Posts by [AF>Le_Pommier] Jerome_C2005

1) Message boards : Number crunching : Colour strips Badges ?? (Message 40)
Posted 21 Jul 2023 by [AF>Le_Pommier] Jerome_C2005
Post:
Thanks for the answer Conan (very difficult to read though ! :) ) but if you click reply it's then very easy to read (you don't have the bloody left blank space)

I actually compared in details the tasks running in that VM in the mac, and those coming from another external (hosted) VM that I have (small and old 2 cores) and that is currently running ramanujan also

Debian VM on my iMac (not old i9 10 cores)
Average running time 	01:10:31
Average CPU time 	01:09:35
Average Credits 	14,3
Credits/hour		12,3

Debian VM on external service provider OVH (old 2 cores)
Average running time 	00:41:45
Average CPU time 	00:41:28
Average Credits 	26,3
Credits/hour		38,0


So the run time is much worse on a supposedly more powerful and recent machine, and obviously the credit/hour ratio is terrible on the iMac VM.

I just ran the benchmarks on the VM and I got

4083 FP MIPS
85904 integer MIPS

and on the project page of that machine I realize that you are most probably right :

Vitesse mesurée pour les calculs en virgule flottante 1 milliard(s) d'opérations par seconde
Vitesse mesurée pour les calculs en nombres entiers 1 milliard(s) d'opérations par seconde

so yes this is some kind of "default" ! I thought that boinc was doing this automatically ?? so hopefully it should change it all !


By curiosity I ran the benchmarks on the iMac itself (macos boinc) and I got

5665 FP MIPS
16195 integer MIPS

which is very surprisingly low for the integer test compared to the VM on that same machine !

Thanks again :)
2) Message boards : Number crunching : Colour strips Badges ?? (Message 38)
Posted 19 Jul 2023 by [AF>Le_Pommier] Jerome_C2005
Post:
I think that "s****d up" was not a bad definition after all :)



+ I get 10 credits for 1h50 of crunching per task, so I'm afraid I won't get any badge before a looooong time :)




©2024 The Ramanujan Machine Project Team