CVS difference for ai12s/ai12-0125-3.txt

Differences between 1.15 and version 1.16
Log of other versions for file ai12s/ai12-0125-3.txt

--- ai12s/ai12-0125-3.txt	2016/12/22 03:29:34	1.15
+++ ai12s/ai12-0125-3.txt	2016/12/28 04:20:39	1.16
@@ -4070,3 +4070,60 @@
 
 ****************************************************************
 
+From: Florian Schanda
+Sent: Wednesday, November 2, 2016  3:09 AM
+
+Yesterday we had the HIS-2016 conference in Bristol. Because one of the
+speakers could not turn up I did a replacement talk on Ada 202x.
+
+Of course I mentioned @ and *nobody* seemed to mind. Nobody complained to me
+in the questions, or privately afterwards. So it can't be that hideous.
+
+On the other hand, people really liked the parallel blocks and that we're
+attacking parallelism.
+
+****************************************************************
+
+From: Tucker Taft
+Sent: Wednesday, November 2, 2016  7:33 AM
+
+...
+> Of course I mentioned @ and *nobody* seemed to mind. Nobody complained
+> to me in the questions, or privately afterwards. So it can't be that hideous.
+
+Glad to hear it.
+
+> On the other hand, people really liked the parallel blocks and that
+> we're attacking parallelism.
+
+Also good to hear.
+
+****************************************************************
+
+From: Florian Schanda
+Sent: Thursday, November 10, 2016  8:50 AM
+
+I've now given another presentation internal in Altran to ~20 Ada programmers.
+People loved the @. A *lot* of discussion on extended Ravenscar and parallel
+blocks (but a lot of it was "why can't the compiler just to it all magically").
+
+I did also include generators and after the initial shock people were quite
+intrigued.
+
+****************************************************************
+
+From: Brad Moore
+Sent: Thursday, November 10, 2016  11:32 PM
+
+I believe compilers will be able to add more and more parallelism implicitly,
+with improved static support from the language. For example, the Global aspect
+and the Potentially Blocking aspect should allow the compiler to safely
+parallelize more stuff, as the compiler will in theory know what data races
+exist and need to be avoided.
+
+However there are cases where I doubt the compiler will be able to parallelize
+without some help from the programmer. In particular, where user defined
+reductions are involved.
+
+****************************************************************
+

Questions? Ask the ACAA Technical Agent