BOINC 6.6.9 is available for Windows. As stated in other threads this is a Development version and may not work correctly.

Changes since 6.6.8...

- client: if "no more work" checked, don't request more work (duh!)

---

BOINC blog

Nice to see such complex and difficult things being sorted out.

Please insert sarcasm at will

---

Big change indeed ;^)
For now, I'll stick to BOINC 6.4.5., I've changed enough for now, when I'll try CUDA, I'll take a higher version, which is more suitable for CUDA.
And it works 'like a charm', gives me enough work and don't go in High Priority Mode, as well.
(IMHO, boinc.exe and boincmgr.exe, use up less CPU time then 5.10.45, does, but hé, can be a lot off reasons, why this happens in one of my hosts)

---

You probably missed the rest of the changes. There's plenty of them around.

The no more work option was really only broken in 6.6.8, hence why 6.6.9 was released quickly enough (20 hours apart). Given that the report had to come in about it (in the correct place!) and that the developers also sleep, eat and do other things with their lives (gasp!), it's worthy of applause, not sarcasm.

---

Ok, question. Was 6.4.5 under-fetching by so much that 6.6.9 had to get 50+ more(before I stopped it)? This on a C2D with an 8500GT. If yes then I will install it again but I was burned once by 6.6.4 and don't want it to happen again.

---

PROUD MEMBER OF Team Starfire World BOINC

The 6.6 versions have a whole new work fetch and client side CPU scheduler, rebuilt from the ground upwards. The way that debts are being used has also radically changed. The only thing that you can compare between the old way and the new way is that they do the same thing, just differently. You will have to watch out now how much extra work you fetch and essentially BOINC will have to learn all over again how long tasks take. And that won't be done within an hour or a day.

---

If I'm reading you right Ageless you are saying I have to decide how much is enough and then set NNT? I'm only running SETI so there is no problem with debt. I have been running with 6 days cache on 6.4.5 but when 6.6.9 started downloading all those it scared me.

Oh and one more question. I pulled out my cc_config file before I loaded 6.6.9. I noticed when it started running it was not running on my GPU, just two CPUs will it pick up on the GPU or should I leasve the cc_config in there?

---

PROUD MEMBER OF Team Starfire World BOINC

If I'm reading you right Ageless you are saying I have to decide how much is enough and then set NNT? I'm only running SETI so there is no problem with debt. I have been running with 6 days cache on 6.4.5 but when 6.6.9 started downloading all those it scared me.

What he's saying is that 6.6.9 does things very differently.

You should never have to set NNT. BOINC is supposed to handle this.

If you are runnning 6.6.9, you have an obligation to let it run and observe what it does -- you should not "help" it.

As was pointed out in the original post, this is a Development version. It is released by the developers so that it can be tried.

If you control it, then we're testing your ability to schedule work, not the BOINC scheduler.

If you're adverse to experimental software, you should not run experimental versions of BOINC.

---

Ned, I'm not adverse to running it I'm just trying to not wind up with 1000 WUs like happened with 6.6.4. I don't want to end up having tasks time out because it downloaded too many.

---

PROUD MEMBER OF Team Starfire World BOINC

What the 6.6 versions have extra is a GPU scheduler. If you are allowing the GPU to fetch work, you'll get a whole lot more than "what you're used to".

As for work not running on the GPU, that depends on what is downloaded.
Seti 6.03 is CPU only, Seti 6.08 is CUDA. We can't see what it says in your Tasks tab what kind of work you have.

The other thing to check though, is the extra preferences for GPU crunching that are included in the web-based preferences and the local advanced preferences (if you use those). I don't mean the If GPU is available etc. preference, but the "Suspend GPU work while computer is in use?" preference in your computing preferences. Yes, it says there it works only with 6.7+, but that option has (secretly) worked since 6.6.7 ;-)

---

Debt calculations are still FUBAR.

