| Age | Commit message (Collapse) | Author |
|---|
|
The amount of instructions should be the same regardless what
method you use (1 or 2) but the cache misses will be very
different. Looking at these statistics side by side clearly
illustrate the problem.
Update the README to also include the instruction count.
Signed-off-by: John Ogness <john.ogness@linutronix.de>
|
|
mention some other usage of perf to track done the problem
[Reformatted by jogness]
Signed-off-by: Manuel Traut <manut@linutronix.de>
Signed-off-by: John Ogness <john.ogness@linutronix.de>
|
|
Different tools have been used by various trainers as demos. Put
all these into master so they are available to all trainers.
ipc_pipe: ipc demo using pipes
ipc_shm: ipc demo using shared memory
libduma: source and instructions for compiling libduma
matrix: demo of good and bad cache access
mtrace: patch and infos for using mtrace with ASLR
rtex: demo of handling page faults
Signed-off-by: John Ogness <john.ogness@linutronix.de>
|
