A week into be Jocelyn-less, I'm wondering what work-arounds people have to view the results of their crunched work units.

Issue is, even when Jocelyn comes back, probably 75% of your own results for the week will have already been cleaned off the reporting database, so we won't be able to browse granularly the points, check error WU's, drill down into triplet & gaussian counts, etc. Only the WU's where Wingmen haven't reported, or were reported in the previous 24h will be viewable.

Ideally (inside Boinc), the property page of completed & uploaded WU's would list some of this stuff, before actual reporting takes place (which we can optionally control anyhow), but it seems the data isn't really in the "_0" files. (Too bad the Boinc standard didn't accomodate a <summary><\summary> freeform field).

I looked at the descriptions of BoincSpy and BoincWatch, but they don't sound like they do what I'm after, which is to approximately duplicate the results page locally minus all aspects of the wingman (afterall, the data was more-or-less here and processed in the first place, and yes I do know misc validation, calculations, & analysis do take place server side -- and that your local view isn't necessarily what the Server will grant either).

Cheers.

A week into be Jocelyn-less, I'm wondering what work-arounds people have to view the results of their crunched work units.

Issue is, even when Jocelyn comes back, probably 75% of your own results for the week will have already been cleaned off the reporting database, so we won't be able to browse granularly the points, check error WU's, drill down into triplet & gaussian counts, etc. Only the WU's where Wingmen haven't reported, or were reported in the previous 24h will be viewable.

Ideally (inside Boinc), the property page of completed & uploaded WU's would list some of this stuff, before actual reporting takes place (which we can optionally control anyhow), but it seems the data isn't really in the "_0" files. (Too bad the Boinc standard didn't accomodate a <summary><\summary> freeform field).

I looked at the descriptions of BoincSpy and BoincWatch, but they don't sound like they do what I'm after, which is to approximately duplicate the results page locally minus all aspects of the wingman (afterall, the data was more-or-less here and processed in the first place, and yes I do know misc validation, calculations, & analysis do take place server side -- and that your local view isn't necessarily what the Server will grant either).

Cheers.

Have you tried BoincView? It has a logging function as well. Altho I am unsure if it records the data you are wanting.

---

SETI@home classic workunits: 93,865 CPU time: 863,447 hours
Join the [url=http://tinyurl.com/8y46zvu]BP6/VP6 User Group[

Thx Hal, I looked at screenshots of BoincView and Mike's Boinc.Net, and everything else I could easily find on Google Images. Nothing seems to capture the key result pieces and nothing seems to even capture the non-key pieces in an easy running log format. Oh well... I guess the addon imagineers figure that without wingman data, no sense in doing the other stuff.

I am missing such functions since the days of SETIwatch (logging top Gaussian/Pulse/Triplet for every single WorkUnit into local database).

By now I originally had expected that some BOINC-wide logging would exist (either per Project Site or with a dedicated BOINC tool). But apart from isolated and project-specialized tools requiring manual work (CPDN) or Projects like Rosetta (detailled results summary) usually nothing is shown concerning results.

Maybe we see something like that when NTPCKR is fully operational but I'm not too hopeful because of other priorities.

---

Personally, I just get mildly annoyed that I cant check exactly what my machines are doing easily. But I can live with it. Backup server will be online in due course and things will be back to normal.

Ian

Have you tried BoincView? It has a logging function as well. Altho I am unsure if it records the data you are wanting.

Here is one BOINCview prepared earlier...

