Author: sewardj Date: 2007-10-12 21:27:33 +0100 (Fri, 12 Oct 2007) New Revision: 6990
Log: Implement compressed shadow memory. Use it to support shadow values at byte granularity instead of 32-bit granularity. Net effect is to allow each byte to be shadowed by a 32-bit value whilst using about 5 bits/byte. Intent is to provide byte-level lockset tracking, whilst keeping memory consumption reasonable. Initial experiments suggest memory consumption is similar to Memcheck, although that can't really be true in the limit case (Memcheck = 2 bits/byte vs 4 bits/byte here). Unfortunately: * the implementation of shadow memory is scarily complex * this slows the tool down massively, by a factor of about 2.5x compared to the situation before this commit. Currently under investigation. Modified: branches/THRCHECK/thrcheck/tc_main.c [... diff too large to include ...] ------------------------------------------------------------------------- This SF.net email is sponsored by: Splunk Inc. Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now >> http://get.splunk.com/ _______________________________________________ Valgrind-developers mailing list Valgrind-developers@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/valgrind-developers