[PATCH 0/5] x86,idle: Enhance menu governor C-state prediction

[Youquan Song]

The prediction for future is difficult and when the cpuidle governor prediction 
fails and govenor possibly choose the shallower C-state than it should. How to 
quickly notice and find the failure becomes important for power saving.    

cpuidle menu governor has a method to predict the repeat pattern if there are 8
C-states residency which are continuous and the same or very close, so it will
predict the next C-states residency will keep same residency time.

This patchset adds a timer when menu governor choose a non-deepest C-state in
order to wake up quickly from shallow C-state to avoid staying too long at 
shallow C-state for prediction failure. The timer is set to a time out value 
that is greater than predicted time and if the timer with the value is triggered 
, we can confidently conclude prediction is failure. When prediction
succeeds, CPU is waken up from C-states in predicted time and the timer is not 
triggered and will be cancelled right after CPU waken up. When prediction fails,
the timer is triggered to wake up CPU from shallow C-states, so menu governor 
will quickly notice that prediction fails and then re-evaluates deeper C-states
 possibility. This patchset can improves cpuidle prediction process for both 
repeat mode and general mode.

The patchset integrates one patch from Rik van Riel <riel@xxxxxxxxxx>, which try
to find a typical interval along with cut the upside outliers depends on
historical sleep intervals. The patch tends to choose a shallow C-state to
achieve better performance and ehancement of prediction failure will advise it
if the deepest C-state should be chosen.  

Testing result:

The whole patchset achieve good result after bunch of testing/tuning. 
Testing on two sockets Sandybridge server, SPECPower2008 get 2%~5% increase
ssj_ops/watt; Running benchmark in phoronix-test-suite: compress-7zip, 
build-linux-kernel, apache, fio etc, it also proves to increase the 
performance/power; What's more, it not only boosts the performance but also
saves power.  
 
There are also 2 cases will clear show this patchset benefit.

One case is turbostat utility (tools/power/x86/turbostat) at kernel 3.3 or early
. turbostat utility will read 10 registers one by one at Sandybridge, so it will
generate 10 IPIs to wake up idle CPUs. So cpuidle menu governor will predict it
 is repeat mode and there is another IPI wake up idle CPU soon, so it keeps idle
 CPU stay at C1 state even though CPU is totally idle. However, in the turbostat
, following 10 registers reading is sleep 5 seconds by default, so the idle CPU
 will keep at C1 for a long time though it is idle until break event occurs.
In a idle Sandybridge system, run "./turbostat -v", we will notice that deep 
C-state dangles between "70% ~ 99%". After patched the kernel, we will notice
deep C-state stays at >99.98%.

Below is another case which will clearly show the patch much benefit:

#include <stdlib.h>
#include <stdio.h>
#include <unistd.h>
#include <signal.h>
#include <sys/time.h>
#include <time.h>
#include <pthread.h>

volatile int * shutdown;
volatile long * count;
int delay = 20;
int loop = 8;

void usage(void)
{
	fprintf(stderr,
		"Usage: idle_predict [options]\n"
		"  --help	-h  Print this help\n"
		"  --thread	-n  Thread number\n"
		"  --loop     	-l  Loop times in shallow Cstate\n"
		"  --delay	-t  Sleep time (uS)in shallow Cstate\n");
}

void *simple_loop() {
	int idle_num = 1;
	while (!(*shutdown)) {
		*count = *count + 1;
	
		if (idle_num % loop)
			usleep(delay);
		else {
			/* sleep 1 second */
			usleep(1000000);
			idle_num = 0;
		}
		idle_num++;
	}

}

static void sighand(int sig)
{
	*shutdown = 1;
}

int main(int argc, char *argv[])
{
	sigset_t sigset;
	int signum = SIGALRM;
	int i, c, er = 0, thread_num = 8;
	pthread_t pt[1024];

	static char optstr[] = "n:l:t:h:";

	while ((c = getopt(argc, argv, optstr)) != EOF)
		switch (c) {
			case 'n':
				thread_num = atoi(optarg);
				break;
			case 'l':
				loop = atoi(optarg);
				break;
			case 't':
				delay = atoi(optarg);
				break;
			case 'h':
			default:
				usage();
				exit(1);
		}

	printf("thread=%d,loop=%d,delay=%d\n",thread_num,loop,delay);
	count = malloc(sizeof(long));
	shutdown = malloc(sizeof(int));
	*count = 0;
	*shutdown = 0;

	sigemptyset(&sigset);
	sigaddset(&sigset, signum);
	sigprocmask (SIG_BLOCK, &sigset, NULL);
	signal(SIGINT, sighand);
	signal(SIGTERM, sighand);

	for(i = 0; i < thread_num ; i++)
		pthread_create(&pt[i], NULL, simple_loop, NULL);

	for (i = 0; i < thread_num; i++)
		pthread_join(pt[i], NULL);

	exit(0);
}

Get powertop v2 from git://github.com/fenrus75/powertop, build powertop.
After build the above test application, then run it.
Test plaform can be Intel Sandybridge or other recent platforms.
#./idle_predict -l 10 &
#./powertop

We will find that deep C-state will dangle between 40%~100% and much time spent
on C1 state. It is because menu governor wrongly predict that repeat mode
is kept, so it will choose the C1 shallow C-state even though it has chance to
sleep 1 second in deep C-state.
 
While after patched the kernel, we find that deep C-state will keep >99.6%. 

Thanks for help from Arjan, Len Brown and Rik!

Thanks
-Youquan
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/