Both run 24/7. The dual core runs the SSE3 opti MB app and the SSE3 AP app. The AMD runs the Modified CUDA app and SSE3 opti AP app. The GPU and modified CUDA app really did make the AMD a true dual core pc. Thanks Raistmer

---

I don't get no new WUs for SETI. I currently have 5 AP WUs and 1 CPDN as backup (paused). I'm using the 4c package with ap_5.00r69_SSE3.exe.

I just allowed new work for GPUgrid and instantly got a WU from them (http://www.gpugrid.net/results.php?hostid=21636).

I was wondering why I can't get any new SETI WUs for CUDA. When I used AK8 it worked like a charm but since you, raimster, put so much work to get CUDA to work halfway properly for us, I don't want to go back when I can use my GPU that is much faster. Any idea why I receive no new work?

By the way, thanks for already helping me/us so much with your work!

---

I hope your app is not more buggy as the stock.. ;-)

wuid=389527115

1) You don't use current build. It's rev396, not 380.
2) Current build less buggy then current stock (6.06)
3) It also faster even in free single core run.

No.. it's not my rig.. my rig is the one with the CPU-result..

Hey.. no 'insult'.. just little kidding..! ;-D

Your work and the work from the optimizer-crew is very much appreciated from the whole SETI@home-community!


BTW.
I'm little curious and very restless.. I want to buy my GTX 2xx GPU.. ;-D
But I'll buy first, if the SETI@home-CUDA-app is BUG-free!
I have no interest to 'babysitting' my PC.. :-(

• When we all could get a BUG-free CUDA-app?
  Maybe in three months?
  
  
• When it'll be possible to let run MB and AP on CPU (optimized) and MB on GPU simultanously?
  Your CUDA-app.. or the stock CUDA-app in the future could be optimized?
  
  
• We'll see in the future maybe a CUDA-AP-app also?
  

---

I don't get no new WUs for SETI. I currently have 5 AP WUs and 1 CPDN as backup (paused). I'm using the 4c package with ap_5.00r69_SSE3.exe.

I just allowed new work for GPUgrid and instantly got a WU from them (http://www.gpugrid.net/results.php?hostid=21636).

I was wondering why I can't get any new SETI WUs for CUDA. When I used AK8 it worked like a charm but since you, raimster, put so much work to get CUDA to work halfway properly for us, I don't want to go back when I can use my GPU that is much faster. Any idea why I receive no new work?

By the way, thanks for already helping me/us so much with your work!

BOINC log ?

BTW.
I'm little curious and very restless.. I want to buy my GTX 2xx GPU.. ;-D
But I'll buy first, if the SETI@home-CUDA-app is BUG-free!
I have no interest to 'babysitting' my PC.. :-(

• When we all could get a BUG-free CUDA-app?
  
  

thats what i wanna know.

i long for the day in a few years when i can buy myself a box for about $200 with 3 or 4 of thees, now topend cards, but by then a rarity, and have myself a 10kw/h folding rig. :)

---

I hope your app is not more buggy as the stock.. ;-)

wuid=389527115

1) You don't use current build. It's rev396, not 380.
2) Current build less buggy then current stock (6.06)
3) It also faster even in free single core run.

No.. it's not my rig.. my rig is the one with the CPU-result..

Hey.. no 'insult'.. just little kidding..! ;-D

Your work and the work from the optimizer-crew is very much appreciated from the whole SETI@home-community!


BTW.
I'm little curious and very restless.. I want to buy my GTX 2xx GPU.. ;-D
But I'll buy first, if the SETI@home-CUDA-app is BUG-free!
I have no interest to 'babysitting' my PC.. :-(

• When we all could get a BUG-free CUDA-app?
  Maybe in three months?
  
  
• When it'll be possible to let run MB and AP on CPU (optimized) and MB on GPU simultanously?
  Your CUDA-app.. or the stock CUDA-app in the future could be optimized?
  
  
• We'll see in the future maybe a CUDA-AP-app also?
  

