Topics

Added: Computer tab MAC address column, this allows dynamic IP addresses to be used.  Use: use_mac for the IP Address and fill in the MAC address int the next column

too geeky and who would remember the magic word?
What about dropping this "use_mac" completely and use the MAC value, if the IP is empty?

Add: Messages: 3 user definable texts, that show a highlighted color.  (V 1.43 )

Works nicely. Sort of "*core of the text to be highlighted*".
Could the text be searched in a case-insensitive way (if it is not intentionally case-sensitive)?

I've noticed that my "BT Data Transfer graph" displays the same fixed amount for each day's upload (1.8 MB) or download (2.6 MB) in the tooltips, although the graphs wildly vary up to 240 MB. Timeline spans one year.

Threads graph - if both actual values and average is displayed, the average line is much thinner and often not visible behind the (much more dense) values. Could the average be thick and values thin?

Tasks graph: I've sometimes noticed, that I can select a few projects and see their tasks graphs, but as soon as I select any of a few other projects, all graphs disappear. No matter whether Combine projects is selected. (This is when the "Over a period of:" (= Timespan) checkbox is deselected, i.e. "All times" data.)
If I select the Timespan and play with the days, I can find a borderline - with less days, all is fine, with more days, all graphs suddenly disappear.

With other words: I can select even a 190 day Timespan. After selecting a few (or all) projects (from a group of 11 "good" projects), their tasks are displayed in graphs, for 7.10.-6.12.2011. Selecting any of the remaining 6 projects (e.g. SETI@home) causes all graphs to disappear (and the graph baseline gets 6.10.-6.12.2011). But after lowering the Timespan to 60 days, the graphs get displayed again (and the graph baseline gets 7.10.-6.12.2011). Then I select one more project (form the "bad" ones - Rosetta), graphs disappear (and the graph baseline gets 6.10.-6.12.2011). After lowering the Timespan to 59 days, the graphs get displayed again (and the graph baseline gets 7.10.-6.12.2011). Now after selecting any (or all) from the remaining "bad" ones, all graphs remain correctly displayed.

(I remember there was already a similar bug in the far past, with Credit Graphs disappearing this way.)

Project rank column - what is the second number? What about right-aligning the ranks?

Find computers - the range is assumed according to the mask? What about environments with huge masks and trying to scan just a small known range (like 50 IP's)?

"-Add: Settings->Extra: Thread/process run time graph."

Just a typo in announcement: should be "Extra->Settings->Expert".

I assume it requires a functional TThrottle connection to the target machine(s)? If yes, which version? What does it present on the machines? The localhost is a flat line at 0% (with occasional tiny spikes).

_BT is around 1% with occasional spikes in 0.5-2.5% area. A minimized BG gets _BT mostly to 0%, a busy GUI forces it to 7%.

Switching between tabs issue:
I've stopped BOINC client (shortly past 19:29:00) to install a newer version. In a few seconds BT dismissed all tasks from the Tasks tab pane. So far all good. But suddenly I could not switch to the Computers tab by clicking on it, neither on any other tab - the "tab button" just flashed after a click and nothing happened. But I was able to switch between tab panes using the "Show / ...." menu items, without any problem.

In a minute or so the newer BOINC version was installed and started (at 19:32:56). Some other minute later I helped the BT's connection to a client to be recovered - suddenly the usual switching between the tab panes was functional as usual.

Code Select Expand

16 november 2011 - 22:29:57 BoincTasks version: 1.27
16 november 2011 - 22:29:57 Language ---- User: 1051 (SKY), System: 1051 (SKY), Selected: 0 () Used: 1051 (SKY)
16 november 2011 - 22:30:01 Maximum number of computers: Unlimited
16 november 2011 - 22:30:01 Computers ---- Vandus , localhost
[....]
17 november 2011 - 19:15:06 Connect ---- Connected to: Host: localhost, Vandus,Port: 31417, TThrottle Version: 5.44
17 november 2011 - 19:29:47 Connect ---- Lost connection to: Host: localhost, Vandus, Port: 31416
17 november 2011 - 19:34:12 BoincTasks is closing down all threads
17 november 2011 - 19:36:14 BoincTasks has closed down all threads
17 november 2011 - 19:36:16 Maximum number of computers: Unlimited
17 november 2011 - 19:36:16 Computers ---- Vandus , localhost
17 november 2011 - 19:36:17 Connect ---- Connected to: Host: localhost, Vandus, Port: 31416, BOINC Version: 6.13.12
17 november 2011 - 19:36:20 Connect ---- Connected to: Host: localhost, Vandus,Port: 31417, TThrottle Version: 5.44

