I've been running AP opencl_ati_100 both here and from the beta site since early May (credit new) without any major malfunctions. There have been the occasional hiccups due to ill advised tweaking (I'm on the wrong side 50 so brain farts are becoming more frequent). Since 26 November my GPU has only been getting opencl_ati_cat132 for S@H and opencl_ati5_cat132/AstroPulse v6 v6.06 (cal_ati) for S@H Beta.


No changes that I am aware of have been made to my setup. However...


I upgraded to BOINC client version 7.2.33 for windows_x86_64 on or about that time and...

The beta site has been tinkering with new ARM/Android Apps...


I realize these are coincidences but I noticed Starman and Variable have voiced the same complaints in regards to AP opencl_ati_100 downloads.


There have been, here in Hooterville, there have been numerous power outages for past two weeks. Thinking files may have been damaged as a result of improper shut downs I reverted back and forth between different BOINC stock app versions as well as AMD driver versions (13.1, 13.4 and 13.9) to no avail. My next thought is to wipe the drive and reinstall Windows and BOINC and...yada...yada...yada.


I would prefer to not have to do that.


Any suggestions?


Thanks,


Keifer



By the way I reside in the US, Zip code 01007. For those of you too young to remember that was a Hooterville reference to Petticoat Junction and Green Acres, not the restaurant.

---

AFAIK opencl_ati_100 and opencl_ati_cat132 is the same application.

Tasks marked 'opencl_ati_100' are sent to Catalyst <13.1
Tasks marked 'opencl_ati_cat132' are sent to Catalyst >13.1
No tasks are sent to Catalyst 13.1 (broken OpenCL compiler)

---

 


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

ati_opencl_100 are AP v6 for both S@H and S@H Beta stock app.

opencl_ati_cat132 are MB v7 for both S@H and S@H Beta stock app.

Someone please correct me if I am wrong.

Keifer

---

Any suggestions?

Thanks,
Keifer

Seems to be a problem with certain machines. I'm not sure why...

Try this;
To see which Preference Set your Host is using, go to the Host Details page and check the Location setting at the bottom. Change the Location Setting to match the desired Preference Set and hit Update.
Go to your preferences, http://setiathome.berkeley.edu/prefs.php?subset=project
Choose the correct set for your host and change it to;
Resource share: 100
Use CPU: no
Use ATI GPU: yes
Use NVIDIA: yes

SETI@home Enhanced: no
SETI@home v7: no
AstroPulse v6: yes
If no work for selected applications is available, accept work from other applications? no

Then go to the BOINC Manager, select the Projects tab, and hit the Update button a couple of times until the Event Log shows it's asking for AstroPulse.
Next, go to the BOINC Manager Menu TOOLS/Computing preferences/network usage
Set the Minimum work buffer to 10.00 days, Set the Max additional work buffer to 0.10 days

Since you don't have any active tasks, Hit the Project Reset Button. This will resend all the Project Files and hopefully wake up the server. The button is in the Projects tab of the BOINC Manage advanced view, named 'Reset project'.

If it works and you receive GPU AP tasks, then change it to;
Use CPU: yes
Use ATI GPU: yes
Use NVIDIA: yes

SETI@home Enhanced: no
SETI@home v7: no
AstroPulse v6: yes
If no work for selected applications is available, accept work from other applications? no

When there are not any APs available, change the bottom line to;
If no work for selected applications is available, accept work from other applications? Yes

See if that works.

I have been using those suggested settings since I began running AP's back in May.

I have tried "resets". When that didn't work I did a complete uninstall of all BOINC files and directories.

Still no AP ati_opencl_100 are sent.

Nothing has changed.

---

Did you Reset the Project?
After you Reset the project, if it still won't send GPU APs, Restart BOINC and post the first 30 lines of the Event Log. Make sure to include the first Work Request the Host sends.

Here's a couple other Hosts I'm watching;
This Host can't receive more than about a dozen GPU APs even though his quota is over a Hundred, http://setiathome.berkeley.edu/results.php?hostid=7137839&offset=0&show_names=0&state=0&appid=0

This Host can't receive over 8 GPU APs even though the quota is over 40+, http://setiathome.berkeley.edu/results.php?hostid=7102466&offset=0&show_names=0&state=0&appid=0

There are others....

Yes I have. Numerous times beginning in late November.

I'm leaning towards newer files being corrupted with older files as a result of the abrupt power failures.
That can happen right? One thing I have just thought of and should have tried days ago is to use the original install disk for my Sapphire HD 7770 combined with the version of BOINC mngr. that I was running back when I built this system (April 2012). Then keep upgrading each version of both AMD and BOINC until if pukes again.

