Browse files

Automatic commit at Tue Mar 20 07:00:04 MDT 2012

  • Loading branch information...
1 parent 7b41478 commit 08a73f50942aa194edc8fe6bf0743b1c2e84f013 @jeapostrophe jeapostrophe committed Mar 20, 2012
Showing with 6 additions and 0 deletions.
  1. +6 −0
@@ -187,3 +187,9 @@
order invariance. If you also ensured that reads only happened once,
then you could effectively collect all reads and then evaluate them
simultaneously on the GPU for efficiency.
+ You make 'write' a blocking operation until the next round.
+ Since you can hold off evaluating the effect of all the writes until
+ every process is at that point, you could use a mutable
+ data-structure for the tuple-space.

0 comments on commit 08a73f5

Please sign in to comment.