Hi, I'm currently using 3 Titan Black to do crunching.
In order to maximize the performance, I added the app_config.xml file to let GPUGrid run 2 tasks on each GPU simultaneously.
At first it seems to be working well, but after finishing the first 6 tasks, Bonic does not send me any new tasks to me.
I have no idea what's going on. Before I added the app_config.xml, everything goes well, so I'm wondering whether I have made any mistakes? Could someone help me on this? My event log and app_config.xlm contents are attached below.
Thank you very much!

[Event Log]
2016-04-12 9:23:59 PM | GPUGRID | Sending scheduler request: To fetch work.
2016-04-12 9:23:59 PM | GPUGRID | Requesting new tasks for CPU and NVIDIA GPU
2016-04-12 9:24:00 PM | GPUGRID | Scheduler request completed: got 0 new tasks
2016-04-12 9:24:00 PM | GPUGRID | No tasks sent
2016-04-12 9:24:00 PM | GPUGRID | Project has no tasks available
2016-04-12 9:24:00 PM | GPUGRID | [work_fetch] backing off CPU 727 sec
2016-04-12 9:24:00 PM | GPUGRID | [work_fetch] backing off NVIDIA GPU 590 sec
2016-04-12 9:24:00 PM | | [work_fetch] Request work fetch: RPC complete
2016-04-12 9:24:05 PM | | [work_fetch] ------- start work fetch state -------
2016-04-12 9:24:05 PM | | [work_fetch] target work buffer: 86400.00 + 86400.00 sec
2016-04-12 9:24:05 PM | | [work_fetch] --- project states ---
2016-04-12 9:24:05 PM | GPUGRID | [work_fetch] REC 1461664.272 prio -345.383 can't request work: scheduler RPC backoff (25.98 sec)
2016-04-12 9:24:05 PM | | [work_fetch] --- state for CPU ---
2016-04-12 9:24:05 PM | | [work_fetch] shortfall 2289631.14 nidle 3.00 saturated 0.00 busy 0.00
2016-04-12 9:24:05 PM | GPUGRID | [work_fetch] share 0.000 project is backed off (resource backoff: 722.42, inc 600.00)
2016-04-12 9:24:05 PM | | [work_fetch] --- state for NVIDIA GPU ---
2016-04-12 9:24:05 PM | | [work_fetch] shortfall 518400.00 nidle 3.00 saturated 0.00 busy 0.00
2016-04-12 9:24:05 PM | GPUGRID | [work_fetch] share 0.000 project is backed off (resource backoff: 584.83, inc 600.00)
2016-04-12 9:24:05 PM | | [work_fetch] ------- end work fetch state -------
2016-04-12 9:24:05 PM | | [work_fetch] No project chosen for work fetch
2016-04-12 9:24:31 PM | | [work_fetch] Request work fetch: Backoff ended for GPUGRID

[app_config]
<app_config>
<app>
<name>acemdlong</name>
<max_concurrent>9999</max_concurrent>
<gpu_versions>
<gpu_usage>0.5</gpu_usage>
<cpu_usage>1.00</cpu_usage>
</gpu_versions>
</app>
<app>
<name>acemdlong</name>
<max_concurrent>9999</max_concurrent>
<gpu_versions>
<gpu_usage>0.5</gpu_usage>
<cpu_usage>1.00</cpu_usage>
</gpu_versions>
</app>
<app>
<name>acemdlong</name>
<max_concurrent>9999</max_concurrent>
<gpu_versions>
<gpu_usage>0.5</gpu_usage>
<cpu_usage>1.00</cpu_usage>
</gpu_versions>
</app>
<app>
<name>acemdlong</name>
<max_concurrent>9999</max_concurrent>
<gpu_versions>
<gpu_usage>0.5</gpu_usage>
<cpu_usage>1.00</cpu_usage>
</gpu_versions>
</app>
<app>
<name>acemdlong</name>
<max_concurrent>9999</max_concurrent>
<gpu_versions>
<gpu_usage>0.5</gpu_usage>
<cpu_usage>1.00</cpu_usage>
</gpu_versions>
</app>
<app>
<name>acemdlong</name>
<max_concurrent>9999</max_concurrent>
<gpu_versions>
<gpu_usage>0.5</gpu_usage>
<cpu_usage>1.00</cpu_usage>
</gpu_versions>
</app>
</app_config>

