Message boards :
Number crunching :
Windows vs. Linux application
Message board moderation
Author | Message |
---|---|
Send message Joined: 15 Sep 07 Posts: 15 Credit: 9,818,265 RAC: 0 |
Hallo, on our team board we discussed about crunching time of Convolution Work Units. We compared two machines and results are interesting. 'bosh - AMD X2 4200 Windows XP WU-------h--m--s gs_72... 00:19:39 gs_70... 00:19:41 gs_62... 00:05:26 gs_60... 00:05:25 LiborA - AMD A64 2800+ OpenSuse 10.3 64b WU--------h--m--s gs_74 ... 00:13:39 gs_73 ... 00:13:31 gs_72 ... 00:13:43 gs_72 ... 00:13:34 gs_71 ... 00:13:29 gs_62 ... 00:05:26 gs_60 ... 00:05:24 It seems that for non-convolution Work Units the application need a similar time for the WU in the same range. But for Convolution Work Units is linux application a more, more effective. |
Send message Joined: 30 Aug 07 Posts: 2046 Credit: 26,480 RAC: 0 |
Hallo, on our team board we discussed about crunching time of Convolution Work Units. We compared two machines and results are interesting. It looks like the windows application isn't being compiled as well as the linux one. The problem behind this might be that the linux is using a 64 bit binary (where i'm pretty sure the windows does not). We're working on getting a 64-bit windows system to compile on which will hopefully improve the times on windows. |
Send message Joined: 15 Sep 07 Posts: 15 Credit: 9,818,265 RAC: 0 |
The problem behind this might be that the linux is using a 64 bit binary (where i'm pretty sure the windows does not). Are you sure? I have downloaded astronomy_1.13_i686-pc-linux-gnu executable, it seems as 32b application? And on download folder isn't x86_64 application available. We're working on getting a 64-bit windows system to compile on which will hopefully improve the times on windows. Its a very good message for all users of 64b Windows OS :-) |
©2024 Astroinformatics Group