TTh's (and also BT's) large Temperature Graph now contains, in addition to "hours ago" (or "minutes ago") markers at thick green vertical lines, also a true wall-clock timeline, which is exactly explanatory regarding the past time. But if the computer gets asleep for some time, after it awakes, it is not true anymore.

I've been thinking about whether (at all) to try to take the interruptions into account (this is actually up to the author, as usual), and how to display them graphically. The simple numbers "hours/minutes ago" could be kept interpreted as "run-rime ago", and a double thick green vertical line would simply mark the place of interruption. But this idea does not play with true wall-clock time. Is it some brainstorming worth?

The changeset for BT 1.25 seems to ne promising going to be a wonderful release!
Hope this thread will be as short as possible

Currently if tasks using some resource (CPU, GPU) are being throttled - be it because of temperature or a lowered max. limit - all tasks from the particular client, which use this resource, are displayed as being throttled by some non-zero amount (1-99%). But some processes might be (and often are) excluded - wrappers, nCi's, explicitly excluded projects and/or executables.

When TTh informs BT about tasks being throttled for resource XY, it sends the throttling % value. Could it also send the PIDs of throttled processes? BT could then be capable to fill the entries in the Throttle column just for tasks, which are really being throttled.
(A task may consist of multiple processes; if any of them is being throttled, then the task can be considered as being throttled (by the largest amount of any throttled process).)

The addition of dynamic target temperature reminded me on my older wish and e-mail discussion with Fred - let's try to revive it. A bit longer, but still possible to follow.

I'm adding a wish: the red line in graph symbolizes the target temperature, I can see whether the real temperature(s) do match it. At a later point (overnight, hot day) I may wish to change the target temperature, but it will be moved in the graph as one line. The historical value gets lost and the temperatures from past suddenly do not match.

Could the target temperature values be also stored and displayed the same way as the measured temperatures are, instead of moving one line over the whole graph?

You have to convince me... The temperatures do not change when you set another max. The temperature difference between what TThrottle reads and a case temperature reading may change.

OK, Fred, then you have possibly misunderstood. It was actually not about temperature values, but just the straight target temp line.

My computer was running overnight with target temperature set to 79ºC (the straight red line):



and TTh was able to keep both CPUs (orange+yellow graphs) at around the same temperature.

After logging me in, I'd like to lower the fan speed (it actually rotates much faster and disturbs at temperatures over 81ºC, but just imagine I'm talking about 85ºC overnight and 69ºC while I'm working).

Therefore I've set the target temperature to 75ºC (the straight red line moved down a bit):



After a while the temperature stabilized, both CPU cores match the target temperature. But it seems that (after a hour or two, I've forgot about, or...) some time ago the temperatures were not matching the target. WTF? 

This could be solved if the target temperature (my straight red line) would be also "dotted" according to the actual value at these previous points in time. (So it would not appear as one straight line, but possibly a set of lines, as the user moves the target over time.)



You can consider it being just a cosmetic issue. Thus the "wish".

Fred actually never responded, IIRC 

Now TTh is capable of moving the target temperature even automatically, according to its settings and user's (in)activity. It would be nice if the target temperatures' lines (yes, on an advanced computer, we have much more targets now) would signalize it at a later time.

After starting BT 1.24 for the first time, opening Setup, I've found the History length set at 120 (or even 60?) days - in 1.22 it was at 365 days!