[Event Log]
2016-04-12 9:23:59 PM | GPUGRID | Sending scheduler request: To fetch work.
2016-04-12 9:23:59 PM | GPUGRID | Requesting new tasks for CPU and NVIDIA GPU
2016-04-12 9:24:00 PM | GPUGRID | Scheduler request completed: got 0 new tasks
2016-04-12 9:24:00 PM | GPUGRID | No tasks sent
2016-04-12 9:24:00 PM | GPUGRID | Project has no tasks available
2016-04-12 9:24:00 PM | GPUGRID | [work_fetch] backing off CPU 727 sec
2016-04-12 9:24:00 PM | GPUGRID | [work_fetch] backing off NVIDIA GPU 590 sec
2016-04-12 9:24:00 PM | | [work_fetch] Request work fetch: RPC complete
2016-04-12 9:24:05 PM | | [work_fetch] ------- start work fetch state -------
2016-04-12 9:24:05 PM | | [work_fetch] target work buffer: 86400.00 + 86400.00 sec
2016-04-12 9:24:05 PM | | [work_fetch] --- project states ---
2016-04-12 9:24:05 PM | GPUGRID | [work_fetch] REC 1461664.272 prio -345.383 can't request work: scheduler RPC backoff (25.98 sec)
2016-04-12 9:24:05 PM | | [work_fetch] --- state for CPU ---
2016-04-12 9:24:05 PM | | [work_fetch] shortfall 2289631.14 nidle 3.00 saturated 0.00 busy 0.00
2016-04-12 9:24:05 PM | GPUGRID | [work_fetch] share 0.000 project is backed off (resource backoff: 722.42, inc 600.00)
2016-04-12 9:24:05 PM | | [work_fetch] --- state for NVIDIA GPU ---
2016-04-12 9:24:05 PM | | [work_fetch] shortfall 518400.00 nidle 3.00 saturated 0.00 busy 0.00
2016-04-12 9:24:05 PM | GPUGRID | [work_fetch] share 0.000 project is backed off (resource backoff: 584.83, inc 600.00)
2016-04-12 9:24:05 PM | | [work_fetch] ------- end work fetch state -------
2016-04-12 9:24:05 PM | | [work_fetch] No project chosen for work fetch
2016-04-12 9:24:31 PM | | [work_fetch] Request work fetch: Backoff ended for GPUGRID

The error is on the project's end, as there's not enough workunits queued at the moment to fulfill every request.
(see server message marked in red).
It seems that GPUGrid's resource share is set to 0 in your preferences (see the messages marked with dark red), so the client will request work only when it has run out of work of every projects with above zero priority. Setting zero resource share is a common practice for backup projects. It's nice to have a backup project for the times when the main project has run out of work.

There's no need to have the acemdlong application definition six times in the app_config.xml file:

<app_config>
 <app>
  <name>acemdlong</name>
  <max_concurrent>9999</max_concurrent>
  <gpu_versions>
   <gpu_usage>0.5</gpu_usage>
   <cpu_usage>1.00</cpu_usage>
  </gpu_versions>
 </app>
</app_config>

Thanks for point out my issue!
However, 7-8 hours has passed and I'm still not getting new tasks unfortunately. :( Is it normal?
I see users with high recent average are running tasks all the time, is there any tricks to keep GPU Grid always running?
For resource share, I believe that I've set it to 100, the same as the only other project I'm running -- World Community Grid(however, there're only CPU tasks for World Community Grid, so I believe the GPU Grid is my only project for my GPUs). No idea why it says 0, do you know how I can fix it?

