<html><body><p><font size="2">---------------------------</font><br><font size="2">Wed, Nov 6, 2019:</font><br><font size="2">---------------------------</font><br><br><font size="2">- LLVM dev conference update</font><br><font size="2"> - Tutorial on how to write loop passes and a panel on Loop Opt Working </font><br><font size="2"> Group went well and were well attended.</font><br><br><font size="2">- Loop Nest Analysis (Ettore Tiotto)</font><br><font size="2"> - What instructions should be allowed/disallowed in between perfect </font><br><font size="2"> loop nests?</font><br><font size="2"> - Take the usual approach of being restrictive and gradually allow </font><br><font size="2"> more instructions.</font><br><font size="2"> - Using isSafeToSpeculativelyExecute() may be a good starting point,</font><br><font size="2"> but need to consider other instructions commonly occurring in loop </font><br><font size="2"> nests such as branches and phis.</font><br><br><font size="2">- `#pragma clang transform` and `#pragma clang loop ivdep` (Michael Kruse)</font><br><font size="2"> - Should it be a hint or an order?</font><br><font size="2"> - Current proposal is to have it dictate and require the specified </font><br><font size="2"> transformations in that order.</font><br><font size="2"> - ivdep is defined in a way that makes it difficult to use because it </font><br><font size="2"> depends on the specific limitations of dependence analysis in a </font><br><font size="2"> given compiler, as such it won't be pursued in llvm.</font><br><br><font size="2">- The "freeze" instruction has landed (</font><a href="https://reviews.llvm.org/D29011"><font size="2">https://reviews.llvm.org/D29011</font></a><font size="2">).</font><br><font size="2"> Watch out for potential impact to loop transformations (Philip Reames).</font><br><br><font size="2">- Status Updates</font><br><font size="2"> - DomTreeUpdater investigation (Kit) </font><br><font size="2"> - No update</font><br><font size="2"> - Data Dependence Graph (Bardia)</font><br><font size="2"> - </font><a href="https://reviews.llvm.org/D68827"><font size="2">https://reviews.llvm.org/D68827</font></a><font size="2"> is ready for further comments </font><br><font size="2"> or approval.</font><br><font size="2"> - Follow up on multi-dimensional array indexing RFC (Siddharth) </font><br><font size="2"> - No update.</font><br><font size="2"> - Impact of Loop Rotation on existing passes (Min-Yih) </font><br><font size="2"> - No update.</font><br><font size="2"> - Bugzilla bugs update </font><br><font size="2"> - No update.</font><br><br><br><font size="2">------------------------------</font><br><font size="2">Tentative Agenda for November 20th</font><br><font size="2">------------------------------</font><br><br><font size="2">Loop Trip Count Value Profiling - Wael Yehia</font><br><br><font size="2"><Please let me know if you have other topics to discuss></font><br><br><font size="2">- Status Updates</font><br><font size="2"> - Loop Nest Analysis (Ettore)</font><br><font size="2"> - Data Dependence Graph (Bardia)</font><br><font size="2"> - DomTreeUpdater investigation (Kit)</font><br><font size="2"> - Follow up on multi-dimensional array indexing RFC (Siddharth)</font><br><font size="2"> - Impact of Loop Rotation on existing passes (Min-Yih)</font><br><font size="2"> - Bugzilla bugs update</font><br><br><br><font size="2">Bardia Mahjour<br>Compiler Optimizations<br>IBM Toronto Software Lab</font><BR>
</body></html>