1*101025ffSozkanonurToDos sorted by priority: 2*101025ffSozkanonur 3*101025ffSozkanonur- Use bitmask functions to parse CPU topology more robust 4*101025ffSozkanonur (current implementation has issues on AMD) 5*101025ffSozkanonur- Try to read out boost states and frequencies on Intel 6*101025ffSozkanonur- Somewhere saw the ability to read power consumption of 7*101025ffSozkanonur RAM from HW on Intel SandyBridge -> another monitor? 8*101025ffSozkanonur- Add another c1e debug idle monitor 9*101025ffSozkanonur -> Is by design racy with BIOS, but could be added 10*101025ffSozkanonur with a --force option and some "be careful" messages 11*101025ffSozkanonur- Add cpu_start()/cpu_stop() callbacks for monitor 12*101025ffSozkanonur -> This is to move the per_cpu logic from inside the 13*101025ffSozkanonur monitor to outside it. This can be given higher 14*101025ffSozkanonur priority in fork_it. 15*101025ffSozkanonur- Fork as many processes as there are CPUs in case the 16*101025ffSozkanonur per_cpu_schedule flag is set. 17*101025ffSozkanonur -> Bind forked process to each cpu. 18*101025ffSozkanonur -> Execute start measures via the forked processes on 19*101025ffSozkanonur each cpu. 20*101025ffSozkanonur -> Run test executable in a forked process. 21*101025ffSozkanonur -> Execute stop measures via the forked processes on 22*101025ffSozkanonur each cpu. 23*101025ffSozkanonur This would be ideal as it will not introduce noise in the 24*101025ffSozkanonur tested executable. 25