What do you think?

Keifer

---

I'll forward my startup log after it runs through a few request cycles.

In the meantime...

Going back to my AP's that were abandoned by both S@H and S@H Beta as a result of my resets, I noticed that none of the resends were sent to systems running and requesting AP opencl_ati_100. Resends were only sent to optimized apps.
Doesn't that seem strange.

---

I'd like to see the Host startup log, including the first work request...
nods head.

ati_opencl_100 are AP v6 for both S@H and S@H Beta stock app.
opencl_ati_cat132 are MB v7 for both S@H and S@H Beta stock app.

Yes
Without looking at 'Applications' page I did it wrong

I wanted to say:
(opencl_ati_sah) == (opencl_ati_cat132) {same app, label depends on Catalyst version}
(ati_opencl_100) == (opencl_ati_100) (same app, label depends on BOINC version)

http://setiathome.berkeley.edu/apps.php

(I use app_info.xml so for me as for many others here how the tasks are labeled doesn't matter.
So better check which .exe files are in fact in SETI@home directory (<BOINC_Data>\projects\setiathome.berkeley.edu\)
and which Processes run in Windows Task Manager (or Process Explorer, ...)
)

---

 


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

12 Dec 13 13:20:55 | | cc_config.xml not found - using defaults
12 Dec 13 13:20:55 | | Starting BOINC client version 7.2.33 for windows_x86_64
12 Dec 13 13:20:55 | | log flags: file_xfer, sched_ops, task
12 Dec 13 13:20:55 | | Libraries: libcurl/7.25.0 OpenSSL/1.0.1 zlib/1.2.6
12 Dec 13 13:20:55 | | Data directory: C:\ProgramData\BOINC
12 Dec 13 13:20:55 | | Running under account Keifer
12 Dec 13 13:20:55 | | CAL: ATI GPU 0: AMD Radeon HD 7700 series (Capeverde) (CAL version 1.4.1848, 1024MB, 984MB available, 3520 GFLOPS peak)
12 Dec 13 13:20:55 | | OpenCL: AMD/ATI GPU 0: AMD Radeon HD 7700 series (Capeverde) (driver version 1268.1 (VM), device version OpenCL 1.2 AMD-APP (1268.1), 1024MB, 984MB available, 3520 GFLOPS peak)
12 Dec 13 13:20:55 | | OpenCL CPU: AMD Phenom(tm) II X4 960T Processor (OpenCL driver vendor: Advanced Micro Devices, Inc., driver version 1268.1 (sse2), device version OpenCL 1.2 AMD-APP (1268.1))
12 Dec 13 13:20:55 | | Host name: Keifer
12 Dec 13 13:20:55 | | Processor: 4 AuthenticAMD AMD Phenom(tm) II X4 960T Processor [Family 16 Model 10 Stepping 0]
12 Dec 13 13:20:55 | | Processor features: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 htt pni cx16 popcnt syscall nx lm svm sse4a osvw ibs skinit wdt page1gb rdtscp 3dnowext 3dnow
12 Dec 13 13:20:55 | | OS: Microsoft Windows 7: Professional x64 Edition, Service Pack 1, (06.01.7601.00)
12 Dec 13 13:20:55 | | Memory: 16.00 GB physical, 32.00 GB virtual
12 Dec 13 13:20:55 | | Disk: 1.82 TB total, 1.75 TB free
12 Dec 13 13:20:55 | | Local time is UTC -5 hours
12 Dec 13 13:20:55 | SETI@home | URL http://setiathome.berkeley.edu/; Computer ID 6689103; resource share 200
12 Dec 13 13:20:55 | SETI@home Beta Test | URL http://setiweb.ssl.berkeley.edu/beta/; Computer ID 58319; resource share 100
12 Dec 13 13:20:55 | SETI@home | General prefs: from SETI@home (last modified 02-Dec-2013 06:34:39)
12 Dec 13 13:20:55 | SETI@home | Host location: none
12 Dec 13 13:20:55 | SETI@home | General prefs: using your defaults
12 Dec 13 13:20:55 | | Reading preferences override file
12 Dec 13 13:20:55 | | Preferences:
12 Dec 13 13:20:55 | | max memory usage when active: 4095.83MB
12 Dec 13 13:20:55 | | max memory usage when idle: 4095.83MB
12 Dec 13 13:20:55 | | max disk usage: 1.50GB
12 Dec 13 13:20:55 | | max CPUs used: 3
12 Dec 13 13:20:55 | | (to change preferences, visit a project web site or select Preferences in the Manager)
12 Dec 13 13:20:55 | | Not using a proxy
12 Dec 13 13:20:58 | SETI@home | Sending scheduler request: To fetch work.
12 Dec 13 13:20:58 | SETI@home | Requesting new tasks for ATI
12 Dec 13 13:21:00 | SETI@home | Scheduler request completed: got 0 new tasks
12 Dec 13 13:21:05 | SETI@home Beta Test | Sending scheduler request: To fetch work.
12 Dec 13 13:21:05 | SETI@home Beta Test | Requesting new tasks for ATI
12 Dec 13 13:21:07 | SETI@home Beta Test | Scheduler request completed: got 0 new tasks
12 Dec 13 13:28:07 | SETI@home Beta Test | Sending scheduler request: To fetch work.
12 Dec 13 13:28:07 | SETI@home Beta Test | Requesting new tasks for ATI
12 Dec 13 13:28:09 | SETI@home Beta Test | Scheduler request completed: got 0 new tasks
12 Dec 13 13:30:04 | SETI@home | update requested by user
12 Dec 13 13:30:07 | SETI@home Beta Test | update requested by user
12 Dec 13 13:30:09 | SETI@home | Sending scheduler request: Requested by user.
12 Dec 13 13:30:09 | SETI@home | Requesting new tasks for ATI
12 Dec 13 13:30:10 | SETI@home | Scheduler request completed: got 0 new tasks
12 Dec 13 13:30:15 | SETI@home Beta Test | Sending scheduler request: Requested by user.
12 Dec 13 13:30:15 | SETI@home Beta Test | Requesting new tasks for ATI
12 Dec 13 13:30:17 | SETI@home Beta Test | Scheduler request completed: got 0 new tasks
12 Dec 13 13:30:27 | SETI@home Beta Test | Sending scheduler request: To fetch work.
12 Dec 13 13:30:27 | SETI@home Beta Test | Requesting new tasks for ATI
12 Dec 13 13:30:29 | SETI@home Beta Test | Scheduler request completed: got 0 new tasks