Well, the more peoples will work on debugging/improving, the faster we all get what we want.
Have any spare time? - join to us ;)

About GPU AP - there is some AP CUDA version but it need more debugging too. It's now in hard position. Devaster, I and other Lunatics worked on it some time but then this CUDA MB quest begins abruptly - now I don't know should I continue debugging of own AP CUDA or just relax and wait 1-2 months to get AP CUDA from Berkeley/nVidia - more cooperation in such questions is needed indeed to work more effective. Will see after this CUDA MB will be more bug free - maybe situation already became much more positive in this aspect, I very hope on this.
At least in CUDA MB aspect it's definitely better now than was on beginning this "quest".

ADDON: And about simultaneous run of CPU MB and CUDA MB - on current stage this question should be adressed to BOINC devs mostly.
W/o their help we will be able to do that only when both apps will be united in single executive IMHO (it's possible, but not very soon).

Raistmer, in the latest build, have you tweeked the AP client as well?

Because not only have CUDA times dropped a little, AP tasks are about 30 minutes quicker.

---

Raistmer, in the latest build, have you tweeked the AP client as well?

Because not only have CUDA times dropped a little, AP tasks are about 30 minutes quicker.

No, no changes in AP in this release.
But don't forget if some apps work simultaneously on CPU they share many of CPU resources (L2 cache common to both cores for example). So, changes in one app really can affect on other processes.

Another possibility - if you have multicore system - it seems AP build very sensible to its location in memory during execution. For example, one of simultaneously running opt AP processes can experience amount of page switches in orders of magnitude bigger than another one. So, some deviation in run time is possible even w/o any changes in CUDA app (other loaded processes and DLLs can affect opt AP too, not only CUDA MB).

...
So, some deviation in run time is possible even w/o any changes in CUDA app (other loaded processes and DLLs can affect opt AP too, not only CUDA MB).

I noticed this when v5 opti came out. My first 4 v5 WUs had a large deviation in CPU time. Two were around 185,000 seconds, and the other two were around 250,000. All four were crunching at the same time.

---

Linux laptop:
record uptime: 1511d 20h 19m (ended due to the power brick giving-up)

I don't get no new WUs for SETI. I currently have 5 AP WUs and 1 CPDN as backup (paused). I'm using the 4c package with ap_5.00r69_SSE3.exe.

I just allowed new work for GPUgrid and instantly got a WU from them (http://www.gpugrid.net/results.php?hostid=21636).

I was wondering why I can't get any new SETI WUs for CUDA. When I used AK8 it worked like a charm but since you, raimster, put so much work to get CUDA to work halfway properly for us, I don't want to go back when I can use my GPU that is much faster. Any idea why I receive no new work?

By the way, thanks for already helping me/us so much with your work!

BOINC log ?

