You CAN crunch MilkyWay ATI GPUS with BOINC 5.10.45.

That could be confusing to some (since no GPU wus are availiable). Here's my version of that (in no offense):

One can crunch MW wus with an ATI GPU using BOINC 5.10.45.

That works for me banditwolf. Probably works for 6.6.36 as well - 'One can crunch MW wus with an ATI GPU using BOINC 6.6.36'. So long as it's clear that 5.10.45 will work for [some] gpus [ATI graphics processing units, a selection of].

I can confirm BOINC V5.10.45 works with my ATI GPUs, and the RAC is slowly recovering from it's recent quite steep fall. Whether I can get the proportionate increase Ice manages will be seen over time.

Interestingly, I find Einstein now only crunches on 1 CPU of my quad, as, I think, BOINC sees 3 CPUs (the GPU) and only runs one Einstein WU at a time. I can confirm the GPU is loaded to 95% with 3 Milkyway WUs using GPU-Z 0.3.4.

---

Go away, I was asleep

Hi
Same thing happened to me, more than 3 WUs started running with 6.6.36 (with a slowdown in progres, and I had n2 defined in app_info) so I went back to 5.4.5 and 5.10.45 to try.

You should be able to overcome the 3 standard WU's on a quad problem by setting the ncpus parameter in the cc_config.xml file.

Setting ncpus = 5 for a quad will fool Boinc into thinking you have 5 cores so it should load 4 cores plus the GPU.

Here's an example:

<options>
<ncpus> 5 </ncpus>
</options>

Disclaimer: I'm not sure how that will interact with the app_info file that drives the ATI GPU.

---

Symington weather report and video feed

I'l leave things running as they currently are, but will update the changes tomorrow and see how things go.

---

Go away, I was asleep

Disclaimer: I'm not sure how that will interact with the app_info file that drives the ATI GPU.

I'm running with ncpus = 5, avg_ncpus for MW GPU as 0.5, and a bit of custom resource shares (50 for MW and 100 for WCG). BOINC (well, Synecdoche now) always seemed to run 4 CPUs and 2 GPUs as long as one or other other didn't get hung and get debts up in the air.

EDIT: Oh, BOINC v6.2.19 and Synecdoche v0.1.1, Windows 7 RC1 64-bit.

---

Let's play nice, friends...this thread is degenerating quickly.

Thx

::Bows to Blurf Almighty::

---

me@rescam.org

http://boinc.berkeley.edu/trac/wiki/VersionHistory

history :)

If you don't have a CUDA card and are running Vista/Win7 you *can* run 5.4.5 -- you just need to manually restart BOINC whenever you restart the the workstation.

I don't know if this matters in the above quote but I'm running Vista, CPU only and I turned UAC off so that if the system reboots I don't have to manually start Boinc up afterwards.

[edit]Running Boinc 5.10.45 if I remember the numbers right, yup just checked...

---

If you don't have a CUDA card and are running Vista/Win7 you *can* run 5.4.5 -- you just need to manually restart BOINC whenever you restart the the workstation.

I don't know if this matters in the above quote but I'm running Vista, CPU only and I turned UAC off so that if the system reboots I don't have to manually start Boinc up afterwards.

[edit]Running Boinc 5.10.45 if I remember the numbers right, yup just checked...

Does enabling UAC cause the 5.x.x versions to misbehave / not function?

It does not allow the manager to open as it is not a registered app.

---

Does enabling UAC cause the 5.x.x versions to misbehave / not function?

The only problem I had with UAC turned on was that when I rebooted the system I had to manually start the manager. Even though it was allowed to run at start up in the settings UAC kept asking permission for it to run. FWIW, I'm running Vista Ultimate 32 bit and there is a setting in the local policies that I changed to disable UAC. It is not available in the home versions. I think it can be disabled though the user account section, but is not as effective as being able to modify the local policy settings.

---

And yes, tons of running WUs despite your settings for avg and max ncpus, all with their timers counting even when not getting resources is how 6.6.36, 37, and 38 all behave. It breaks it even worse than it already is, because not only is the scheduler still trying to schedule it against CPU time, but now it thinks every WU is going to take hours instead of minutes to run. Seems like nothing but negative progress since 6.4.7 which is the last version I've found to behave at all predictably.

Not saying it's better, just saying it's broken in a way that's predictable and can be worked around. Unfortunately it's very touchy about how projects schedule WUs, so I have to tweak it for every project I'm mix and match with MW. Was running it with Enigma last month for instance, and they like sending batches of very short deadline units mixed with normal deadlines, causing the manager to panic "high priority" stuff breaking all scheduling and resource sharing. Sooo, had to set very short queues, and script a debt reset to positive debt values for MW which was also running 205 resource share vs 50 for Enigma, and forced 30 minute project updates. Took awhile to figure that mess out, but once I did 80k/day smooth as could be. Now I'm on Yoyo and we're back to it being a mess again. Ah well :-)

Official CAL support isn't due until 6.10 or 6.12 "near the end of the year" at the earliest.

---