OK - Farragut, the new i5-6400 system with the built in Intel 530 card is exhibiting some strange behavior. It keeps Postponing the running task for 30 seconds, then runs for a minute, then Postpones again... WHY??? Below is the FULL Event Log.


12/7/2015 1:41:59 PM | | cc_config.xml not found - using defaults
12/7/2015 1:41:59 PM | | Starting BOINC client version 7.4.42 for windows_x86_64
12/7/2015 1:41:59 PM | | log flags: file_xfer, sched_ops, task
12/7/2015 1:41:59 PM | | Libraries: libcurl/7.39.0 OpenSSL/1.0.1j zlib/1.2.8
12/7/2015 1:41:59 PM | | Data directory: C:\ProgramData\BOINC
12/7/2015 1:41:59 PM | | Running under account Carl_Kuehner
12/7/2015 1:41:59 PM | | OpenCL: Intel GPU 0: Intel(R) HD Graphics 530 (driver version 10.18.15.4271, device version OpenCL 2.0, 1514MB, 1514MB available, 175 GFLOPS peak)
12/7/2015 1:41:59 PM | | OpenCL CPU: Intel(R) Core(TM) i5-6400 CPU @ 2.70GHz (OpenCL driver vendor: Intel(R) Corporation, driver version 5.2.0.10063, device version OpenCL 2.0 (Build 10063))
12/7/2015 1:41:59 PM | SETI@home | Found app_info.xml; using anonymous platform
12/7/2015 1:41:59 PM | | Host name: Farragut
12/7/2015 1:41:59 PM | | Processor: 4 GenuineIntel Intel(R) Core(TM) i5-6400 CPU @ 2.70GHz [Family 6 Model 94 Stepping 3]
12/7/2015 1:41:59 PM | | 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 pbe fsgsbase bmi1 smep bmi2
12/7/2015 1:41:59 PM | | OS: Microsoft Windows 10: x64 Edition, (10.00.10240.00)
12/7/2015 1:41:59 PM | | Memory: 11.89 GB physical, 13.71 GB virtual
12/7/2015 1:41:59 PM | | Disk: 918.76 GB total, 835.90 GB free
12/7/2015 1:41:59 PM | | Local time is UTC -8 hours
12/7/2015 1:41:59 PM | | A new version of BOINC is available. (7.6.9) <a href=http://boinc.berkeley.edu/download.php>Download</a>
12/7/2015 1:41:59 PM | SETI@home | URL http://setiathome.berkeley.edu/; Computer ID 7066606; resource share 365
12/7/2015 1:41:59 PM | SETI@home Beta Test | URL http://setiweb.ssl.berkeley.edu/beta/; Computer ID 64975; resource share 0
12/7/2015 1:41:59 PM | | General prefs: from http://boinc.bakerlab.org/rosetta/ (last modified 09-Apr-2015 15:32:33)
12/7/2015 1:41:59 PM | | Computer location: home
12/7/2015 1:41:59 PM | | General prefs: no separate prefs for home; using your defaults
12/7/2015 1:41:59 PM | | Reading preferences override file
12/7/2015 1:41:59 PM | | Preferences:
12/7/2015 1:41:59 PM | | max memory usage when active: 4262.95MB
12/7/2015 1:41:59 PM | | max memory usage when idle: 6089.93MB
12/7/2015 1:41:59 PM | | max disk usage: 30.00GB
12/7/2015 1:41:59 PM | | (to change preferences, visit a project web site or select Preferences in the Manager)
12/7/2015 1:41:59 PM | | Not using a proxy
12/7/2015 2:06:42 PM | SETI@home | Message from task: 0
12/7/2015 2:06:42 PM | SETI@home | Computation for task 11my11ab.13973.5388.13.12.244_1 finished
12/7/2015 2:06:42 PM | SETI@home | Starting task 11my11ab.13973.5388.13.12.218_1
12/7/2015 2:06:45 PM | SETI@home | Started upload of 11my11ab.13973.5388.13.12.244_1_0
12/7/2015 2:06:47 PM | SETI@home | Finished upload of 11my11ab.13973.5388.13.12.244_1_0
12/7/2015 2:25:48 PM | SETI@home | Sending scheduler request: To report completed tasks.
12/7/2015 2:25:48 PM | SETI@home | Reporting 2 completed tasks
12/7/2015 2:25:48 PM | SETI@home | Not requesting tasks: don't need (CPU: ; Intel GPU: job cache full)
12/7/2015 2:25:49 PM | SETI@home | Scheduler request completed
12/7/2015 2:51:36 PM | SETI@home | Message from task: 0
12/7/2015 2:51:36 PM | SETI@home | Computation for task 11my11ab.13973.5388.13.12.218_1 finished
12/7/2015 2:51:36 PM | SETI@home | Starting task ap_03oc11ac_B1_P0_00243_20151204_09452.wu_1
12/7/2015 2:51:39 PM | SETI@home | Started upload of 11my11ab.13973.5388.13.12.218_1_0
12/7/2015 2:51:41 PM | SETI@home | Finished upload of 11my11ab.13973.5388.13.12.218_1_0
12/7/2015 2:52:37 PM | SETI@home | task postponed 30.000000 sec:
12/7/2015 3:21:35 PM | SETI@home | task postponed 30.000000 sec:
12/7/2015 3:24:13 PM | SETI@home | task postponed 30.000000 sec:
12/7/2015 3:26:15 PM | SETI@home | task postponed 30.000000 sec:
12/7/2015 3:28:17 PM | SETI@home | task postponed 30.000000 sec:
12/7/2015 3:30:54 PM | SETI@home | task postponed 30.000000 sec:
12/7/2015 3:33:32 PM | SETI@home | task postponed 30.000000 sec:
12/7/2015 3:35:33 PM | SETI@home | task postponed 30.000000 sec:
12/7/2015 3:38:11 PM | SETI@home | task postponed 30.000000 sec:
12/7/2015 3:40:50 PM | SETI@home | task postponed 30.000000 sec:
12/7/2015 3:43:32 PM | SETI@home | task postponed 30.000000 sec:
12/7/2015 3:46:12 PM | SETI@home | task postponed 30.000000 sec:
12/7/2015 3:48:13 PM | SETI@home | task postponed 30.000000 sec:
12/7/2015 3:50:49 PM | SETI@home | task postponed 30.000000 sec:
12/7/2015 3:51:51 PM | SETI@home | Sending scheduler request: To report completed tasks.
12/7/2015 3:51:51 PM | SETI@home | Reporting 1 completed tasks
12/7/2015 3:51:51 PM | SETI@home | Requesting new tasks for Intel GPU
12/7/2015 3:51:52 PM | SETI@home | Scheduler request completed: got 1 new tasks
12/7/2015 3:51:54 PM | SETI@home | Started download of 16ap11ah.15244.15609.12.12.93
12/7/2015 3:51:55 PM | SETI@home | Finished download of 16ap11ah.15244.15609.12.12.93
12/7/2015 3:53:22 PM | SETI@home | task postponed 30.000000 sec:
12/7/2015 3:53:22 PM | SETI@home | Starting task 16ap11ah.15244.15609.12.12.93_1
12/7/2015 4:45:48 PM | SETI@home | Message from task: 0
12/7/2015 4:45:48 PM | SETI@home | Computation for task 16ap11ah.15244.15609.12.12.93_1 finished
12/7/2015 4:45:50 PM | SETI@home | Started upload of 16ap11ah.15244.15609.12.12.93_1_0
12/7/2015 4:45:52 PM | SETI@home | Finished upload of 16ap11ah.15244.15609.12.12.93_1_0
12/7/2015 4:46:05 PM | SETI@home | update requested by user
12/7/2015 4:46:09 PM | SETI@home | Sending scheduler request: Requested by user.
12/7/2015 4:46:09 PM | SETI@home | Reporting 1 completed tasks
12/7/2015 4:46:09 PM | SETI@home | Requesting new tasks for Intel GPU
12/7/2015 4:46:10 PM | SETI@home | Scheduler request completed: got 1 new tasks
12/7/2015 4:46:12 PM | SETI@home | Started download of 18fe11ad.19837.4975.7.12.198
12/7/2015 4:46:14 PM | SETI@home | Finished download of 18fe11ad.19837.4975.7.12.198
12/7/2015 4:47:25 PM | SETI@home | task postponed 30.000000 sec:
12/7/2015 4:47:25 PM | SETI@home | Starting task 18fe11ad.19837.4975.7.12.198_1
12/7/2015 5:49:55 PM | SETI@home | task postponed 30.000000 sec:
12/7/2015 6:17:55 PM | SETI@home | Message from task: 0
12/7/2015 6:17:55 PM | SETI@home | Computation for task 18fe11ad.19837.4975.7.12.198_1 finished
12/7/2015 6:17:57 PM | SETI@home | Started upload of 18fe11ad.19837.4975.7.12.198_1_0
12/7/2015 6:17:59 PM | SETI@home | Finished upload of 18fe11ad.19837.4975.7.12.198_1_0
12/7/2015 6:19:30 PM | SETI@home | task postponed 30.000000 sec:
12/7/2015 6:22:06 PM | SETI@home | task postponed 30.000000 sec:
12/7/2015 6:24:07 PM | SETI@home | task postponed 30.000000 sec:
12/7/2015 6:26:08 PM | SETI@home | task postponed 30.000000 sec:
12/7/2015 6:28:44 PM | SETI@home | task postponed 30.000000 sec:
12/7/2015 6:31:20 PM | SETI@home | task postponed 30.000000 sec:
12/7/2015 6:33:56 PM | SETI@home | task postponed 30.000000 sec:
12/7/2015 6:35:56 PM | SETI@home | task postponed 30.000000 sec:
12/7/2015 6:37:57 PM | SETI@home | task postponed 30.000000 sec:
12/7/2015 6:40:33 PM | SETI@home | task postponed 30.000000 sec:
12/7/2015 6:43:09 PM | SETI@home | task postponed 30.000000 sec:
12/7/2015 6:45:45 PM | SETI@home | task postponed 30.000000 sec:
12/7/2015 6:48:20 PM | SETI@home | task postponed 30.000000 sec:
12/7/2015 6:50:55 PM | SETI@home | task postponed 30.000000 sec:
12/7/2015 6:53:30 PM | SETI@home | task postponed 30.000000 sec:
12/7/2015 6:56:06 PM | SETI@home | task postponed 30.000000 sec:
12/7/2015 6:58:43 PM | SETI@home | task postponed 30.000000 sec:
12/7/2015 7:01:20 PM | SETI@home | task postponed 30.000000 sec:


HELP!!!


TL

---

TimeLord04
Have TARDIS, will travel...
Come along K-9!
Join Calm Chaos

Just installed CPU-Z, GPU-Z, and Hardware Monitor. All show the system as Normal. No excess heat anywhere. No problems...

Yet, it still keeps Postponing work on the current WU. It is also stuck at just over 12.6% Done and 16:55:26 to go; while Elapsed Time keeps climbing.

Is this a bad AP WU???

Computer ID is 7066606


TL

---

TimeLord04
Have TARDIS, will travel...
Come along K-9!
Join Calm Chaos

OK - Farragut, the new i5-6400 system with the built in Intel 530 card is exhibiting some strange behavior. It keeps Postponing the running task for 30 seconds, then runs for a minute, then Postpones again... WHY??? Below is the FULL Event Log.

Dig down into the slot directory and find out why from the application's stderr.txt

Claggy

OK - Farragut, the new i5-6400 system with the built in Intel 530 card is exhibiting some strange behavior. It keeps Postponing the running task for 30 seconds, then runs for a minute, then Postpones again... WHY??? Below is the FULL Event Log.

Dig down into the slot directory and find out why from the application's stderr.txt

Claggy

This is what I find - NOTHING... [EDIT: Rechecked after 9 PM - PST, Found LOTS of NEW info in the Stderr Report that WAS NOT THERE when this Post was created.]

Stderr Output.

As for the Slot Directory, I have NO CLUE how to do this...


TL

---

TimeLord04
Have TARDIS, will travel...
Come along K-9!
Join Calm Chaos

I have just aborted the Unit, and new work has downloaded. Since it is now past 9 PM - Pacific, BOINC will pick up and crunch tomorrow when Farragut is turned back on.