The event log for today after I restarted my computer is as follow:
2016-04-13 8:55:10 AM | | Starting BOINC client version 7.6.22 for windows_x86_64
2016-04-13 8:55:10 AM | | log flags: file_xfer, sched_ops, task, work_fetch_debug
2016-04-13 8:55:10 AM | | Libraries: libcurl/7.45.0 OpenSSL/1.0.2d zlib/1.2.8
2016-04-13 8:55:10 AM | | Data directory: C:\ProgramData\BOINC
2016-04-13 8:55:10 AM | | Running under account Larfy-MainDeskTop
2016-04-13 8:55:12 AM | | CUDA: NVIDIA GPU 0: GeForce GTX TITAN Black (driver version 364.72, CUDA version 8.0, compute capability 3.5, 4096MB, 4009MB available, 5645 GFLOPS peak)
2016-04-13 8:55:12 AM | | CUDA: NVIDIA GPU 1: GeForce GTX TITAN Black (driver version 364.72, CUDA version 8.0, compute capability 3.5, 4096MB, 4009MB available, 5645 GFLOPS peak)
2016-04-13 8:55:12 AM | | CUDA: NVIDIA GPU 2: GeForce GTX TITAN Black (driver version 364.72, CUDA version 8.0, compute capability 3.5, 4096MB, 4009MB available, 5645 GFLOPS peak)
2016-04-13 8:55:12 AM | | OpenCL: NVIDIA GPU 0: GeForce GTX TITAN Black (driver version 364.72, device version OpenCL 1.2 CUDA, 6144MB, 4009MB available, 5645 GFLOPS peak)
2016-04-13 8:55:12 AM | | OpenCL: NVIDIA GPU 1: GeForce GTX TITAN Black (driver version 364.72, device version OpenCL 1.2 CUDA, 6144MB, 4009MB available, 5645 GFLOPS peak)
2016-04-13 8:55:12 AM | | OpenCL: NVIDIA GPU 2: GeForce GTX TITAN Black (driver version 364.72, device version OpenCL 1.2 CUDA, 6144MB, 4009MB available, 5645 GFLOPS peak)
2016-04-13 8:55:12 AM | | OpenCL CPU: Intel(R) Core(TM) i7-5960X CPU @ 3.00GHz (OpenCL driver vendor: Advanced Micro Devices, Inc., driver version 1800.7 (sse2,avx), device version OpenCL 1.2 AMD-APP (1800.7))
2016-04-13 8:55:12 AM | | Host name: DESKTOP-SUULGDL
2016-04-13 8:55:12 AM | | Processor: 16 GenuineIntel Intel(R) Core(TM) i7-5960X CPU @ 3.00GHz [Family 6 Model 63 Stepping 2]
2016-04-13 8:55:12 AM | | Processor features: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss htt tm pni ssse3 fma cx16 sse4_1 sse4_2 movebe popcnt aes f16c rdrandsyscall nx lm avx avx2 vmx tm2 dca pbe fsgsbase bmi1 smep bmi2
2016-04-13 8:55:12 AM | | OS: Microsoft Windows 10: Enterprise x64 Edition, (10.00.10586.00)
2016-04-13 8:55:12 AM | | Memory: 15.83 GB physical, 18.21 GB virtual
2016-04-13 8:55:12 AM | | Disk: 87.71 GB total, 56.64 GB free
2016-04-13 8:55:12 AM | | Local time is UTC -4 hours
2016-04-13 8:55:12 AM | | VirtualBox version: 5.0.12
2016-04-13 8:55:12 AM | GPUGRID | Found app_config.xml
2016-04-13 8:55:12 AM | | Config: report completed tasks immediately
2016-04-13 8:55:12 AM | GPUGRID | URL http://www.gpugrid.net/; Computer ID 316193; resource share 100
2016-04-13 8:55:17 AM | GPUGRID | General prefs: from GPUGRID (last modified 12-Apr-2016 22:08:40)
2016-04-13 8:55:17 AM | GPUGRID | Host location: none
2016-04-13 8:55:17 AM | GPUGRID | General prefs: using your defaults
2016-04-13 8:55:17 AM | | Preferences:
2016-04-13 8:55:17 AM | | max memory usage when active: 15400.92MB
2016-04-13 8:55:17 AM | | max memory usage when idle: 16049.38MB
2016-04-13 8:55:17 AM | | max disk usage: 56.76GB
2016-04-13 8:55:17 AM | | (to change preferences, visit a project web site or select Preferences in the Manager)
2016-04-13 8:55:17 AM | | [work_fetch] Request work fetch: Prefs update
2016-04-13 8:55:17 AM | | [work_fetch] Request work fetch: Startup
2016-04-13 8:55:18 AM | | [work_fetch] ------- start work fetch state -------
2016-04-13 8:55:18 AM | | [work_fetch] target work buffer: 180.00 + 4320.00 sec
2016-04-13 8:55:18 AM | | [work_fetch] --- project states ---
2016-04-13 8:55:18 AM | GPUGRID | [work_fetch] REC 1413826.093 prio -1.994 can request work
2016-04-13 8:55:18 AM | | [work_fetch] --- state for CPU ---
2016-04-13 8:55:18 AM | | [work_fetch] shortfall 63000.00 nidle 14.00 saturated 0.00 busy 0.00
2016-04-13 8:55:18 AM | GPUGRID | [work_fetch] share 0.500
2016-04-13 8:55:18 AM | | [work_fetch] --- state for NVIDIA GPU ---
2016-04-13 8:55:18 AM | | [work_fetch] shortfall 13500.00 nidle 3.00 saturated 0.00 busy 0.00
2016-04-13 8:55:18 AM | GPUGRID | [work_fetch] share 1.000
