MT and single-CPU point differences

Author	Message
MellowedWithAge Send message Joined: 23 Mar 13 Posts: 2 Credit: 6,031 RAC: 0	Message 59062 - Posted: 23 Jun 2013, 16:01:33 UTC When I run an MT N-Body Simulation WU on 2 or more cores I average under 6 points per CPU per hour. Example WU: http://milkyway.cs.rpi.edu/milkyway/workunit.php?wuid=385527079 (#520156 is me) When I run single CPU MilkyWay@Home WU's I average over 40 points per hour. Example WU: http://milkyway.cs.rpi.edu/milkyway/workunit.php?wuid=386935266 Why the huge difference?? ID: 59062 · Rating: 0 · rate: / Reply Quote

Alinator Send message Joined: 7 Jun 08 Posts: 464 Credit: 56,639,936 RAC: 0	Message 59064 - Posted: 23 Jun 2013, 16:39:52 UTC Mostly because the current nBody runs have been as much a beta test for the new app as much as anything else. Thus there have been bigger fish to fry lately than getting the credit rate firmed up. ID: 59064 · Rating: 0 · rate: / Reply Quote

jdzukley Send message Joined: 26 May 11 Posts: 32 Credit: 43,959,896 RAC: 121	Message 59157 - Posted: 28 Jun 2013, 19:09:44 UTC appreciate that there are bigger fish to fry, but also I would appreciate that someone take just a bit of time to adjust the point calcs for mt tasks... and I know I do not do this for the points (so at least I tell my self).... 12 cores for 59 minutes run time = 1.55 points... 24k, 35k and 39k CPU seconds for the 3 different computers that crunched this task... http://milkyway.cs.rpi.edu/milkyway/workunit.php?wuid=382923304 like I keep telling myself, your not in this for the points... ID: 59157 · Rating: 0 · rate: / Reply Quote

Alinator Send message Joined: 7 Jun 08 Posts: 464 Credit: 56,639,936 RAC: 0	Message 59167 - Posted: 29 Jun 2013, 17:34:31 UTC Hmmmm... Something strange is going on. I've recently started having some that went through with more or less the same amount of net CPU time, but paid less than one credit!!?? ID: 59167 · Rating: 0 · rate: / Reply Quote

xsedlacp Send message Joined: 5 Sep 12 Posts: 2 Credit: 216,778 RAC: 0	Message 59174 - Posted: 30 Jun 2013, 7:41:20 UTC The credit system is probably corrupted. MT task ran about 148 minutes on 8 cores (more than 17 hours of CPU time) and I received 0.64 credits. See wu 388738664 ID: 59174 · Rating: 0 · rate: / Reply Quote

Conan Send message Joined: 2 Jan 08 Posts: 123 Credit: 69,522,130 RAC: 2,002	Message 59873 - Posted: 10 Sep 2013, 23:11:43 UTC Last modified: 10 Sep 2013, 23:28:31 UTC Re-started running some CPU Milkyway tasks and found that the Milkyway@home N-BODY Simulation v1.36 (mt) tasks are giving a lot less points than the other Milkyway@home applications. On 6 Core Linux 64 bit computer Milkyway@home N-BODY Simulation (mt) v1.36, Run Time= 1,260.27s, CPU Time= 6,598.42s gives 70.84 points (See WU 424213680) Milkyway@home v1.01, Run Time= 3852.00s, CPU Time= 3816.61s gives 106.88 points (See WU 424348905) Milkyway@home Separation (Modified Fit) v1.26, Run Time= 14,881.84s, CPU Time= 14,763.36s gives 213.76 points (See WU 419384816) Milkyway@home Separation (Modified Fit) v1.26, Run Time= 9,136.92s, CPU Time= 9,071.21s gives 267.19 points (See WU 423885082) A 4 core computer is similar. It is not adding up to me when a Multi-Thread Application uses all available CPU resources (stopping other work from taking place) but awards less for doing so. Conan ID: 59873 · Rating: 0 · rate: / Reply Quote

pmassey Send message Joined: 22 Dec 10 Posts: 8 Credit: 19,792,404 RAC: 0	Message 59929 - Posted: 19 Sep 2013, 4:25:25 UTC So how long is it planned for them to be in beta? I do like that it is a MT/MP application, which none of the other tasks I run on BOINC are. Just wish the credits gained faster. "Apollo was astonished, Dionysus thought me mad. But they heard my story further and they wondered and were sad." ID: 59929 · Rating: 0 · rate: / Reply Quote