Move loading of atomic vs. cstdatomic in atomic_compat.h.

[loba.git] / TODO

diff --git a/TODO b/TODO
index c68c38dd43db977fdbb4516b09ad81f8fbb880fd..77dbb42286c0cd98c42d99fdbbe411529fbbaf54 100644 (file)
--- a/TODO
+++ b/TODO
@@ -1,11 +1,25 @@
-* implement loba_* algorithms (start with some trivial one)
-* add loba algorithm selection (-a number ?)
+* Change simulator to use discrete tasks.
+   Each task must then be iterated a minimum of times.
+   At least two strategies to balance the tasks:
+    (i) send the tasks that have the most count of iterations;
+   (ii) send the tasks that have the least count of iterations.

 
 

-* fix process::run when load is 0
-   -> wait for a message...
-   -> how does it work with opt::bookkeeping ???
+* Support heterogeneous platforms?
+   Not sure yet.
+   Should be doable if each process also sends its speed to its neighbors.

 
 

-* implement automatic process topology 
-   (line, ring, star, btree, clique, hypercube, etc..)
-* implement automatic platform generation
-   (number of hosts, all connected, constant bandwidth/latency)
+* Add options -j/-J: minimum number of iterations?
+
+* Add options -f/-F: termination based on flops achieved?
+
+* Choose a license before making the code public.
+   CeCILL-B?
+
+* Translate french parts in english (README, QUESTIONS, ALGORITHMS, ...)
+   Not urgent.
+
+* Add synchronized mode?
+   No.
+
+* Add a variant to (not) change neighbor load information at send?
+   No.