aboutsummaryrefslogtreecommitdiff
path: root/parameters/learn.cpp
AgeCommit message (Expand)Author
2010-09-18make it work by having all indexes go via a central APIarjan
2010-09-17turn the parameters into a vector -- good for speed, and is getting ready for...arjan
2010-09-14fan works better if we use a quadratic equation to go from rpms to powerarjan
2010-09-13do a "fit with the to plot parameter assumed 0" for plottingarjan
2010-09-10slight visual/print improvementsarjan
2010-09-09make learning converge MUCH faster by having a minimum improvement thresholdarjan
2010-09-09drop the too selective tuning during calibration; bad ideaarjan
2010-09-08Add devices to the power guilty listarjan
2010-09-08learn smarter by doing one vector at a timearjan
2010-09-08more targeted learning during calibration in the hope to converge betterarjan
2010-09-05only show real alsa devicesarjan
2010-09-05start using the measurements to determine orderingarjan
2010-09-02typo fixesarjan
2010-09-02fix a lot of the accounting stuffarjan
2010-09-02more pstate workarjan
2010-08-29add C state data as wellarjan
2010-08-29dump actual measurement data for nowarjan
2010-08-28more USB device params.. and autolearn along the wayarjan
2010-08-27do a random-disturb pass during learning to home in on correlated parameters ...arjan
2010-08-27improve learning,arjan
2010-08-24learning frequencies seems to work quite well nowarjan
2010-08-22less memory leaks, more efficient learningarjan
2010-08-25use C++ strings not C strings to make memory management easierArjan van de Ven
2010-08-22uh oh.. associative arrays are breaking up on mearjan
2010-08-21use "Actual" not "Effective"; feedback from Lenarjan
2010-08-21add some very rudimentary learning logicarjan