<html><body><p><font size="2">---------------------------</font><br><font size="2">Wed, Aug 28, 2019:</font><br><font size="2">---------------------------</font><br><br><font size="2">- Bump normalization (to make loop induction variables start from 0 and</font><br><font size="2"> increment by 1)</font><br><font size="2"> - History behind why it was removed: some targets were not able to </font><br><font size="2"> reverse that transformation, and the transformation pessimized the</font><br><font size="2"> input (eg. normalizing a loop that iterates over a pointer inserts</font><br><font size="2"> a new variable for IV but the original pointer is still needed). </font><br><font size="2"> Loop Strength Reduce should be able to clean it up it in theory.</font><br><br><font size="2">- Delinearization/Multi-dimensional array indexing RFC </font><br><font size="2"> (Siddharth)</font><br><font size="2"> - A nice summary of the points raised so far prior to today's meeting </font><br><font size="2"> are available at: </font><a href="https://github.com/bollu/llvm-multidim-array-indexing-proposal/blob/master/discussion-first-round.md"><font size="2">https://github.com/bollu/llvm-multidim-array-indexing-proposal/blob/master/discussion-first-round.md</font></a><br><font size="2"> - Another possible idea was raised as a result of discussion in the </font><br><font size="2"> meeting to extend the GEP instruction with operand bundles.</font><br><font size="2"> - The bundle on GEP would be optional (like it is on calls) but it </font><br><font size="2"> would also be "droppable" meaning that an optimization who doesn't </font><br><font size="2"> know about it can drop it without changing the observable </font><br><font size="2"> behaviour of the program. Siddharth to investigate the idea and </font><br><font size="2"> share his findings at the next meeting.</font><br><br><br><font size="2">----------------------------</font><br><font size="2">Tentative Agenda for Sept 11</font><br><font size="2">----------------------------</font><br><br><font size="2">- LICM vs Loop Sink Strategy (Whitney)</font><br><br><font size="2">- Status Updates</font><br><font size="2"> - Follow up on multi-dimensional array indexing RFC (Siddharth)</font><br><font size="2"> - Data Dependence Graph (</font><a href="https://reviews.llvm.org/D65350"><font size="2">https://reviews.llvm.org/D65350</font></a><font size="2">) (Bardia)</font><br><font size="2"> - Impact of Loop Rotation on existing passes (Min-Yih)</font><br><font size="2"> - Bugzilla bugs update (Vivek)</font><br><font size="2"> - Others?</font><br><br><br><font size="2">Bardia Mahjour<br>Compiler Optimizations<br>IBM Toronto Software Lab<br>bmahjour@ca.ibm.com (905) 413-2336</font><BR>
</body></html>