App 4.78

Message boards : Number crunching : App 4.78

To post messages, you must log in.

AuthorMessage
mrwizer
Avatar

Send message
Joined: 18 Sep 05
Posts: 23
Credit: 507,085
RAC: 0
Message 1673 - Posted: 24 Oct 2005, 4:19:46 UTC

David (or anyone else that knows), what exactly was changed in 4.78? I know the memory foot print was reduced, and OS X 10.3.9 is now supported. But has anything been added for WU hangs or errors?

Not a push, just trying to get a grasp on what was modified. And if there is a change log, etc., just point me in the right direction. Thanks.
ID: 1673 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Herk

Send message
Joined: 28 Sep 05
Posts: 1
Credit: 3,606
RAC: 0
Message 1676 - Posted: 24 Oct 2005, 5:49:43 UTC

What is memory footprint?
ID: 1676 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
mrwizer
Avatar

Send message
Joined: 18 Sep 05
Posts: 23
Credit: 507,085
RAC: 0
Message 1677 - Posted: 24 Oct 2005, 5:52:16 UTC

Memory usage of the Rosetta application. although realted to the WU type, it can also be affected by the application. For instance, 4.75 had a much larger memory usage.
ID: 1677 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Paul D. Buck

Send message
Joined: 17 Sep 05
Posts: 815
Credit: 1,812,737
RAC: 0
Message 1681 - Posted: 24 Oct 2005, 8:46:19 UTC - in response to Message 1676.  

What is memory footprint?

The amount of space taken up by the program and data, as mrwizer said. The executable program for Rosetta@Home is 4.392 Mega-Bytes. So, when run, the absolute minimum amount of memory space used is about that much.

To this amount you have to add in the amount of memory used to hold the data to be processed, the intermediate result/data structures, and the output data structures.

Very early computers used a divided memory architecture where the program was stored in one memory and the data in a physically separate memory bank. This was later called the "Harvard" architecture.

The problem came up was that at times you may have a very small program with a very large data set. Or just the opposite... with the memory in separate physically isolated banks it meant that one or the other was always underutilized.

A different memory architecure was proposed by John von Neuman where programs and data would reside in one memory space and the computer would then have less "wastage" ... of course, we now had a new problem, called the "von Neuman Bottleneck" where the memory to CPU bus speed became a new limiting factor ... :)

Now, we use a modified scheme where we use integrated memory systems outside the CPU and level 1 cache (level 2 and 3 cache are almost always still integrated) with level 1 cache being constructed on the principles of the "Harvard" architecture with separate locations for the progam and data.

Some of this is discussed in the Wiki, and you can also look up some of these other terms there and in the general Wikipedia:

Physical memory
Virtual Memory
Cache

ID: 1681 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote

Message boards : Number crunching : App 4.78



©2024 University of Washington
https://www.bakerlab.org