TL

---

TimeLord04
Have TARDIS, will travel...
Come along K-9!
Join Calm Chaos

Let's preserve just one cycle from stderr in case we get the opportunity to debug it.

Running on device number: 0
Priority of worker thread raised successfully
Priority of process adjusted successfully, below normal priority class used
OpenCL platform detected: Intel(R) Corporation
BOINC assigns device 0
Info: BOINC provided OpenCL device ID used
Used GPU device parameters are:
Number of compute units: 23
Single buffer allocation size: 256MB
Total device global memory: 1488MB
max WG size: 256
local mem type: Real
-unroll default value used: 18
-ffa_block default value used: 5888
-ffa_block_fetch default value used: 2944

Build features: Non-graphics BLANKIT OpenCL USE_OPENCL_INTEL TWIN_FFA OCL_SYNCHED OCL_ZERO_COPY COMBINED_DECHIRP_KERNEL FFTW USE_INCREASED_PRECISION USE_SSE2 x86
CPUID: Intel(R) Core(TM) i5-6400 CPU @ 2.70GHz

Cache: L1=64K L2=256K

CPU features: FPU TSC PAE CMPXCHG8B APIC SYSENTER MTRR CMOV/CCMP MMX FXSAVE/FXRSTOR SSE SSE2 HT SSE3 SSSE3 FMA3 SSE4.1 SSE4.2 AVX
### Restart at 11.71 percent.
state.fold_buf_size_short=65536; state.fold_buf_size_long=262144
ERROR: OpenCL kernel/call 'ReadBuffer(gpu_power->power_range_ptr)' call failed (-5) in file ..\..\ap_client_main.cpp near line 4351.
Waiting 30 sec before restart...

I would suggest to run fail save params like i had to do on my GPU.
Maybe thats an issue with OpenCL 2.0 based GPU`s.

Search a file called ap_cmdline_win_x86_SSE2_OpenCL_Intel.txt and add the following with notepad

-unroll 8 -ffa_block 2048 -ffa_block_fetch 1024.

Save as text.

This will run the next task with reduced values and see if its getting better.

---

With each crime and every kindness we birth our future.

I've got the same issue i.e task postponed for 30 secs with an Intel HD 4400 in my Core i3 4030U Haswell (OpenCL 1.2).
With Astropulse task.

Marco.

---

I've got the same issue i.e task postponed for 30 secs with an Intel HD 4400 in my Core i3 4030U Haswell (OpenCL 1.2).
With Astropulse task.

Marco.

Not much i can say since your host is hidden.
You can also try my suggestion but most important do you have a CPU core reserved ?

---

With each crime and every kindness we birth our future.

state.fold_buf_size_short=65536; state.fold_buf_size_long=262144
ERROR: OpenCL kernel/call 'ReadBuffer(gpu_power->power_range_ptr)' call failed (-5) in file ..\..\ap_client_main.cpp near line 4351.
Waiting 30 sec before restart...

I need CLinfo output for affected hosts.

state.fold_buf_size_short=65536; state.fold_buf_size_long=262144
ERROR: OpenCL kernel/call 'ReadBuffer(gpu_power->power_range_ptr)' call failed (-5) in file ..\..\ap_client_main.cpp near line 4351.
Waiting 30 sec before restart...

I need CLinfo output for affected hosts.

For those who are prepared to find the BOINC data directory, and from that, the current slot directory, CLinfo can be found towards the top of stderr.txt

Otherwise, if you notice behaviour like this, please abort the task within 20 minutes (so before 40 restarts are possible) - this should allow the first entries in stderr.txt to be visible on this website.

And please remember to give us the task name or number so we can find it!

I would suggest to run fail save params like i had to do on my GPU.
Maybe thats an issue with OpenCL 2.0 based GPU`s.

Search a file called ap_cmdline_win_x86_SSE2_OpenCL_Intel.txt and add the following with notepad

-unroll 8 -ffa_block 2048 -ffa_block_fetch 1024.

Save as text.

This will run the next task with reduced values and see if its getting better.

