[estudiante2]

He encontrado un manual del SetCpu. En él se habla de los modos, está en ingles:

 Cita:

7. CPU Scaling Governors
CPU governors control exactly how the CPU scales between your "max" and "min" set frequencies. Most kernels have "ondemand" and "performance." The availability

• ondemand - Available in most kernels, and the default governor in most kernels. When the CPU load reaches a certain point (see "up threshold" in Advanced Settings), ondemand will rapidly scale the CPU up to meet demand, then gradually scale the CPU down when it isn't needed.
• conservative - Available in some kernels. It is similar to the ondemand governor, but will scale the CPU up more gradually to better fit demand. Conservative provides a less responsive experience than ondemand, but can save battery.
• performance - Available in most kernels. It will keep the CPU running at the "max" set value at all times. This is a bit more efficient than simply setting "max" and "min" to the same value and using ondemand because the system will not waste resources scanning for the CPU load. This governor is recommended for stable benchmarking.
• powersave - Available in some kernels. It will keep the CPU running at the "min" set value at all times.
• userspace- A method for controlling the CPU speed that isn't currently used by SetCPU. For best results, do not use the userspace governor.

Sacado de aquí: http://www.pokedev.com/setcpu

Faltaría el modo Interactive que deben haber añadido los cocineros de Xda, (-Snq concretamente creo), que debe mejorar la eficiencia del SetCpu.
Luego del modo Smartass, no tengo ni idea.
Edito para añadir el comentario del autor de las modificaciones snq- del foro Xda

 Cita:

Originally Posted by snq-
a few thoughts about the new kernel introduced with 2.2a (2.6.32.15-leedroid_2.2a):

as lee has already mentioned, it was built without perflock, allows over/underclocking (see available frequencies below [kHz]), and sets the cpu supply voltage based on the cpu frequency scaling
a wide range of cpufreq governors is also included:
Code:
$adb shell grep "." /sys/devices/system/cpu/cpu0/cpufreq/*avail* /sys/devices/system/cpu/cpu0/cpufreq/scaling_available_frequencies: 128000 245000 384000 422400 460800 499200 537600 576000 614400 652800 691200 729600 768000 806400 844800 883200 921600 960000 998400 1036800 1075200 1113600 /sys/devices/system/cpu/cpu0/cpufreq/scaling_available_governors: smartass interactive conservative ondemand powersave userspace performance
as some of you might prefer not to over/underclock (at least not with tools like setcpu), the kernel defaults to userspace:998400. in case you have setcpu installed and configured, this will be overwritten as soon as it starts.
you can still select your preferred governor/frequencies without setcpu through the sysfs interface (/sys/devices/system/cpu/cpu0/cpufreq: scaling_governor, scaling_setspeed, scaling_*_freq)
hacks like perflock-disable, kallsyms grepping when attaching externals governor modules, undervolting-hack have become redundant, lee removed the modules and the corresponding init scripts

popular drivers/file systems like cifs (with iso8859-1 and utf8 character set support), ext4, tun were compiled into the kernel, the only external module under /system/lib/modules is the wifi driver (modified based on the idea of myshkinbob)
still, due to some size-related optimizations, the boot image (kernel image + root fs) takes up only 1843200 bytes (900 blocks) in boot:. this is a bfd - and not only for people with a device that has bad blocks

in addition to the anticipatory, deadlineand and noop i/o scheduler, the cfq scheduler is also available (set as the default one)
(/sys/block//queue/scheduler)


all in all, in my opinion, having our own customized kernel specifically suited to this rom and supported by lee is a good thing (tm), and it could open up new, interesting possibilities in the future

Saludos.