I'm also having a lot of invalids.

In fact most WU's are returning as Invalid, very few errors and very few accepted.


A Sample Invalid:

Result Name: X0930100641201200806191153_ 1--
<core_client_version>7.0.44</core_client_version>
<![CDATA[
<stderr_txt>
Commandline: projects/www.worldcommunitygrid.org/wcg_hcc1_img_7.05_windows_intelx86__ati_hcc1 --zipfile X0930100641201200806191153.zip --imagelist images.txt --device 0
<app_init_data>
<major_version>7</major_version>
<minor_version>0</minor_version>
<release>44</release>
<app_version>705</app_version>
<app_name>hcc1</app_name>
<acct_mgr_url>http://bam.boincstats.com/</acct_mgr_url>
<project_preferences>


<color_scheme>Tahiti Sunset</color_scheme>
<max_frames_sec>7</max_frames_sec>
<max_gfx_cpu_pct>5.0</max_gfx_cpu_pct>
</project_preferences>

<project_dir>C:\ProgramData\BOINC/projects/www.worldcommunitygrid.org</project_dir>
<boinc_dir>C:\ProgramData\BOINC</boinc_dir>
<wu_name>X0930100641201200806191153</wu_name>
<result_name>X0930100641201200806191153_1</result_name>
<comm_obj_name>boinc_0</comm_obj_name>
<slot>4</slot>
<wu_cpu_time>0.000000</wu_cpu_time>
<starting_elapsed_time>0.000000</starting_elapsed_time>
<using_sandbox>0</using_sandbox>
<user_total_credit>2061187.026927</user_total_credit>
<user_expavg_credit>227.315472</user_expavg_credit>
<host_total_credit>374911.291978</host_total_credit>
<host_expavg_credit>227.315487</host_expavg_credit>
<resource_share_fraction>1.000000</resource_share_fraction>
<checkpoint_period>60.000000</checkpoint_period>
<fraction_done_start>0.000000</fraction_done_start>
<fraction_done_end>1.000000</fraction_done_end>
<gpu_type>ATI</gpu_type>
<gpu_device_num>0</gpu_device_num>
<gpu_opencl_dev_index>0</gpu_opencl_dev_index>
<ncpus>1.000000</ncpus>
<rsc_fpops_est>25520135981107.000000</rsc_fpops_est>
<rsc_fpops_bound>510402719622140.000000</rsc_fpops_bound>
<rsc_memory_bound>78643200.000000</rsc_memory_bound>
<rsc_disk_bound>50000000.000000</rsc_disk_bound>
<computation_deadline>1359133609.000000</computation_deadline>
<vbox_window>0</vbox_window>
</app_init_data>
INFO: gpu_type set in init_data.xml to ATI
INFO: gpu_device_num set in init_data.xml to 0
Boinc requested ATI gpu device number0
Unzipping input images ../../projects/www.worldcommunitygrid.org/X0930100641201200806191153_X0930100641201200806191153.zip
Processing jobdescription
Number of Images defined in image list is 2
Found compute platform Advanced Micro Devices, Inc.
Selecting this platform
CL_DEVICE_NAME: Cypress
CL_DEVICE_VENDOR: Advanced Micro Devices, Inc.
CL_DEVICE_VERSION: 1084.4 (VM)
CL_DEVICE_MAX_COMPUTE_UNITS: 
CL_DEVICE_MAX_WORK_ITEM_DIMENSIONS: 3
CL_DEVICE_MAX_WORK_ITEM_SIZES: 256 / 256 / 256
CL_DEVICE_MAX_WORK_GROUP_SIZE: 256
CL_DEVICE_MAX_CLOCK_FREQUENCY: 600 MHz
CL_DEVICE_ADDRESS_BITS: 32
CL_DEVICE_MAX_MEM_ALLOC_SIZE: 512 MByte
CL_DEVICE_GLOBAL_MEM_SIZE: 1024 MByte
CL_DEVICE_ERROR_CORRECTION_SUPPORT: no
CL_DEVICE_LOCAL_MEM_TYPE: local
CL_DEVICE_LOCAL_MEM_SIZE: 32 KByte
CL_DEVICE_MAX_CONSTANT_BUFFER_SIZE: 64 KByte
CL_DEVICE_QUEUE_PROPERTIES: CL_QUEUE_PROFILING_ENABLE
CL_DEVICE_EXTENSIONS:
cl_khr_fp64
cl_amd_fp64
cl_khr_global_int32_base_atomics
cl_khr_global_int32_extended_atomics
cl_khr_local_int32_base_atomics
cl_khr_local_int32_extended_atomics
cl_khr_3d_image_writes
cl_khr_byte_addressable_store
cl_khr_gl_sharing
cl_ext_atomic_counters_32
cl_amd_device_attribute_query
cl_amd_vec3
cl_amd_printf
cl_amd_media_ops
cl_amd_media_ops2
cl_amd_popcnt
cl_khr_d3d10_sharing
Estimated kernel execution time = 0.35996 [sec]
Starting analysis of X0930100641201200806191153.jp2...
Extracting GLCM features...
Total kernel time: 205.902649 (1026 kernel executions)
Total memory transfer time: 1.707382
Average kernel time: 0.200685
Min kernel time: 0.189129 (dx=11 dy=23 sample_dist=24 )
Max kernel time: 0.215030 dx=2 dy=1 sample_dist=1
INFO: GPU calculations complete.
Total time for X0930100641201200806191153.jp2: 297 seconds
Finished Image #0, pctComplete = 0.500000
Starting analysis of X0930100640389200806191206.jp2...
Extracting GLCM features...
Total kernel time: 248.168808 (1026 kernel executions)
Total memory transfer time: 3.387018
Average kernel time: 0.241880
Min kernel time: 0.217186 (dx=25 dy=3 sample_dist=24 )
Max kernel time: 0.257577 dx=2 dy=1 sample_dist=1
INFO: GPU calculations complete.
Total time for X0930100640389200806191206.jp2: 342 seconds
Finished Image #1, pctComplete = 1.000000
CPU time used = 188.402408
15:25:54 (3476): called boinc_finish

