----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: http://reviews.gem5.org/r/1145/ -----------------------------------------------------------
Review request for Default. Description ------- Ruby: Ensure order-dependent iteration uses an ordered map This patch fixes a bug in Ruby that caused non-deterministic simulation when changing the underlying hash map implementation. The reason is order-dependent behaviour in combination with iteration over the hash map contents. The two locations where a sorted container is assumed are now changed to make use of a std::map instead of the unordered hash map. With this change, the stats changes slightly and the follow-on changeset will update the relevant statistics. Diffs ----- src/mem/ruby/buffers/MessageBuffer.hh 5534a564f6a0 src/mem/ruby/system/TimerTable.hh 5534a564f6a0 src/mem/slicc/symbols/StateMachine.py 5534a564f6a0 Diff: http://reviews.gem5.org/r/1145/diff/ Testing ------- Confirmed that the changes caused the following two regressions to fail: ***** build/ALPHA_MOESI_hammer/tests/opt/quick/se/50.memtest/alpha/linux/memtest-ruby-MOESI_hammer FAILED! ***** build/ALPHA_MOESI_CMP_token/tests/opt/quick/se/50.memtest/alpha/linux/memtest-ruby-MOESI_CMP_token FAILED! This is in line with what was observed when using <unordered_map> instead of <ext/hash_map> Thanks, Andreas Hansson _______________________________________________ gem5-dev mailing list [email protected] http://m5sim.org/mailman/listinfo/gem5-dev