<log>
<tool>
  <name>BoincView</name>
  <version>1.4.2</version>
</tool>
<result>
  <name>02se09ab.22745.10701.13.10.121_0</name>
  <final_cpu_time>43.5000000000</final_cpu_time>
  <completed_time>0.000000</completed_time>
  <last_access>2455116.2152365740</last_access>
  <app>
    <name>setiathome_enhanced</name>
    <major_version>6</major_version>
    <minor_version>08</minor_version>
  </app>
  <client>
  </client>
</result>
</log>
<platform_name>windows_intelx86</platform_name>
<core_client_major_version>6</core_client_major_version>
<core_client_minor_version>10</core_client_minor_version>
<core_client_release>13</core_client_release>
<cpu_sched_period>0</cpu_sched_period>
<cpu_sched_work_done_this_period>0.000000</cpu_sched_work_done_this_period>
<user_run_request>1</user_run_request>
<user_network_request>2</user_network_request>
<work_fetch_no_new_work>0</work_fetch_no_new_work>
<cpu_benchmarks_pending>0</cpu_benchmarks_pending>
<network_status>2</network_status>
<task_suspend_reason>0</task_suspend_reason>
<network_suspend_reason>8</network_suspend_reason>
<host_venue>home</host_venue>
<host_info>
  <timezone>39600</timezone>
  <domain_name>Maul</domain_name>
  <ip_addr>10.1.1.6</ip_addr>
  <host_cpid>ca0239ef733b8e8ffec44c814777de32</host_cpid>
  <p_vendor>GenuineIntel</p_vendor>
  <p_model>Intel(R) Core(TM) i7 CPU         920  @ 2.67GHz</p_model>
  <p_capabilities>x86 Family 6 Model 26 Stepping 4</p_capabilities>
  <p_ncpus>8</p_ncpus>
  <p_fpops>2674765617.7786598200</p_fpops>
  <p_iops>6859032871.2105770100</p_iops>
  <p_membw>125000000.0000000000</p_membw>
  <p_calculated>12899560724.4375000000</p_calculated>
  <m_nbytes>3210784768.0000000000</m_nbytes>
  <m_cache>262144.0000000000</m_cache>
  <m_swap>5188280320.0000000000</m_swap>
  <d_total>320061046784.0000000000</d_total>
  <d_free>310871326720.0000000000</d_free>
  <os_name>Microsoft Windows XP</os_name>
  <os_version>Professional x86 Edition, Service Pack 3, (05.01.2600.00)</os_version>
  <accelerators></accelerators>
</host_info>
<time_stats>
  <last_update>12899714984.647375</last_update>
  <on_frac>0.5927040000</on_frac>
  <active_frac>0.9998940000</active_frac>
  <connected_frac>0.4993970000</connected_frac>
  <cpu_efficiency>1.0000000000</cpu_efficiency>
</time_stats>
<net_stats>
  <bwup>16368.3884250000</bwup>
  <bwdown>174996.6496500000</bwdown>
</net_stats>
<project>
  <master_url>http://setiathome.berkeley.edu/</master_url>
  <project_name>SETI@home</project_name>
  <sym_store></sym_store>
  <user_name>MarkJ</user_name>
  <team_name></team_name>
  <host_venue>home</host_venue>
  <email_hash>*****</email_hash>
  <cross_project_id>a489f8719b30db60d6c793d095111a3f</cross_project_id>
  <user_total_credit>7061755.077247</user_total_credit>
  <user_expavg_credit>28199.765717</user_expavg_credit>
  <user_create_time>1203242636.000000</user_create_time>
  <rpc_seqno>6239</rpc_seqno>
  <hostid>4848148</hostid>
  <host_total_credit>1662005.065958</host_total_credit>
  <host_expavg_credit>11303.514473</host_expavg_credit>
  <host_create_time>1237553440.000000</host_create_time>
  <nrpc_failures>0</nrpc_failures>
  <master_fetch_failures>0</master_fetch_failures>
  <min_rpc_time>12899706750.834875</min_rpc_time>
  <disk_usage>185117339.000000</disk_usage>
  <debt>69979.078527</debt>
  <long_term_debt>-254812.488178</long_term_debt>
  <resource_share>25.000000</resource_share>
  <duration_correction_factor>0.266899</duration_correction_factor>
  <sched_rpc_pending>0</sched_rpc_pending>
  <ams_resource_share>0.000000</ams_resource_share>
  <last_rpc_time>0.000000</last_rpc_time>
  <project_files_downloaded_time>0.000000</project_files_downloaded_time>
  <gui_urls>
<gui_url>
  <name>Message boards</name>
  <description>Correspond with other users on the SETI@home message boards</description>
  <url>http://setiathome.berkeley.edu/forum_index.php</url>
</gui_url>
<gui_url>
  <name>Help</name>
  <description>Ask questions and report problems</description>
  <url>http://setiathome.berkeley.edu/forum_help_desk.php</url>
</gui_url>
<gui_url>
  <name>Your account</name>
  <description>View your account information and credit totals</description>
  <url>http://setiathome.berkeley.edu/show_user.php?userid=8823491</url>
</gui_url>
<gui_url>
  <name>Your preferences</name>
  <description>View and modify your SETI@home account profile and preferences</description>
  <url>http://setiathome.berkeley.edu/home.php</url>
</gui_url>
<gui_url>
  <name>Your results</name>
  <description>View your last week (or more) of computational results and work</description>
  <url>http://setiathome.berkeley.edu/results.php?userid=8823491</url>
</gui_url>
<gui_url>
  <name>Your computers</name>
  <description>View a listing of all the computers on which you are running SETI@Home</description>
  <url>http://setiathome.berkeley.edu/hosts_user.php?userid=8823491</url>
</gui_url>
<gui_url>
  <name>Donate</name>
  <description>Donate to SETI@home</description>
  <url>http://setiathome.berkeley.edu/sah_donate.php</url>
</gui_url>
  </gui_urls>
</project>