There you go:
11.01.2009 16:40:02||Starting BOINC client version 6.4.5 for windows_intelx86
11.01.2009 16:40:02||log flags: task, file_xfer, sched_ops
11.01.2009 16:40:02||Libraries: libcurl/7.19.0 OpenSSL/0.9.8i zlib/1.2.3
11.01.2009 16:40:02||Data directory: C:\BOINCdata
11.01.2009 16:40:02||Running under account Home
11.01.2009 16:40:02|Einstein@Home|Found app_info.xml; using anonymous platform
11.01.2009 16:40:02|SETI@home|Found app_info.xml; using anonymous platform
11.01.2009 16:40:02||Processor: 2 AuthenticAMD AMD Athlon(tm) 64 X2 Dual Core Processor 4800+ [x86 Family 15 Model 35 Stepping 2]
11.01.2009 16:40:02||Processor features: fpu tsc pae nx sse sse2 3dnow mmx
11.01.2009 16:40:02||OS: Microsoft Windows XP: Professional x86 Editon, Service Pack 3, (05.01.2600.00)
11.01.2009 16:40:02||Memory: 2.00 GB physical, 3.85 GB virtual
11.01.2009 16:40:02||Disk: 136.38 GB total, 64.99 GB free
11.01.2009 16:40:02||Local time is UTC +1 hours
11.01.2009 16:40:02||Not using a proxy
11.01.2009 16:40:02||CUDA devices found
11.01.2009 16:40:02||Coprocessor: GeForce 9800 GT (1)
11.01.2009 16:40:02|rosetta@home|URL: http://boinc.bakerlab.org/rosetta/; Computer ID: 292643; location: (none); project prefs: default
11.01.2009 16:40:02|HashClash|URL: http://boinc.banaan.org/hashclash/; Computer ID: 1665; location: (none); project prefs: default
11.01.2009 16:40:02|boincsimap|URL: http://boinc.bio.wzw.tum.de/boincsimap/; Computer ID: 68912; location: school; project prefs: default
11.01.2009 16:40:02|climateprediction.net|URL: http://climateprediction.net/; Computer ID: 414899; location: home; project prefs: default
11.01.2009 16:40:02|Einstein@Home|URL: http://einstein.phys.uwm.edu/; Computer ID: 581852; location: home; project prefs: default
11.01.2009 16:40:02|Predictor @ Home|URL: http://predictor.scripps.edu/; Computer ID: 212464; location: (none); project prefs: default
11.01.2009 16:40:02|SETI@home|URL: http://setiathome.berkeley.edu/; Computer ID: 1955611; location: work; project prefs: default
11.01.2009 16:40:02|GPUGRID|URL: http://www.gpugrid.net/; Computer ID: 21636; location: (none); project prefs: default
11.01.2009 16:40:02||General prefs: from boincsimap (last modified 28-Dec-2007 19:22:25)
11.01.2009 16:40:02||Computer location: school
11.01.2009 16:40:02||General prefs: using separate prefs for school
11.01.2009 16:40:02||Reading preferences override file
11.01.2009 16:40:02||Preferences limit memory usage when active to 2046.42MB
11.01.2009 16:40:02||Preferences limit memory usage when idle to 2046.42MB
11.01.2009 16:40:02||Preferences limit disk usage to 65.42GB
11.01.2009 16:40:02|SETI@home|Restarting task ap_20no08ad_B0_P0_00336_20090103_22066.wu_0 using astropulse version 500
11.01.2009 16:40:02|GPUGRID|Restarting task LHw8190-GPUTEST6-1-5-acemd_0 using acemd version 655

Ignore the 'school' and 'work' computer location. Don't know why I have those names.. University would be better, I guess. :)

Edit: I have the latest nVidia driver (181.20), by the way.

---

Bei den ganzen Projekten die du da am laufen hast, hat der Scheduler überhaupt noch Luft, Arbeit von Seti anzufordern? ;)

---

...
So, some deviation in run time is possible even w/o any changes in CUDA app (other loaded processes and DLLs can affect opt AP too, not only CUDA MB).

I noticed this when v5 opti came out. My first 4 v5 WUs had a large deviation in CPU time. Two were around 185,000 seconds, and the other two were around 250,000. All four were crunching at the same time.

There is preliminary exit for AP too (as result overflow -9 for SETI MB), maybe one of tasks had that overflow...

I don't get no new WUs for SETI. I currently have 5 AP WUs and 1 CPDN as backup (paused). I'm using the 4c package with ap_5.00r69_SSE3.exe.

I just allowed new work for GPUgrid and instantly got a WU from them (http://www.gpugrid.net/results.php?hostid=21636).

I was wondering why I can't get any new SETI WUs for CUDA. When I used AK8 it worked like a charm but since you, raimster, put so much work to get CUDA to work halfway properly for us, I don't want to go back when I can use my GPU that is much faster. Any idea why I receive no new work?

By the way, thanks for already helping me/us so much with your work!

BOINC log ?