By the way, I also noticed that if I elect to use "If no work for selected applications is available, accept work from other applications?" under preferences the response I get from the server is "No tasks are available for the projects I selected" even though thousands upon thousand of AP's are showing on the server status page as being "Results ready to send"

---

The startup looks fine to me. We could try something else....

Since you don't have any tasks;

Change the Minimum Work Buffer to 1 day to keep from receiving many tasks.
Then change your Preferences to also ask for CPU APs.
If you receive CPU APs, change the Preferences back to just the GPU. Hit the Update button a few times to make sure it's still not asking for CPU work.

Then, Hit the Project Reset Button again.
This should force the server to resend the CPU APs as GPU APs.

See what happens.

I received four S@H AP's back on 7 Dec. They were abandoned when I reset earlier.

http://setiathome.berkeley.edu/results.php?hostid=6689103&offset=0&show_names=0&state=0&appid=12

---

Thu Dec 12 14:20:49 2013 | SETI@home | Project is temporarily shut down for maintenance...

Try it again after the tweaks are finished.

Will do. Thanks for your input.

Keifer

---

Why i get no AP opencl_ati_100 since today?
Greetings

I have the same issue, but mostly only on one rig. I followed TBar's suggestion, and once the AP starting rolling out the other day, I have lots (264) of CPU and Nvidia AP's but not one ATI AP. Has anyone else received ATI AP WU's in past couple of days.

Brett

---

Why i get no AP opencl_ati_100 since today?
Greetings

Post the first 40 to 50 lines of the event log.

Since 26 November my GPU has only been getting opencl_ati_cat132 for S@H and opencl_ati5_cat132/AstroPulse v6 v6.06 (cal_ati) for S@H Beta.

A new scheduler was deployed on Seti Beta on the 22nd Nov at Seti Beta to fix the Intel GPU anonymous platform Bug, and a few days later at Seti Main, either we're jumped a few untested changesets,
or we're reverted back a bit (There was a problem with AMD GPUs where the the APR was too high and wasn't getting updated because the Wu was unfeasible, and because the APR wasn't updated the APR remained too high)

Claggy

Has anyone else received ATI AP WU's in past couple of days.

Just checked my only ATI rig, #6912878 and sure enough, I haven't received any APs for the GPU since November 24th. That machine usually only gets a half dozen or so of those tasks every month, but it's now been 18 days since the last one.