---

BOINC blog

Thx All

Have you tried BoincView? It has a logging function as well. Altho I am unsure if it records the data you are wanting.

I will look closer at BoincView and BoincSpy by installing them (along with .net 2.0 and other stuff I had been hoping to avoid on my cruncher system, oh well).

An old project called BoincLogX looked promising, but after installing it doesn't list things like Spikes, Gaussians, & triplets. Also, it's from 2006, and I had to hand edit the INI file since it's GUI directory picker has no notion of the virtual directories used by Boinc (not a problem once you figure out why nothing in it seemed to work)(plus it seems to periodically poll the work directory which really is very expensive and very out of lock step with the the multi cores of today not to mention the speed at which CUDA can now complete and move on to the next WU).

Thx All

Have you tried BoincView? It has a logging function as well. Altho I am unsure if it records the data you are wanting.

I will look closer at BoincView and BoincSpy by installing them (along with .net 2.0 and other stuff I had been hoping to avoid on my cruncher system, oh well).

An old project called BoincLogX looked promising, but after installing it doesn't list things like Spikes, Gaussians, & triplets. Also, it's from 2006, and I had to hand edit the INI file since it's GUI directory picker has no notion of the virtual directories used by Boinc (not a problem once you figure out why nothing in it seemed to work)(plus it seems to periodically poll the work directory which really is very expensive and very out of lock step with the the multi cores of today not to mention the speed at which CUDA can now complete and move on to the next WU).

Probably the best approach is run with No New Tasks turned on so the core client doesn't report completed work too soon, save a copy of client_state.xml before each time you allow new tasks. The stderr sections in those saved files are the same as shown on a Task details page, and the <fpops_cumulative> value can be converted to an approximation of the credit claim (not exact since we don't know the server-side credit_multiplier value for any specific "sent" date).

The interval at which BoincLogX checks files can be adjusted. I'm not sure its default checking is particularly burdensome compared to BOINC Manager which does a complete set of RPCs every second, I run without the Manager 99.9% of the time but do have BoincLogX always running. I don't have a fast multicore system though. BoincLogX doesn't save the fpops_cumulative value so those particularly interested in credit might be disappointed, but it does save all signals it extracts from the MB result files; there's no built-in display but the csv files are there for viewing in a spreadsheet program.

The ideal would of course be a version of the BOINC core client which would log the pertinent details before deleting them. But that's a feature that only a minority of users would want or understand so is not likely to be added by the BOINC devs.

                                                                Joe

Thx J.S, all points valid and well taken.

The interval at which BoincLogX checks files can be adjusted. I'm not sure its default checking is particularly burdensome compared to BOINC Manager which does a complete set of RPCs every second

I would think the directory dredge burden takes over after you have a few hundred WU's in inventory (certainly before a few thousand), or at least I would hope the RPCs are efficient in internally tracking state changes (and yes, I know BMgr starts to wobble itself around 2000 WUs, so the RPCs are not heaven either on frequent large data sets, although I have a feeling most of that wobble is in the UI control itself). If LogX had used a filter driver (yes, much harder code), then that might have been the best of both sides.