There you go:
11.01.2009 16:40:02||Starting BOINC client version 6.4.5 for windows_intelx86
11.01.2009 16:40:02||log flags: task, file_xfer, sched_ops
11.01.2009 16:40:02||Libraries: libcurl/7.19.0 OpenSSL/0.9.8i zlib/1.2.3
11.01.2009 16:40:02||Data directory: C:\BOINCdata
11.01.2009 16:40:02||Running under account Home
11.01.2009 16:40:02|Einstein@Home|Found app_info.xml; using anonymous platform
11.01.2009 16:40:02|SETI@home|Found app_info.xml; using anonymous platform
11.01.2009 16:40:02||Processor: 2 AuthenticAMD AMD Athlon(tm) 64 X2 Dual Core Processor 4800+ [x86 Family 15 Model 35 Stepping 2]
11.01.2009 16:40:02||Processor features: fpu tsc pae nx sse sse2 3dnow mmx
11.01.2009 16:40:02||OS: Microsoft Windows XP: Professional x86 Editon, Service Pack 3, (05.01.2600.00)
11.01.2009 16:40:02||Memory: 2.00 GB physical, 3.85 GB virtual
11.01.2009 16:40:02||Disk: 136.38 GB total, 64.99 GB free
11.01.2009 16:40:02||Local time is UTC +1 hours
11.01.2009 16:40:02||Not using a proxy
11.01.2009 16:40:02||CUDA devices found
11.01.2009 16:40:02||Coprocessor: GeForce 9800 GT (1)
11.01.2009 16:40:02|rosetta@home|URL: http://boinc.bakerlab.org/rosetta/; Computer ID: 292643; location: (none); project prefs: default
11.01.2009 16:40:02|HashClash|URL: http://boinc.banaan.org/hashclash/; Computer ID: 1665; location: (none); project prefs: default
11.01.2009 16:40:02|boincsimap|URL: http://boinc.bio.wzw.tum.de/boincsimap/; Computer ID: 68912; location: school; project prefs: default
11.01.2009 16:40:02|climateprediction.net|URL: http://climateprediction.net/; Computer ID: 414899; location: home; project prefs: default
11.01.2009 16:40:02|Einstein@Home|URL: http://einstein.phys.uwm.edu/; Computer ID: 581852; location: home; project prefs: default
11.01.2009 16:40:02|Predictor @ Home|URL: http://predictor.scripps.edu/; Computer ID: 212464; location: (none); project prefs: default
11.01.2009 16:40:02|SETI@home|URL: http://setiathome.berkeley.edu/; Computer ID: 1955611; location: work; project prefs: default
11.01.2009 16:40:02|GPUGRID|URL: http://www.gpugrid.net/; Computer ID: 21636; location: (none); project prefs: default
11.01.2009 16:40:02||General prefs: from boincsimap (last modified 28-Dec-2007 19:22:25)
11.01.2009 16:40:02||Computer location: school
11.01.2009 16:40:02||General prefs: using separate prefs for school
11.01.2009 16:40:02||Reading preferences override file
11.01.2009 16:40:02||Preferences limit memory usage when active to 2046.42MB
11.01.2009 16:40:02||Preferences limit memory usage when idle to 2046.42MB
11.01.2009 16:40:02||Preferences limit disk usage to 65.42GB
11.01.2009 16:40:02|SETI@home|Restarting task ap_20no08ad_B0_P0_00336_20090103_22066.wu_0 using astropulse version 500
11.01.2009 16:40:02|GPUGRID|Restarting task LHw8190-GPUTEST6-1-5-acemd_0 using acemd version 655

Ignore the 'school' and 'work' computer location. Don't know why I have those names.. University would be better, I guess. :)

Edit: I have the latest nVidia driver (181.20), by the way.

I don't see any request for work at all here.
Try to manual update project.

I don't see any request for work at all here.
Try to manual update project.

If I could manually update this project..........hoo boy.......you would see some big changes here.....LOL......

---

"Freedom is just Chaos, with better lighting." Alan Dean Foster

I don't get no new WUs for SETI. I currently have 5 AP WUs and 1 CPDN as backup (paused). I'm using the 4c package with ap_5.00r69_SSE3.exe.

I just allowed new work for GPUgrid and instantly got a WU from them (http://www.gpugrid.net/results.php?hostid=21636).

I was wondering why I can't get any new SETI WUs for CUDA. When I used AK8 it worked like a charm but since you, raimster, put so much work to get CUDA to work halfway properly for us, I don't want to go back when I can use my GPU that is much faster. Any idea why I receive no new work?

By the way, thanks for already helping me/us so much with your work!

BOINC log ?

