I've changed the preference on my SETI page and on the BOINC manager, so I'll see how the PC copes. It's relatively old, and can be RAM-bound as I run 4Gb on Vista 32-bit - I understand I'd need to change to 64-bit to use any more RAM. (Reminds me of having to use QEMM to access above 640K on DOS!)

To turn off the GPU, I take it I just do it from the activity menu in BOINC rather than via the SETI online preferences?

Yes, from the (local) activity menu - that takes effect immediately, whereas the website change might not be noticed for several hours (at the next update/work fetch/work report).

It's running in the background as I write. I chose Suspend GPU from the Activity menu at 12:57 after the "resuming computation" entry, and nothing appeared in the log. What should be written when I turn off GPU? This is the log:

10/01/2013 12:57:01 |  | Resuming computation
10/01/2013 12:57:01 | SETI@home | [cpu_sched] Resuming 24oc12ad.10465.2930.12.10.127_1
10/01/2013 12:57:01 | SETI@home | [cpu_sched] Resuming 25oc12aa.15614.4157.14.10.30_0
10/01/2013 12:57:07 | SETI@home | [checkpoint] result 25oc12aa.15614.4157.14.10.30_0 checkpointed
10/01/2013 12:57:31 | SETI@home | [checkpoint] result 24oc12ad.10465.2930.12.10.127_1 checkpointed
10/01/2013 12:58:20 | SETI@home | [checkpoint] result 25oc12aa.15614.4157.14.10.30_0 checkpointed
10/01/2013 12:58:40 | SETI@home | [checkpoint] result 24oc12ad.10465.2930.12.10.127_1 checkpointed
10/01/2013 12:59:33 | SETI@home | [checkpoint] result 25oc12aa.15614.4157.14.10.30_0 checkpointed
10/01/2013 12:59:45 | SETI@home | [checkpoint] result 24oc12ad.10465.2930.12.10.127_1 checkpointed
10/01/2013 13:00:44 | SETI@home | [checkpoint] result 25oc12aa.15614.4157.14.10.30_0 checkpointed
10/01/2013 13:00:51 | SETI@home | [checkpoint] result 24oc12ad.10465.2930.12.10.127_1 checkpointed
10/01/2013 13:01:53 | SETI@home | [checkpoint] result 25oc12aa.15614.4157.14.10.30_0 checkpointed
10/01/2013 13:01:56 | SETI@home | [checkpoint] result 24oc12ad.10465.2930.12.10.127_1 checkpointed
10/01/2013 13:03:03 | SETI@home | [checkpoint] result 24oc12ad.10465.2930.12.10.127_1 checkpointed
10/01/2013 13:03:04 | SETI@home | [checkpoint] result 25oc12aa.15614.4157.14.10.30_0 checkpointed

---

YJ
You may have speed, but I have momentum

Something like mine does, when it comes to time-of-day switch off:
I run 3 Seti tasks on 3 CPU cores and one Einstein task on the GPU.

11/01/2013 07:00:00 |  | Suspending computation - time of day
11/01/2013 07:00:00 | SETI@home | [cpu_sched] Preempting 24oc12ae.11734.17245.11.10.223_0 (left in memory)
11/01/2013 07:00:00 | SETI@home | [cpu_sched] Preempting 24oc12ae.11734.20108.11.10.115_1 (left in memory)
11/01/2013 07:00:00 | SETI@home | [cpu_sched] Preempting 25oc12aa.8907.33112.11.10.104_1 (left in memory)
11/01/2013 07:00:00 | Einstein@Home | [cpu_sched] Preempting p2030.20120222.G194.47+00.27.S.b3s0g0.00000_208_0 (removed from memory)

Why do the Seti tasks stay in memory? Because I've got 'Leave in memory while suspended' set to On.

---

It would only log the GPU switch-off if there was an actual GPU task running at the time.

It's running in the background as I write. I chose Suspend GPU from the Activity menu at 12:57 after the "resuming computation" entry, and nothing appeared in the log. What should be written when I turn off GPU?

You have to have Running GPU task (in your case AstroPulse v6 v6.04 (opencl_nvidia_100))
to see any effect from "Suspend GPU from the Activity menu"

Do you use any program to monitor GPU load and temperature? (and know that 'something' is using it):

http://rh-software.com/
http://www.techpowerup.com/gpuz/
http://www.ozone3d.net/gpu_caps_viewer/

http://event.msi.com/vga/afterburner/download.htm
http://www.almico.com/speedfan.php

TThrottle (can also control the temperature to avoid overheating):
http://www.efmer.eu/boinc/

P.S.
You lose big processing time when you "Kill the app/process":
http://setiathome.berkeley.edu/results.php?hostid=4499642&offset=0&show_names=0&state=6&appid=

Is the computer too unresponsive when AstroPulse runs on your GPU?

---

- ALF - "Find out what you don't do well ..... then don't do it!" :)

I've been running with GPU off, and suspending from the Activity menu is immediate. I'll try the same with GPU on next. Afraid I'm away from this afternoon, so I won't be able to try anything more till after the weekend.

@BilBg Thanks for the info - I'd never considered the impact running task would have on the GPU, so I haven't been running anything to check on my it. I will have a look at that as soon as I get back after the weekend. With GPU off, 2 SETI tasks run and there is a slight but noticeable effect in general usage. With AP running in GPU, there is a significant impact in that everything is much, much slower and the PC is almost unusable - that's how I know it's been left running...

Is it either 1 task in GPU or 2 on my twin cores, or could I have 3 running when idle (1xGPU + 2xCPU)?

---

YJ
You may have speed, but I have momentum

Got back after leaving PC running for weekend. The popup text on the BOINC icon in the system tray says

Computing is enabled
GPU computing is suspended - computer in use
Network is enabled

On checking the task manager, astropulse is running. Changing the GPU option in the Activity menu to Suspend GPU has no effect, When I close the BOINC manager, the task does stop, so it looks like the AP module isn't behaving as it should within BOINC.

The good news is that I had to check, as it seems to be running manageably and not hogging the CPU. It may have done this before, and I just hadn't noticed, as I only checked when the PC was sluggish...

---

YJ
You may have speed, but I have momentum

Any thoughts on why AP is still running when GPU is suspended?

---

YJ
You may have speed, but I have momentum

I recently noticed the following logging options for cc_config.xml:

<app_msg_receive> (Shared-memory messages received from applications.)
<app_msg_send> (Shared-memory messages sent to applications.)

If you turn them on you might be able to see whether the client is or is not sending a suspend order and whether the application acknowledges receiving it or not, etc. I don't know the details of the protocol between client and apps but someone around here probably will.

Any thoughts on why AP is still running when GPU is suspended?

AP != GPU
(not every AP runs on your GPU)

You recently killed (probably) an CPU AP task (or was this just 'natural' error?):
http://setiathome.berkeley.edu/workunit.php?wuid=1150387841

---

- ALF - "Find out what you don't do well ..... then don't do it!" :)

Try to use -exit_check additional option via cmd line or ap_cmdline.txt file.

---