|
1)
Message boards :
Number crunching :
gnfs-lasieve processes and cpu time
(Message 1217)
Posted 4 Jun 2018 by C. Ketelsen Post: Another example: https://yafu.myfirewall.org/yafu/result.php?resultid=2448537 runtime: 100,336.00 sec; cpu time: 546,383.60 running on 40 cores the overall cpu time should be around 30 to 35 times the runtime, which is clearly not the case. people contribute to this project for free, at least try to credit their contribution correctly |
|
2)
Message boards :
Number crunching :
gnfs-lasieve processes and cpu time
(Message 1216)
Posted 31 May 2018 by C. Ketelsen Post: oh and what the heck is this? These tasks ran on a 40-core system: https://yafu.myfirewall.org/yafu/result.php?resultid=2424269 https://yafu.myfirewall.org/yafu/result.php?resultid=2428966 First Task: CPU time roundabout 3 times the runtime Second Task: CPU time roundabout 9 times the runtime But the best: second task cpu time is over 9.5 times as long as first, but the first task gets over 22 times the credit, like WTF? |
|
3)
Message boards :
Number crunching :
gnfs-lasieve processes and cpu time
(Message 1215)
Posted 30 May 2018 by C. Ketelsen Post: maybe someone can clarify on this. as far as i can see, the gnfs-lasieve* processes are not counted towards total cpu-time although they consume alot of cpu time? when i observe a workunit in boinctui current_cpu_time does not increase when the sieve processes run. |