There you go:
11.01.2009 16:40:02||Starting BOINC client version 6.4.5 for windows_intelx86
11.01.2009 16:40:02||log flags: task, file_xfer, sched_ops
11.01.2009 16:40:02||Libraries: libcurl/7.19.0 OpenSSL/0.9.8i zlib/1.2.3
11.01.2009 16:40:02||Data directory: C:\BOINCdata
11.01.2009 16:40:02||Running under account Home
11.01.2009 16:40:02|Einstein@Home|Found app_info.xml; using anonymous platform
11.01.2009 16:40:02|SETI@home|Found app_info.xml; using anonymous platform
11.01.2009 16:40:02||Processor: 2 AuthenticAMD AMD Athlon(tm) 64 X2 Dual Core Processor 4800+ [x86 Family 15 Model 35 Stepping 2]
11.01.2009 16:40:02||Processor features: fpu tsc pae nx sse sse2 3dnow mmx
11.01.2009 16:40:02||OS: Microsoft Windows XP: Professional x86 Editon, Service Pack 3, (05.01.2600.00)
11.01.2009 16:40:02||Memory: 2.00 GB physical, 3.85 GB virtual
11.01.2009 16:40:02||Disk: 136.38 GB total, 64.99 GB free
11.01.2009 16:40:02||Local time is UTC +1 hours
11.01.2009 16:40:02||Not using a proxy
11.01.2009 16:40:02||CUDA devices found
11.01.2009 16:40:02||Coprocessor: GeForce 9800 GT (1)
11.01.2009 16:40:02|rosetta@home|URL: http://boinc.bakerlab.org/rosetta/; Computer ID: 292643; location: (none); project prefs: default
11.01.2009 16:40:02|HashClash|URL: http://boinc.banaan.org/hashclash/; Computer ID: 1665; location: (none); project prefs: default
11.01.2009 16:40:02|boincsimap|URL: http://boinc.bio.wzw.tum.de/boincsimap/; Computer ID: 68912; location: school; project prefs: default
11.01.2009 16:40:02|climateprediction.net|URL: http://climateprediction.net/; Computer ID: 414899; location: home; project prefs: default
11.01.2009 16:40:02|Einstein@Home|URL: http://einstein.phys.uwm.edu/; Computer ID: 581852; location: home; project prefs: default
11.01.2009 16:40:02|Predictor @ Home|URL: http://predictor.scripps.edu/; Computer ID: 212464; location: (none); project prefs: default
11.01.2009 16:40:02|SETI@home|URL: http://setiathome.berkeley.edu/; Computer ID: 1955611; location: work; project prefs: default
11.01.2009 16:40:02|GPUGRID|URL: http://www.gpugrid.net/; Computer ID: 21636; location: (none); project prefs: default
11.01.2009 16:40:02||General prefs: from boincsimap (last modified 28-Dec-2007 19:22:25)
11.01.2009 16:40:02||Computer location: school
11.01.2009 16:40:02||General prefs: using separate prefs for school
11.01.2009 16:40:02||Reading preferences override file
11.01.2009 16:40:02||Preferences limit memory usage when active to 2046.42MB
11.01.2009 16:40:02||Preferences limit memory usage when idle to 2046.42MB
11.01.2009 16:40:02||Preferences limit disk usage to 65.42GB
11.01.2009 16:40:02|SETI@home|Restarting task ap_20no08ad_B0_P0_00336_20090103_22066.wu_0 using astropulse version 500
11.01.2009 16:40:02|GPUGRID|Restarting task LHw8190-GPUTEST6-1-5-acemd_0 using acemd version 655

Ignore the 'school' and 'work' computer location. Don't know why I have those names.. University would be better, I guess. :)

Edit: I have the latest nVidia driver (181.20), by the way.

I don't see any request for work at all here.
Try to manual update project.

*sigh*
Believe me, I tried that a dozen times. Here is the latest:
11.01.2009 18:36:03|SETI@home|Sending scheduler request: Requested by user. Requesting 0 seconds of work, reporting 0 completed tasks
11.01.2009 18:36:08|SETI@home|Scheduler request completed: got 0 new tasks


I don't get why Seti doesn't even request work when I update it manually. Other projects do that automatically when WUs are below the buffer. I set the buffer to 10 days to see if there is a change, but there isn't. :(

---