Hi Korey, I'm confused. The miss_latency calculated by the sequencer is the miss latency of the particular request, not just L1 cache hits.
If you're seeing a bunch of minimum latency requests, I suspect something else is wrong. For instance, is "issued_time" a cycle value or a tick value? Brad > -----Original Message----- > From: [email protected] [mailto:[email protected]] > On Behalf Of Korey Sewell > Sent: Wednesday, April 20, 2011 11:38 AM > To: M5 Developer List > Subject: [m5-dev] Defining Miss Latencies in Ruby > > Hi all, > I've been working on miss latencies and stats in Ruby Caches and I noticed > something that might be a bug in tracking miss stats. > > The code in Sequencer.cc has the following check for looking at a miss: > "Time miss_latency = g_eventQueue_ptr->getTime() - issued_time; > > // Profile the miss latency for all non-zero demand misses > if (miss_latency != 0) {" > <track miss stats> > } > " > > Should this not instead be "L1_cache_latency" or "2 * L1_cache_latency" (if > it has to be buffered both ways)??? > > The effect of this I think is a saturation of the miss latency histogram in > the > 1st bucket. > > If anyone has any thoughts, let me know, as I could be missing something > here ... :) > > -- > - Korey > _______________________________________________ > m5-dev mailing list > [email protected] > http://m5sim.org/mailman/listinfo/m5-dev _______________________________________________ m5-dev mailing list [email protected] http://m5sim.org/mailman/listinfo/m5-dev