---

BOINC WIKI

Ok, I cut my cache back to 4 days, put the cc_config back in and reloadeed 6.6.9. Here are the starting messages...
2/18/2009 6:50:44 PM Starting BOINC client version 6.6.9 for windows_intelx86
2/18/2009 6:50:44 PM log flags: task, file_xfer, sched_ops
2/18/2009 6:50:44 PM Libraries: libcurl/7.19.2 OpenSSL/0.9.8i zlib/1.2.3
2/18/2009 6:50:44 PM Data directory: C:\ProgramData\BOINC
2/18/2009 6:50:44 PM Running under account perry
2/18/2009 6:50:44 PM SETI@home Found app_info.xml; using anonymous platform
2/18/2009 6:50:44 PM Processor: 2 GenuineIntel Intel(R) Celeron(R) CPU E1400 @ 2.00GHz [x86 Family 6 Model 15 Stepping 13]
2/18/2009 6:50:44 PM Processor features: fpu tsc pae nx sse sse2 pni mmx
2/18/2009 6:50:44 PM OS: Microsoft Windows Vista: Home Premium x86 Editon, Service Pack 1, (06.00.6001.00)
2/18/2009 6:50:44 PM Memory: 3.00 GB physical, 6.23 GB virtual
2/18/2009 6:50:44 PM Disk: 288.09 GB total, 221.76 GB free
2/18/2009 6:50:44 PM Local time is UTC -5 hours
2/18/2009 6:50:44 PM CUDA device: GeForce 8500 GT (driver version 18122, CUDA version 1.1, 512MB, est. 5GFLOPS)
2/18/2009 6:50:44 PM Not using a proxy
2/18/2009 6:50:44 PM Version change (6.4.5 -> 6.6.9)
2/18/2009 6:50:44 PM SETI@home URL: http://setiathome.berkeley.edu/; Computer ID: 4731257; location: home; project prefs: default
2/18/2009 6:50:44 PM General prefs: from SETI@home (last modified 27-Jan-2009 10:20:06)
2/18/2009 6:50:44 PM Computer location: home
2/18/2009 6:50:44 PM General prefs: no separate prefs for home; using your defaults
2/18/2009 6:50:44 PM Reading preferences override file
2/18/2009 6:50:44 PM Preferences limit memory usage when active to 1535.26MB
2/18/2009 6:50:44 PM Preferences limit memory usage when idle to 2763.46MB
2/18/2009 6:50:44 PM Preferences limit disk usage to 100.00GB
2/18/2009 6:50:44 PM Preferences limit # CPUs to 3
2/18/2009 6:50:44 PM Running CPU benchmarks
2/18/2009 6:50:44 PM Suspending computation - running CPU benchmarks
2/18/2009 6:51:15 PM Benchmark results:
2/18/2009 6:51:15 PM Number of CPUs: 3
2/18/2009 6:51:15 PM 1849 floating point MIPS (Whetstone) per CPU
2/18/2009 6:51:15 PM 4020 integer MIPS (Dhrystone) per CPU
2/18/2009 6:51:16 PM SETI@home Restarting task 17ja09aa.29585.14383.10.8.221_1 using setiathome_enhanced version 608
2/18/2009 6:51:16 PM SETI@home Restarting task 17ja09aa.29585.14383.10.8.213_1 using setiathome_enhanced version 608
2/18/2009 6:51:16 PM SETI@home Restarting task 23ja09ab.25968.1708.6.8.241_1 using setiathome_enhanced version 608
2/18/2009 6:51:19 PM SETI@home Sending scheduler request: To fetch work.
2/18/2009 6:51:19 PM SETI@home Reporting 1 completed tasks, requesting new tasks
2/18/2009 6:51:24 PM SETI@home Scheduler request completed: got 0 new tasks
2/18/2009 6:52:40 PM SETI@home Sending scheduler request: To fetch work.
2/18/2009 6:52:40 PM SETI@home Requesting new tasks
2/18/2009 6:52:45 PM SETI@home Scheduler request completed: got 0 new tasks
2/18/2009 6:55:00 PM SETI@home Sending scheduler request: To fetch work.
2/18/2009 6:55:00 PM SETI@home Requesting new tasks
2/18/2009 6:55:05 PM SETI@home Scheduler request completed: got 0 new tasks

---

PROUD MEMBER OF Team Starfire World BOINC

2/18/2009 6:50:44 PM CUDA device: GeForce 8500 GT (driver version 18122, CUDA version 1.1, 512MB, est. 5GFLOPS)

Nice to see this one working, finally we have easy information. ;-)

Anyway, I see that you are using Raistmer's optimized applications, both for CPU and GPU work. It may be that there's something in that that's now having problems. I'd say, wait until Gecko_R or Raistmer come by this thread, they can explain it better than I can. Even I have my limitations. :-)

---

Ok, it's settled down and seems to be running smooth now. Sorry I panicked but I just didn't want to end up with way more than I could handle.

---

PROUD MEMBER OF Team Starfire World BOINC

Ned, I'm not adverse to running it I'm just trying to not wind up with 1000 WUs like happened with 6.6.4. I don't want to end up having tasks time out because it downloaded too many.

Perryjay,

If you are going to run a test version, this is one of the dangers. You may wind up with 1000 work units -- and that needs to be reported to the BOINC developers.

Also, you need to let it run until it is obvious that there are consequences, and document those.

Otherwise, you really should stick to a released version.

-- Ned

---

Just when you thought it safe to try, they release a newer version :-)

6.6.10 is now out for Mac and Windows.

---

BOINC blog

Just when you thought it safe to try, they release a newer version :-)

6.6.10 is now out for Mac and Windows.

Running a development version is never "safe" -- it is interesting, and it a whole different way to contribute, but if you want safe, stick to a released version.

---

I don't understand why my GPU WU is "waiting to run". Is it competing for resources? It ran for roughly 5 minutes and is held up.

---

I don't understand why my GPU WU is "waiting to run". Is it competing for resources? It ran for roughly 5 minutes and is held up.

If you run development versions, you have to accept that s**t happens. This question is more appropriate to the Boinc dev boards, really.

I believe that it is the 6.6.9 Boinc Manager that causes this - I am now running 6.6.9 with the 6.4.5 BM and no longer get this problem.

F.

---

They skipped the 6.6.6 - LOL

---

--
Get up, stand up! Don\'t give up the fight!
Credits will make everybody feel high! ;-)