The Graphs toolbar does really contain just graphs. But I believe the LT-History, with its window having very similar layout,  would fit there too

One concern with "Display active tasks" - if it is set and there are no more active tasks, the Tasks tab is empty. It is then not possible to open the tasks' context menu and display the inactive ones - context menu is the only way the control it...

[Question]

As of BT 1.22 (for those who enable it), the majority of finished tasks are being shifted into Long-term history, thus allowing the Short-term History tab & data to be much more lightweight, with less overhead. Would it be possible to relax the module and let it react faster or more often, to capture finishing tasks? At the moment, even if BT is open, with the Tasks tab active and refreshing each 3 seconds, I can observe some tiny tasks finishing, then uploading, being reported and vanishing in a rapid succession, but the History nearly always lists these task as "Missed". That's at 3-20 seconds History refresh, Smart mode being on. I've now tried to lower the upper bound to 5 seconds and will try to observe the History tab next time such tiny task will finish.

But my (maybe already asked, in the past, or just not answered or not understood by myself) Question is, that regardless of any timing issues and smart mode: If BT already clearly notices (and tells me through the Tasks tab), that a task is already being uploaded, why can't it trigger an event in the History code area, that this task has already transitioned from Running into Uploading state? When I see a task waiting to be reported (in the Tasks tab), how can be that History meand that BT has missed the moment?? Why can then BT tell me (in the History tab), that it has completely missed this task's Uploading+ReadyToReport states, when it already displayed the task in these states? Is it necessary to capture the tasks' states in two separate loops? One module might notify another module about a task's state, instead of letting it temporaruily snoop much faster... (And miss anyway.)

Add: If a user switch occurs, the BOINC client will be stopped and restarted when the other user logs in.

Do you mean "suspend computation", or really, literally "stop client"   

Changed: Projects: Tasks a week now shows 2 digits.

Maybe unnecessarily too much, 1 digit would be enough? But still, thanks.

Fixed: The icon isn't blinking while snoozing.

Have you finally found the culprit??


A re-test from 1.20:

• When (repeatedly) trying to trigger displaying the <ClientNotStartedBT> dialog ("The BOINC client wasn't started by BoincTasks and is still running"), I've (always) succeeded to terminate my client (launched by BOINC Manager) without any warning. Should this have happened?

Still happens in 1.21, is it intentional?

Checked again. After stopping the local client, I do see a STOP sign to the left of the computer, on the comp select sidebar. However, an automatic reconnect does not happen, even many minutes after starting the client again.

Please activate connection logging: C:\Program Files\eFMer\BoincTasks\examples\log
<log> <options> <connecting> 1 </connecting> </options> </log>
And see what is logged.

I've checked it just once, BT shut down the client, with messages:

17:12:36 Shut down BoincTasks ---- The BOINC client is shutting down
17:13:01 Shut down BoincTasks ---- The BOINC client has shut down

then some 3:40 minutes later (the timeout is set to 120 seconds) it successfully tried and connected to a newly started client:

17:16:41 Connect ---- Lost connection to: Host: localhost, Pavilon6, Port: 31416
17:16:41 Connect ---- Try to connect to: Host: localhost, Pavilon6, Port: 31416, Thread ID: 9484
17:16:41 Connect ---- Connected to: Host: localhost, Pavilon6, Port: 31416, BOINC Version: 6.12.34

Other two machines are regularly attempted for a connection in other 12-minute cycle.

I'll try to check it again later. But it was the first time it jumped in immediately 

Very good step to place Graphs' Computers and Projects selection side-by-side instead vertically! Saves a lot of (opposite) scrolling among projects. (Tasks Graphs might follow it.)
However, I'm still missing the meaning of "Expanded selection box".

I personally like the new "tube-skinned" stripes more than the former alternating ones - they were changing the underlying colors too much, differently for similar tasks. Now equally highlighted tasks also share the same color. (However they are possibly a bit too dark in the middle of rows.)

Tasks a day/week: the majority of my projects show 26/day + 235/33/week, others have 0 + 0/0. Mostly projects with no tasks on board show 0 + 0/0 and those with any tasks show 26 + 235/33, but it is not an absolute pattern, exceptionally CPDN beta with 1/0 tasks has 0+0/0.