I was about to suggest the same. Isn't -unroll 18 a little aggressive for an iGPU? My 6970 Cayman uses unroll 18 and it's a much more high end GPU than a CPU iGPU.

OK - I went to the BOINC Data Folder, to the Slots Folder in Farragut and found this for the current unit being worked on:

--------------------------------------------------------------

Running on device number: 0
Priority of worker thread raised successfully
Priority of process adjusted successfully, below normal priority class used
OpenCL platform detected: Intel(R) Corporation
BOINC assigns device 0
Info: BOINC provided OpenCL device ID used

Build features: SETI7 Non-graphics OpenCL USE_OPENCL_INTEL OCL_SYNCHED OCL_ZERO_COPY OCL_ZERO_COPY_APU OCL_CHIRP3 FFTW USE_SSSE3 x86
CPUID: Intel(R) Core(TM) i5-6400 CPU @ 2.70GHz

Cache: L1=64K L2=256K

CPU features: FPU TSC PAE CMPXCHG8B APIC SYSENTER MTRR CMOV/CCMP MMX FXSAVE/FXRSTOR SSE SSE2 HT SSE3 SSSE3 FMA3 SSE4.1 SSE4.2 AVX
OpenCL-kernels filename : MultiBeam_Kernels_r2929.cl
ar=0.431053 NumCfft=195053 NumGauss=1096563930 NumPulse=226476462757 NumTriplet=452883965687
Currently allocated 128 MB for GPU buffers
In v_BaseLineSmooth: NumDataPoints=1048576, BoxCarLength=8192, NumPointsInChunk=32768

Windows optimized S@H v7 application
Based on Intel, Core 2-optimized v8-nographics V5.13 by Alex Kan
SSSE3xj Win32 Build 2929 , Ported by : Raistmer, JDWhale

SETI7 update by Raistmer

OpenCL version by Raistmer, r2929

Number of OpenCL platforms: 1


OpenCL Platform Name: Intel(R) OpenCL
Number of devices: 1
Max compute units: 23
Max work group size: 256
Max clock frequency: 950Mhz
Max memory allocation: 390280806
Cache type: Read/Write
Cache line size: 64
Cache size: 524288
Global memory size: 1561123226
Constant buffer size: 65536
Max number of constant args: 8
Local memory type: Scratchpad
Local memory size: 65536
Queue properties:
Out-of-Order: No
Name: Intel(R) HD Graphics 530
Vendor: Intel(R) Corporation
Driver version: 10.18.15.4271
Version: OpenCL 2.0
Extensions: cl_intel_accelerator cl_intel_advanced_motion_estimation cl_intel_ctz cl_intel_d3d11_nv12_media_sharing cl_intel_dx9_media_sharing cl_intel_motion_estimation cl_intel_simultaneous_sharing cl_intel_subgroups cl_khr_3d_image_writes cl_khr_byte_addressable_store cl_khr_d3d10_sharing cl_khr_d3d11_sharing cl_khr_depth_images cl_khr_dx9_media_sharing cl_khr_fp16 cl_khr_gl_depth_images cl_khr_gl_event cl_khr_gl_msaa_sharing cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_gl_sharing cl_khr_icd cl_khr_image2d_from_buffer cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_mipmap_image cl_khr_mipmap_image_writes cl_khr_spir


Work Unit Info:
...............
Credit multiplier is : 2.85
WU true angle range is : 0.431053
Used GPU device parameters are:
Number of compute units: 23
Single buffer allocation size: 64MB
Total device global memory: 1488MB
max WG size: 256
local mem type: Real
period_iterations_num=20
Spike: peak=24.42871, time=97.31, d_freq=1418849979.9, chirp=0.91687, fft_len=64k
Spike: peak=24.25371, time=87.24, d_freq=1418851529.15, chirp=-13.573, fft_len=128k
Spike: peak=24.94658, time=87.24, d_freq=1418851529.14, chirp=-13.574, fft_len=128k
Spike: peak=24.43636, time=87.24, d_freq=1418851529.14, chirp=-13.575, fft_len=128k
GPU device sync requested... ...GPU device synched
Termination request detected or computations are finished. GPU device synched, exiting...

--------------------------------------------------------

The failed and aborted unit's info had already been deleted; but, I hope that this info helps Raistmer.


