Error: In graphics screen no percentage or time figures.

Observation: After first 25% it looks like time to completion could be as much as 20% longer than V 5.27.

Andy

The identification in stderr.txt still shows
setiathome_enhanced 5.27 DevC++/MinGW

With the focus on a new graphics module, it would be nice if this project had the same range of preferences for graphics as the main project does. Choosing one of two color schemes doesn't seem like a very thorough test.

Result 2856748 completed in 17754.63sec.
That is about 15% slower that using V5.27, ave 15,400sec approx, at very similar AR.

Andy

Error: In graphics screen no percentage or time figures...

Same with the %s & time.

Second result completed, 2860124 This time on other computer. Didn't look at graphics, will do on next unit, plus some stop/starts and shutdown/restart of BOINC.
Again time was longer, 9,855sec, compared to ave of 8,900sec with V 5.27. Thats about 10% slower.

Andy

First result rid 2861088 has finished. Runtime was longer for me, too.
Approximately 13400 secs (AR=0.405) with 5.27, now 15500 secs (AR=0.405) with v6.00 .
____________
_\|/_
Urs

Same experience as other reporters. Seven tasks have completed, one validated so far. Seem to be 10%/15% slower than 5.27, but I wouldn't take exact timings from this machine - it depends too much what's running on the other cores.

I was able to display the graphics, but it wasn't a good experience - running BOINC v5.10.13 under Vista Business, service install, system logon with desktop interaction allowed.

Requesting display of graphics resulted in an "Interactive services dialog detection" prompt from Vista. When I accepted the invitation to 'Show me da message', I got a new blank desktop with - for the hell of it - seven different SETI mini-windows showing different graphics, unsynchronised, for the seven different Beta tasks that happened to be running at the time. (I can't remember what core 8 was doing at the time - CPDN, probably). I could see fields for %age done and elapsed time, but both were displaying as zero.

When I do this normally, I get a little prompt dialog which monitors what's going on on the graphics desktop, and allows me to return to the normal display when I'm finished. Somehow during testing I got into a state where there were no graphics windows on the graphics desktop, but no return prompt either: I had to use ctrl-alt-del to get back to the normal screen. I suspect that this may be a problem related to a task finishing while displaying graphics - I'll try and check the next one more carefully.

Does anyone know whether there's a BOINC v6 client that's safe to test yet? If so, it would obviously be a good idea to test this SETI v6 in its design environment under Vista, and I don't mind giving it a try.

Edit - @ Joe S - watch out for mail - RID 2858696

Does anyone know whether there's a BOINC v6 client that's safe to test yet? If so, it would obviously be a good idea to test this SETI v6 in its design environment under Vista, and I don't mind giving it a try.

The application is very finicky as to where specific routines are located, and that is what appears to be causing the 10-15% variation between compiles.

I'm going to look into GCC compile time flags that might help stabilize it. I'd appreciate advice from any experts out there...

Eric
____________

My current Beta WU is still crunching and when I look at the graphics it shows no % done or CPU time.

I run a AMD 3000+ witn Windows XP.
____________

WU’s crunching well verification on some this one below is -9 overflow 31 pulses I have the log file if you require it Joe but co host has not return his/her result yet to confirm.
802858

____________
A computer program will always do what you tell it to do, but rarely what you want it to do.

WU’s crunching well verification on some this one below is -9 overflow 31 pulses I have the log file if you require it Joe but co host has not return his/her result yet to confirm.
802858

All of my 6.0 got Redundant result. All of them was completed, but i got Zero Credit.

http://setiweb.ssl.berkeley.edu/beta/result.php?resultid=2859543
http://setiweb.ssl.berkeley.edu/beta/result.php?resultid=2859434
http://setiweb.ssl.berkeley.edu/beta/result.php?resultid=2857223


What means redundant result?

And i want my credits!!!!!!!!
____________
Save the Gouldian Finch

All of my 6.0 got Redundant result. All of them was completed, but i got Zero Credit.

http://setiweb.ssl.berkeley.edu/beta/result.php?resultid=2859543
http://setiweb.ssl.berkeley.edu/beta/result.php?resultid=2859434
http://setiweb.ssl.berkeley.edu/beta/result.php?resultid=2857223


What means redundant result?

And i want my credits!!!!!!!!

The application is very finicky as to where specific routines are located, and that is what appears to be causing the 10-15% variation between compiles.

I'm going to look into GCC compile time flags that might help stabilize it. I'd appreciate advice from any experts out there...

Returned first V6 took about an hour longer then 5.27, graphics

the same as reported by others time & % not updating.

pete.
____________

I can not say that run times are all that much greater. However when graphics are started they take a great amount of cpu cycles. For example on my PII, 400 Mhz, Win 98SE laptop this is the data collected by process explorer:

98% V6 running from command prompt

95% v6 running from Bonic Manager

60-30% v6 running from Bonic Manager and graphics running, on screen

50-25% V6 running from Bonic Manager and graphics running, Minimized

On this slow box the graphics application uses as much CPU as the science application, and uses even more resources when minamized. I have seen this behavior on my faster boxes and even on 3.2Ghz P4 graphics make a big hit. Perhaps those reporting longer run-times are really seeing the cpu cycles lost to the graphics app.

I can not say that run times are all that much greater. However when graphics are started they take a great amount of cpu cycles. For example on my PII, 400 Mhz, Win 98SE laptop this is the data collected by process explorer:

98% V6 running from command prompt

95% v6 running from Bonic Manager

60-30% v6 running from Bonic Manager and graphics running, on screen

50-25% V6 running from Bonic Manager and graphics running, Minimized

On this slow box the graphics application uses as much CPU as the science application, and uses even more resources when minamized. I have seen this behavior on my faster boxes and even on 3.2Ghz P4 graphics make a big hit. Perhaps those reporting longer run-times are really seeing the cpu cycles lost to the graphics app.

Difficult to get exact figures - on my 8-core, an application running flat out shows as 12%/13% - but while I was testing the graphics, the separate thread seemed to average about 4%, with a lot of variability. So SETI with graphics running would be very roughly two-thirds science, one-third eye candy.

Like Andy, I only ran the graphics briefly for testing - most of the time, it runs as a service in the background and does pure science. No change in the experimental conditions, so the slowdown between 5.27 and 6.00 seems to be down to the science app only.