From a30c5371bf2885334a99446fdb975fa4bf5c7128 Mon Sep 17 00:00:00 2001 From: Sven Verdoolaege Date: Tue, 21 Mar 2017 14:44:12 +0100 Subject: [PATCH] doc: explain must dependences in the result of dataflow analysis Reported-by: Siddharth Bhat Signed-off-by: Sven Verdoolaege --- doc/user.pod | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/doc/user.pod b/doc/user.pod index 9b237376..31689598 100644 --- a/doc/user.pod +++ b/doc/user.pod @@ -8726,9 +8726,12 @@ pairs of corresponding sink iterations and accessed data elements. To compute standard flow dependences, the sink should be a read, while the sources should be writes. If any of the source accesses are marked as being I -accesses, then there will be a dependence from the last +accesses, then there will be a (may) dependence from the last I access B from any I access that follows -this last I access. +this last I access, but still precedes the sink access. +Only dependences originating in a must access and without +any may accesses between the must access and the sink access +are considered to be must dependences. In particular, if I sources are I accesses, then memory based dependence analysis is performed. If, on the other hand, all sources are I accesses, -- 2.11.4.GIT