TL

---

TimeLord04
Have TARDIS, will travel...
Come along K-9!
Join Calm Chaos

I would suggest to run fail save params like i had to do on my GPU.
Maybe thats an issue with OpenCL 2.0 based GPU`s.

Search a file called ap_cmdline_win_x86_SSE2_OpenCL_Intel.txt and add the following with notepad

-unroll 8 -ffa_block 2048 -ffa_block_fetch 1024.

Save as text.

This will run the next task with reduced values and see if its getting better.

Mike,

I found the file, and opened it. At present, there's NOTHING in the file. To make sure, I closed Notepad without doing anything in Notepad, and looked at the File Size. Current File Size is 0 K.

I put your text line in there, and saved as text per your request. File Size is now 1 K.

Thanks Mike, I hope this works. :-)


TL

PS: Did I need the period at the end of the line??? Right now, it is NOT there, I can put it in if it is needed.

---

TimeLord04
Have TARDIS, will travel...
Come along K-9!
Join Calm Chaos

I would suggest to run fail save params like i had to do on my GPU.
Maybe thats an issue with OpenCL 2.0 based GPU`s.

Search a file called ap_cmdline_win_x86_SSE2_OpenCL_Intel.txt and add the following with notepad

-unroll 8 -ffa_block 2048 -ffa_block_fetch 1024.

Save as text.

This will run the next task with reduced values and see if its getting better.

I was about to suggest the same. Isn't -unroll 18 a little aggressive for an iGPU? My 6970 Cayman uses unroll 18 and it's a much more high end GPU than a CPU iGPU.

That`s what i feared when it was implemented.
On discret GPU`s it`s no problem.
One unroll for each compute unit with max of 18.
On my R9 380 i also had to reduce unroll to get it working properly.
On iGPU its much worse but the first iGPU`s only had 4 to 10 CU`s.

Sometimes i`m talking against walls.

---

With each crime and every kindness we birth our future.

Thanks, Mike, for the PM response.

New Command Line is " -unroll 8 -ffa_block 2048 -ffa_block_fetch 1024"; without the quotes, and with the space before the -unroll as you have instructed.


Thanks,


TL

---

TimeLord04
Have TARDIS, will travel...
Come along K-9!
Join Calm Chaos

The failed and aborted unit's info had already been deleted; but, I hope that this info helps Raistmer.


TL

It's from MB stderr, error was reported for AP. Do you have issues with MB tasks also?
And reduce unroll to less than 10 for AP then check if AP error remains.

The failed and aborted unit's info had already been deleted; but, I hope that this info helps Raistmer.


TL

It's from MB stderr, error was reported for AP. Do you have issues with MB tasks also?
And reduce unroll to less than 10 for AP then check if AP error remains.

Unfortunately, the AP info was deleted when the MBs started running. This is all I could obtain from the Stderr.txt file...

I will wait for another AP to come in, and then I will grab the Stderr.txt from that unit. It may be awhile, though...

...And, no, I'm not having issues with MB, just the AP. I've implemented Mike's Command Line, and now Unroll should be 8 for Farragut, ID 7066606.

Thanks Raistmer. :-)


TL

---

TimeLord04
Have TARDIS, will travel...
Come along K-9!
Join Calm Chaos

That`s what i feared when it was implemented.
On discret GPU`s it`s no problem.
One unroll for each compute unit with max of 18.
On my R9 380 i also had to reduce unroll to get it working properly.
On iGPU its much worse but the first iGPU`s only had 4 to 10 CU`s.

Sometimes i`m talking against walls.

Just right pass phrase required to get through the wall ;) Sometimes on Elvish though

If default unroll selection too high for modern iGPU devices it can be corrected with next release.
Keep in mind that all development and testing was done on entry-level CU 6 IvyBridge class iGPU. I use what I have :P iGPU device in question has 23 CU and another generation...

...And, no, I'm not having issues with MB, just the AP. I've implemented Mike's Command Line, and now Unroll should be 8 for Farragut, ID 7066606.

Thanks Raistmer. :-)


TL

So lets await till new task come in to check if unroll adjustment solves issue.