GPU throttle graph in BT is again correct (BT 1.17 + TTh 4.30). (Maybe a longer time already, I've been using 4.20 until today.)

The sorting wasn't done on group.

Sorting should be fixed, now on group (from server, client, BoincTasks, TThrottle), project, created time.
You can collapse read messages, or delete them to clear up the long list.

Even with 1.08, the notices are still not thoroughly sorted according to any apparent key:

--------------------------------------------------------------------------------
  05.06.11 17:50  DNETC@HOME: Summer heat up - challenge   viac...     
--------------------------------------------------------------------------------
  03.06.11 20:12  DNETC@HOME: Project maintenance   viac...     
--------------------------------------------------------------------------------
  01.06.11 14:16  PrimeGrid: Mega Prime found for The Riesel Problem (Déjà vu Déjà vu)   viac...     
--------------------------------------------------------------------------------
  29.05.11 15:18  PrimeGrid: Mega Prime found for The Riesel Problem   viac...     
--------------------------------------------------------------------------------
  28.05.11 15:30  SETI@home: storage service is back up   viac...     
--------------------------------------------------------------------------------
  27.05.11 20:48  SETI@home: More problems with workunit storage server...   viac...     
--------------------------------------------------------------------------------
  21.05.11 00:49  AQUA@home: A new test application   viac...     
--------------------------------------------------------------------------------
  20.05.11 21:45  AQUA@home: Please consider upgrading your BOINC client   viac...     
--------------------------------------------------------------------------------
  25.05.11 18:50  SETI@home: Change of management at Arecibo.   viac...     
--------------------------------------------------------------------------------
  18.05.11 18:12  World Community Grid: Nutritious Rice for the World project publishes paper   viac...     
--------------------------------------------------------------------------------
  13.06.11 04:03  PrimeGrid: World Record Factorial Prime   viac...     
--------------------------------------------------------------------------------
  15.06.11 17:48  AQUA@home: Oznam od servera     
--------------------------------------------------------------------------------
  16.06.11 02:30  Milkyway@home: Oznam od servera     
--------------------------------------------------------------------------------
  17.06.11 13:50  ibercivis: Oznam od servera     
--------------------------------------------------------------------------------
  15.06.11 17:47  Oznam od BOINC   viac...   
--------------------------------------------------------------------------------
  15.06.11 17:47  Oznam od BOINC   viac...   
--------------------------------------------------------------------------------
  15.06.11 17:47  Oznam od BOINC   viac...   
--------------------------------------------------------------------------------
  16.06.11 03:01  SETI@home: New posts in subscribed thread   viac...     
--------------------------------------------------------------------------------
  14.06.11 15:14  SETI@home: New posts in subscribed thread   viac...     
--------------------------------------------------------------------------------
  14.06.11 22:10  World Community Grid: Computing for Clean Water project publishes paper   viac...     
--------------------------------------------------------------------------------

A few minutes ago (around 16:40) I've noticed a cryptic "20100-state" in History. [...] a few seconds later the line completely disappeared!

Seen again in 1.07 (however I'm not exactly sure with the "20100" - I've copied it in clipboard and later overwrote the clipboard contents with something else (I'm occasionally missing a clipboard history)

No torrent for 1.05 this time? Apparently not very useful because of small file size?

On one machine (Win 7, Avast AV free/Home), installation went fine.
On other machine 6 hours later (Win 7, Avast AV Pro) the AV complained about it being suspicious, that the BT_1.05.exe tries to start BT_1.05.tmp and whether I really want to allow it... Twice, for both BT_1.05.tmp files. But no complaints regarding any ill contents of the file.

Upon the first start on the second computer, BT 1.05 said I've 19 Notices, after switching to the tab it was empty and keeps empty. I suspect some inconsistency between my own notices templates and the update from BT 1.05, will investigate and report. On the first mentioned machine (without customized notices' templates), the Notices were displayed correctly, including their "+" and "x" marks.

