CVS difference for ai12s/ai12-0111-1.txt

Differences between 1.8 and version 1.9
Log of other versions for file ai12s/ai12-0111-1.txt

--- ai12s/ai12-0111-1.txt	2016/07/11 22:54:58	1.8
+++ ai12s/ai12-0111-1.txt	2016/08/19 03:02:36	1.9
@@ -2492,11 +2492,58 @@
 From: Tucker Taft
 Sent: Tuesday, June 7, 2016  9:54 PM
 
-> ... Overall, I can't figure out how to get from your suggested change 
+> ... Overall, I can't figure out how to get from your suggested change
 > to your intended result.
 
 Actually, you seemed to have figured out my intent.  Hopefully it will also be
 clear to the rest of the ARG.  We can fix up the wording after the meeting, if
 there is support for the approach.
+
+***************************************************************
+
+From: Erhard Ploedereder
+Sent: Sunday, June 12, 2016  8:53 AM
+
+In response to questions raised...
+
+It was the reading from Containers that was exceptionally slow. Writing was
+reasonably competitive with other languages. Hashed sets were called out as
+culprits (this may well be an "e.g." rather than pointing out the only sinner);
+the difference to other languages was 4x to 8x execution time. Other containers
+used were Vectors, Doubly_linked_lists and Hash maps. (Orthogonally: Use of
+indefinite containers in contrast to the "normal" ones added a factor of ~2.)
+
+***************************************************************
+
+From: Randy Brukardt
+Sent: Monday, July 11, 2016  9:17 PM
+
+Could you find out *which* operations that were used to read from Containers?
+
+I'd expect the performance of Element, Query_Element, and Constant_Reference
+(the three main ways to read) to be very different (with Query_Element being the
+fastest and Element being the slowest for most element types). (Of course,
+regular indexed notation expands into calls on Constant_Reference.) After all,
+Element essentially makes a copy of an element, whereas the other two use a
+direct reference into the element.
+
+***************************************************************
+
+From: Jeff Cousins
+Sent: Tuesday, August 16, 2016  6:21 AM
+
+Erhard - do you know specifically which reads were slow?  Presumably you've
+already done the legwork of taking the measurements, but which did you measure?
+
+***************************************************************
+
+From: Erhard Ploedereder
+Sent: Wednesday, August 17, 2016  7:18 AM
+
+A while ago, I cleaned the experiment from my disk, as changes to the containers
+made the results obsolete anyway. As such, I have to see whether I can retrieve
+this from backup somehow and I have not got around to doing the archeology. It
+was not my experiment, but one of my assistants, so mmy memory alone does not
+serve.
 
 ***************************************************************

Questions? Ask the ACAA Technical Agent