Basic HTML version of Foils prepared
August 7 98
Foil 4 Data Mapping, cont.
From
DoD HPF Training -- 4. HPF Data Mapping Facilities DoD Training and Others --
1995-98
.
by
Chuck Koelbel -- Rice University
1
Goals:
Avoiding Contention: Data updated in parallel should be on different processors
Locality of Reference: Data used together should be on the same processor
2
These goals are sometimes in conflict
To avoid all contention, put every data item on its own processor
To maximize locality, put all data on one processor
HPF gives you the tools to resolve these conflicts, but doesnąt solve them for you
in Table To:
©
Northeast Parallel Architectures Center, Syracuse University, npac@npac.syr.edu
If you have any comments about this server, send e-mail to
webmaster@npac.syr.edu
.
Page produced by
wwwfoil
on Sun Aug 9 1998