2016-04-13 8:55:18 AM | | [work_fetch] ------- end work fetch state -------
2016-04-13 8:55:19 AM | | [work_fetch] Request work fetch: RPC complete
2016-04-13 8:55:24 AM | | [work_fetch] ------- start work fetch state -------
2016-04-13 8:55:24 AM | | [work_fetch] target work buffer: 180.00 + 4320.00 sec
2016-04-13 8:55:24 AM | | [work_fetch] --- project states ---
2016-04-13 8:55:24 AM | GPUGRID | [work_fetch] REC 1413826.093 prio -1.000 can request work
2016-04-13 8:55:24 AM | | [work_fetch] --- state for CPU ---
2016-04-13 8:55:24 AM | | [work_fetch] shortfall 63000.00 nidle 14.00 saturated 0.00 busy 0.00
2016-04-13 8:55:24 AM | GPUGRID | [work_fetch] share 1.000
2016-04-13 8:55:24 AM | | [work_fetch] --- state for NVIDIA GPU ---
2016-04-13 8:55:24 AM | | [work_fetch] shortfall 13500.00 nidle 3.00 saturated 0.00 busy 0.00
2016-04-13 8:55:24 AM | GPUGRID | [work_fetch] share 1.000
2016-04-13 8:55:24 AM | | [work_fetch] ------- end work fetch state -------
2016-04-13 8:55:24 AM | GPUGRID | [work_fetch] set_request() for CPU: ninst 16 nused_total 0.00 nidle_now 14.00 fetch share 1.00 req_inst 16.00 req_secs 63000.00
2016-04-13 8:55:24 AM | GPUGRID | [work_fetch] set_request() for NVIDIA GPU: ninst 3 nused_total 0.00 nidle_now 3.00 fetch share 1.00 req_inst 3.00 req_secs 13500.00
2016-04-13 8:55:24 AM | GPUGRID | [work_fetch] request: CPU (63000.00 sec, 16.00 inst) NVIDIA GPU (13500.00 sec, 3.00 inst)
2016-04-13 8:55:24 AM | GPUGRID | Sending scheduler request: To fetch work.
2016-04-13 8:55:24 AM | GPUGRID | Requesting new tasks for CPU and NVIDIA GPU
2016-04-13 8:55:26 AM | GPUGRID | Scheduler request completed: got 0 new tasks
2016-04-13 8:55:26 AM | GPUGRID | No tasks sent
2016-04-13 8:55:26 AM | GPUGRID | Project has no tasks available
2016-04-13 8:55:26 AM | GPUGRID | [work_fetch] backing off CPU 664 sec
2016-04-13 8:55:26 AM | GPUGRID | [work_fetch] backing off NVIDIA GPU 506 sec
2016-04-13 8:55:26 AM | | [work_fetch] Request work fetch: RPC complete
2016-04-13 8:55:31 AM | | [work_fetch] ------- start work fetch state -------
2016-04-13 8:55:31 AM | | [work_fetch] target work buffer: 180.00 + 4320.00 sec
2016-04-13 8:55:31 AM | | [work_fetch] --- project states ---
2016-04-13 8:55:31 AM | GPUGRID | [work_fetch] REC 1413826.093 prio 0.000 can't request work: scheduler RPC backoff (25.98 sec)
2016-04-13 8:55:31 AM | | [work_fetch] --- state for CPU ---
2016-04-13 8:55:31 AM | | [work_fetch] shortfall 63000.00 nidle 14.00 saturated 0.00 busy 0.00
2016-04-13 8:55:31 AM | GPUGRID | [work_fetch] share 0.000 project is backed off (resource backoff: 659.14, inc 600.00)
2016-04-13 8:55:31 AM | | [work_fetch] --- state for NVIDIA GPU ---
2016-04-13 8:55:31 AM | | [work_fetch] shortfall 13500.00 nidle 3.00 saturated 0.00 busy 0.00
2016-04-13 8:55:31 AM | GPUGRID | [work_fetch] share 0.000 project is backed off (resource backoff: 500.65, inc 600.00)
2016-04-13 8:55:31 AM | | [work_fetch] ------- end work fetch state -------
2016-04-13 8:55:31 AM | | [work_fetch] No project chosen for work fetch
2016-04-13 8:55:57 AM | | [work_fetch] Request work fetch: Backoff ended for GPUGRID
2016-04-13 8:56:01 AM | | [work_fetch] ------- start work fetch state -------
2016-04-13 8:56:01 AM | | [work_fetch] target work buffer: 180.00 + 4320.00 sec
2016-04-13 8:56:01 AM | | [work_fetch] --- project states ---
2016-04-13 8:56:01 AM | GPUGRID | [work_fetch] REC 1413826.093 prio -1.000 can request work
2016-04-13 8:56:01 AM | | [work_fetch] --- state for CPU ---
2016-04-13 8:56:01 AM | | [work_fetch] shortfall 63000.00 nidle 14.00 saturated 0.00 busy 0.00
2016-04-13 8:56:01 AM | GPUGRID | [work_fetch] share 0.000 project is backed off (resource backoff: 629.13, inc 600.00)
2016-04-13 8:56:01 AM | | [work_fetch] --- state for NVIDIA GPU ---
2016-04-13 8:56:01 AM | | [work_fetch] shortfall 13500.00 nidle 3.00 saturated 0.00 busy 0.00
2016-04-13 8:56:01 AM | GPUGRID | [work_fetch] share 0.000 project is backed off (resource backoff: 470.65, inc 600.00)
2016-04-13 8:56:01 AM | | [work_fetch] ------- end work fetch state -------
2016-04-13 8:56:01 AM | | [work_fetch] No project chosen for work fetch