I'd have a few questions to new TTh 4.10:

• Is the driver started when the OS boots, or with the first TTh startup? If the former, how far is it functional alone? Under which process is the driver executed, System(4)?
• Does TTh work just as the driver's GUI, or does it do some more work (obtaining list of BOINC+other processes - just the TTh process (well, except BT, BM and WUProp) seems to have an open TCP connection to the client, etc.)? What amount of the former TTh's work is done by the driver?
• The driver seems to be throttling even while TTh process is terminated. Does it just keep the same set of processes and temp levels?
• In the very past, the complete processes (i.e. all their threads at once) were being suspended, which could be visually observed in Process Explorer as grayed-out tasks' processes. Later due to the randomization, usually (or mostly/always?) at least some threads were not suspended and kept ready/running, and the process did not appear grayed-out. Now with 4.10 the processes again appear grayed, which suggests that their threads are being suspended in unisono. Is the randomization not in use?
• Starting TTh as a plain user UA or an admin user UB seem to deliver different preferences (throttling temperatures and levels, T_Junction, not sure about Rules and other settings) and immediately modify the throttling behavior. Shouldn't the settings be stored by the driver and TTh be just accessing and modifying them like a GUI?
• I'm working on my Win7 machines like a plain user and using UAC to do any administrative tasks. But if my TTh runs as a plain user, it refuses to do anything except Graphs and About... neither exit the GUI, even if I'm able to authenticate myself as an admin. Could TTh be able to request authentication to open its window or exit itself? Or at least allow it to exit without the need to be killed (if it is not required for communication - which unfortunately seems to be the case)?
• My system (well, TCPView) lists around 40 TCP connection to port 31417 in TIME_WAIT state, new ones are appearing each second and the oldest ones disappear with the same frequency, apparently all initiated from BT. Could BT keep one open connection instead?
• Do I correctly remember some message thread about someone connecting from Linux machine to TTh? (I'm not able to find it.) Could WUProp task's process be able to periodically connect to TTh to obtain and report CPU + GPU temperatures + throttling levels?

Two issues found, one major:

• Add: A rule that triggers when a computer or all computers loose connection.

  A "Connection =<>" rule was probably not intended - just the equal sign?
  
  
• On one computer, my local "history_ComputerName.cvs" files got emptied! All three files (just the running tasks are there), I'm not sure when exactly, but evidently after the upgrade. History files for a not connected remote machine are partially untouched - all contain the historical data.
  Today I had to cold-reset the machine, afterwards with BT 1.03 I could see the most recently logged results from two days ago at the morning - the complete history was there. Later the day (after 15:22 - the download time) I've replaced BT 1.03 with 1.04. When I've checked the History tab, it contained just entries for a remote host.
  After using a 3 weeks old copy of history file, it was accepted, but the initial lines suddenly contain one overlong string in form

Code Select Expand

1.04
TheMachineNameTheMachineNameTheMachineName............TheMachineNameheMachineNameTheMachineName
BoincTasks History
Project Application Version .......

where the "machine name" line is 0x07DB2 bytes long!!
The remote machine's history files - just the "current" got modified this way, the  "machine name" line is "merely" 0x1A0 bytes long! The two older files (cvs1+cvs2) are still dated prior to 15:22 - the installation moment, the current file gets regularly refreshed.[/li]

Additionally, it is possibly already some longer time ago (more than 3 weeks - my backup copy), but my localhost's history files begin on 22.10.2010, although my history settings were always 365+23. The remote computer's log starts correctly on 29.5.2010.

Regarding Tasks refresh times - on my machine it looks like Fast=2s, Normal=3s, Slow=7s - maybe the Normal time is too tight on the Fast time?

Sometimes when I work with true memory hog applications (0.4-2.8 GB) and try to (literally) survive on the same computer ) when there are more of them, I'd appreciate being able to see the tasks' memory requirements as separate column(s) (at least "Virtual memory", possibly also "RAM" - the same values as in the Properties window) directly in the Tasks tab.

I'm also investigating, whether it is possible to find out task's initial estimate on memory requirements, prior to starting it for the first time.