The ideal would of course be a version of the BOINC core client which would log the pertinent details before deleting them.

Yes! The tricky part is how to meaningfully represent the succint progress or the succint summation of interest per WU when every project is different (the notion of "credits" is basically a universal leveler which abstracts away that problem). That's kind of why I suggested this inside the existing Task Tab property page (or a single column at the end) as a free form data field (ie: XML of course, but stripped at display time), at least then human eyes could scan it (which is what I do today by clicking down into the web Results page when it's up).

So what I was hoping for was a local log of the STATE files from the \slot directories. IE: after a WU completes but before Upload takes place, a process fetched and parsed a copy (at least before the next WU wipes it out). The complicator as said before is now how do you present that meaningfully... a tabbed pivot table is kind of heavy weight.

But that's a feature that only a minority of users would want or understand.

And that's the stumper for me. I would think at least with this project (S@H), well over 80% ARE interested in knowing if one of their WU's hit triple digits for Gaussians, Triplets, Pulses, and Spikes. Sort of a local NTPCkr. I see tools which are close designed for Seti Classic, but were never rev'd for Boinc.

And that's the stumper for me. I would think at least with this project (S@H), well over 80% ARE interested in knowing if one of their WU's hit triple digits for Gaussians, Triplets, Pulses, and Spikes. Sort of a local NTPCkr. I see tools which are close designed for Seti Classic, but were never rev'd for Boinc.

According to BOINCstats, there are more than 1 million participants at SETI@Home.

Last time I asked, I was told that 8,800 of those had posted in the forums at least once.

No idea how many read the forums, but I'd bet that number is surprisingly small.

Our picture of how BOINC and SETI@Home should work is colored by the fact that we in the forums represent less than 1% of all SETI@Home crunchers.

---

According to BOINCstats, there are more than 1 million participants at SETI@Home.

Take a closer look. More than 900,000 of those have a current monthly credit of zero ... which in my book is dead (or at least profoundly hibernating). They'd probably have no opinion on if the "advanced" UI view were completely eliminated.

Our picture of how BOINC and SETI@Home should work is colored by the fact that we in the forums represent less than 1% of all SETI@Home crunchers.

Point is very well taken, but don't undervalue that 1% by body count too much, as they probably represent 40% of the cobblestones and 75% of the "tinker & ponder" time spent this year by the active S@H community.

Our picture of how BOINC and SETI@Home should work is colored by the fact that we in the forums represent less than 1% of all SETI@Home crunchers.

Point is very well taken, but don't undervalue that 1% by body count too much, as they probably represent 40% of the cobblestones and 75% of the "tinker & ponder" time spent this year by the active S@H community.

The design goals for BOINC are not for the "tinker and ponder" set but the set-and-forget user.

1 million is high, but so is 8,800 -- it's the number who have posted just once.

There are many who are not forum-fanatics.

---

there are also many who read the forums but do not post. lurkers as we tend to call them.

there are also many who read the forums but do not post. lurkers as we tend to call them.

I am aware of the term, and the behaviour. The project doesn't know how many people read the forums, only those who log in and post.

---

i dont know what type of forum software we use, but on some, perhaps most bbs, the forum software is sophisticated enough to distinguish anon's from registered users based on ip addy's, maybe cookies? and such. so a registered user that is not logged in will not be counted as an anon when they browse before logging in to post. this type of bbs software can show you the number of unique guests, regular guests, registered users, registered users that are not logged in, etc.

i dont know if seti@home forum software can do this or not.

i dont know what type of forum software we use, but on some, perhaps most bbs, the forum software is sophisticated enough to distinguish anon's from registered users based on ip addy's, maybe cookies? and such. so a registered user that is not logged in will not be counted as an anon when they browse before logging in to post. this type of bbs software can show you the number of unique guests, regular guests, registered users, registered users that are not logged in, etc.

i dont know if seti@home forum software can do this or not.

The forum software is unique to (and integrated with) BOINC.

I asked for the statistics, and Eric said "we don't know how many read without posting."

Could they figure something out through the Apache logs? Maybe.

---