06-22-2015, 11:58 PM
Interesting (maybe) followup to this. FYI - this was running the 'naxxatoe' wordlist - which has a huge swath of < 8 char in the middle.
I noticed once my run hit a 'dry spell' (i.e. < 8 char for WPA) and therefore no hashing going on - the temp dropped. This was expected. However - afterward - once the run hit valid words in the dict and hashing resumed - the fans were never adjusted (or at least quick enough). My cards hit 90c and abort kicked in.
I'm guessing that the internal calculation for temp / fan isn't reacting fast enough to a pattern like this (extended period of no hashing followed by resume). I.e. in whatever way the average is getting calculated - theres too much low temp data dragging it down and reaction is too slow. I even tried setting the target temp on the command line - but once the 'dry spell' starts - it still didn't kick up the fans after.
I noticed once my run hit a 'dry spell' (i.e. < 8 char for WPA) and therefore no hashing going on - the temp dropped. This was expected. However - afterward - once the run hit valid words in the dict and hashing resumed - the fans were never adjusted (or at least quick enough). My cards hit 90c and abort kicked in.
I'm guessing that the internal calculation for temp / fan isn't reacting fast enough to a pattern like this (extended period of no hashing followed by resume). I.e. in whatever way the average is getting calculated - theres too much low temp data dragging it down and reaction is too slow. I even tried setting the target temp on the command line - but once the 'dry spell' starts - it still didn't kick up the fans after.