2016-04-13 8:55:12 AM | GPUGRID | URL http://www.gpugrid.net/; Computer ID 316193; resource share 100

Your GPUGrid resource share is ok (100). The previous share value was 0 because the GPUGrid project was in backoff state.
Now there is work in the queue, so your host received 6 workunits, which is the maximum (2 per GPU) no matter if you've set a longer queue.

Yay I'm getting new tasks now as you said. Thank you so much for your help! :D

Just to make sure that nothing is wrong with my system: I have not got new tasks since yesterday (April 13) late afternoon.
Is same thing true with other crunchers?

No WUs available: see 'Server status'

Just to make sure that nothing is wrong with my system: I have not got new tasks since yesterday (April 13) late afternoon.
Is same thing true with other crunchers?

No WUs available: see 'Server status'

this is nothing new though. Most of the time, you can see "0" unter "unsent", although tasks are being downloaded on a regular basis.

It appears the queue of new tasks is empty. I show a task completed 22:04:36 UTC.

Also not getting any tasks. Only getting resends now.

You can check the number of workunits by clicking on the "Server status" at the top right corner. The project's queues are low on workunits at the moment (e.g. there are more crunchers / hosts than workunits). There are 800-some workunits in progress, some of them will result in a new workunit, but when this number is below ~2000 (~the number of active hosts doubled) then there's a very low possibility to get new task. You should be patient while the queues are filled up with new work, or set up a backup project (by setting 0 resource share in your profile at the given project).

Server says:

Tasks ready to send 1
Tasks in progress 1,714

Boinc Manager says:

26/04/2016 08:46:51 | GPUGRID | Scheduler request completed: got 0 new tasks
26/04/2016 08:46:51 | GPUGRID | No tasks sent
26/04/2016 08:46:51 | GPUGRID | No tasks are available for Long runs (8-12 hours on fastest card)
26/04/2016 08:46:51 | GPUGRID | Project has no tasks available

Have not received a GPUGrid task in 3 days...

---

FAQ's

HOW TO:
- Opt out of Beta Tests
- Ask for Help

I still get tasks some times. Some of the tasks generate new ones right?

I've been running dry for the last couple of days as well

Got 1 on a 2 GPU system.

The Tasks in progress has dropped down into the 1500's:

Tasks ready to send 1
Tasks in progress 1,571

---

FAQ's

HOW TO:
- Opt out of Beta Tests
- Ask for Help

I've been running dry for the last couple of days as well

what we need to be aware of is that while in the past year there were quite a number of researchers active at GPUGRID, since around beginning of 2016 there is only one left: Gerard.
And of course, just one person is bound to be unable to fill the WU queue 24 hours a day.
As we all make the experience, many times it is impossible for several hours, or even up to a day or longer (like recently) to get new tasks - which is a pitty, but that's the way it is for the time being.
From what I have read somewhere in the BOINC forum the other day, this situation might remain unchanged at least until end of summer, if not beyond :-(