</stderr_txt>
]]>


System is i7-920 with ATI 5830.

Win 7/64 Ult with Catalyst 13.1

Nothing is OC'd (in fact, based on suggestions in similar threads, I have reduced the GPU clocks to the lowest values Catalyst supports - 600/900 to no avail).

Any Suggestions?

Is this why my pages of PVs are going up?

Hi andzgrid

I am not sure if that is a critique of my post or you wish to explore other avenues.

I would say that in general I tend to post my experiences when running my rigs. I don't intend these posts to do anything more than indicate to others that I found a solution to a similar problem that may, just may, help someone with issues.

I understand that there may be another reason but your theory of the reference set? well in this particular instance I find myself finding it hard to believe that the simple comparison of work done by different computers, well the results anyway, could be prone to error...... Unless, of course, the validation system only matches the first pair of similar results then rejects all others. Is this the case? Are we seeing invalids in sufficiently large numbers to warrant having 3 matching results? Were this the case then surely those that run the system would make the appropriate adjustments

But I am an open minded sort who acknowledges that others with greater skills and knowledge may find other, better, solutions.

In line with the ethic of most who contribute here I just proffer what I can, when I can and hope that I am not seen as doing something other than this. Just in case you feel otherwise..... feel free to criticise.

From what has been said in the past, if I remember correctly, the validity of a work unit is set by some parameters set by the researchers. For some sciences of more than a quorum of one, the reference set, as I understand it, is how closely the two(or more) results match, provided they both conform to originally set parameters. From personal experience I have noticed the invalids for me have come from either malformed work units or a hardware glitch on my end. I do not overclock, but overclocking too much will cause an invalid condition to occur (as many have attested in the forums.) Another cause is a failing power supply, which I have personally experienced. I recall seeing only a few references to faulty memory, but it too can cause this condition as well as overheating. There are undoubtedly other causes. Perhaps one of the techs or researchers could chime in to further illuminate this topic.
Cheers

Validation, even for invalids on zero redundant sciences are often 3 copies, where the agreement of 2 singles out the 3rd. For quorum 2 which HCC1 is *of course*, there's a 3 way distribution at minimum to determine which of the 3 is invalid. This cycle can actually go up to 5 or 7 before a task is considered an out-take [put aside to the review list].

Validation rules have a few basics:

1) Must meet a minimum set of output conditions and pass an included minitest at start of task [zero redundancy sciences and for GPU a performance test], lets call that first level quality control.
2) Must closely match wingman, by again another set of checks.

Errors are outright during execution, rarely will they get to a Pending Validation state. Invalids always pass through PVal and PVer... Pending Verification is a second /subsequent phase waiting for revalidation.

In a nutshell.