1)
Message boards :
Extended Sierpinski Problem :
IT'S ALIVE!
(Message 76996)
Posted 3278 days ago by FrankHagen
looks good so far.
|
2)
Message boards :
Extended Sierpinski Problem :
The Extended Sierpinski Problem Double Check
(Message 76967)
Posted 3279 days ago by FrankHagen
Sometime Saturday or Sunday, we expect to be starting the ESP project on the BOINC server.
well then it's about time to make it selectable in account settings.
|
3)
Message boards :
Number crunching :
Perseid Shower Challenge
(Message 68328)
Posted 3576 days ago by FrankHagen
Oops. My mistake.
109.22 is the "official" credit rate for TRP Sieve, so that's what you see on the challenge leader boards.
When it comes to actually counting credit, there's a 10% "conjecture bonus" added to the credit, which brings you to 120.14.
For sieve challenges, since all sieves are fixed credit, perhaps we should switch the challenge scoring to simply be a count of the tasks completed. "Long" tasks would count as 5, and short tasks would count as 1. The credit is irrelevant in a challenge with fixed credit.
well - or my mistake to forget about the bonus.
anyway - a point per task sounds logical in this case.
|
4)
Message boards :
Number crunching :
Perseid Shower Challenge
(Message 68326)
Posted 3576 days ago by FrankHagen
any explanation for 109.22 points per WU - i thought boinc-credits would be counted?
You may have missed where the tasks were made 5 times shorter, and therefore get 1/5th the credit. 109.22 is the BOINC credit per task, unless you happen to get one of the old, longer tasks that are still floating around the system.
nope - didn't miss it.
i see 120.14 boinc-credits for the short ones.
most likely no effect for the challenge, but weird. ;)
|
5)
Message boards :
Number crunching :
Perseid Shower Challenge
(Message 68323)
Posted 3576 days ago by FrankHagen
any explanation for 109.22 points per WU - i thought boinc-credits would be counted?
|
6)
Message boards :
Number crunching :
Valentine's Day Challenge
(Message 62607)
Posted 3756 days ago by FrankHagen
no stats, no fun!
|
7)
Message boards :
Problems and Help :
outch!
(Message 52965)
Posted 4064 days ago by FrankHagen
Did you inform the ubuntu-dev's about this?
PS: I am using the older Lubuntu11.10 with client 7.0.24 and had no problems in YoYo, PG and in the last DistrRTgen challenge.
nope- already a long list of bug reports about this..
|
8)
Message boards :
Problems and Help :
outch!
(Message 52962)
Posted 4064 days ago by FrankHagen
This thread suggests a Ubuntu-related client problem
So have you tried to use a boinc 6 series client or a non-Ubuntu provided client?
Before downgrading must sure your cache is empty because, as you are probably aware, boinc client 7 series has a different format than the 6 series!
Or does Ubuntu have updated libC libraries or something (not an Ubuntu person)?
since it also seems to affect apps which have no connection to boinc, it's probably totally worthless to try to rollback to boinc 6.x.
|
9)
Message boards :
Problems and Help :
outch!
(Message 52929)
Posted 4065 days ago by FrankHagen
i upgraded one host to ubuntu 12.04 today and now all SGS-LLR tasks are crashing right away:
<core_client_version>7.0.24</core_client_version>
<![CDATA[
<message>
process exited with code 193 (0xc1, -63)
</message>
<stderr_txt>
Stack trace (11 frames):
/usr/bin/boinc(boinc_catch_signal+0x65)[0x48f675]
/lib/x86_64-linux-gnu/libc.so.6(+0x364c0)[0x7f3afa10c4c0]
/lib/x86_64-linux-gnu/libc.so.6(__strcat_chk+0x39)[0x7f3afa1dcdb9]
/usr/bin/boinc[0x412a34]
/usr/bin/boinc[0x412b00]
/usr/bin/boinc[0x43163d]
/usr/bin/boinc[0x431adb]
/usr/bin/boinc[0x41a728]
/usr/bin/boinc[0x4712a4]
/lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xed)[0x7f3afa0f776d]
/usr/bin/boinc[0x406a65]
</stderr_txt>
any ideas?
|
10)
Message boards :
Number crunching :
The Marie Sophie Germain Challenge 2012
(Message 52559)
Posted 4073 days ago by FrankHagen
The best we can do now is to do some bigger work so the validator can get more CPU power. I run TRP LLR now to easy the load on Boinc server.
you know that' it's not lack of CPU-power, but transaction time due to a huge DB on a host which has not the I/O capacity to handle that.
|
