Hi, everyone,
I don't really know if it is proper to post this topic here.
In order to accelerate our applications, I adopted the PREFETCH
instruction of UltraSparc. However, the results are not expected: the
one with prefetch works even a little bit slower. The whole program is
listed below.
I read the Ultrasparc specification and found that "when executed in
nonprivileged or privileged mode, PREFETCH has the same observable
effect as a NOP". I don't know how to understand this sentence. What
are the nonprivileged, privileged, and hyperprivileged modes in Linux?
Could that be user mode and kernel mode? I mean, does it mean I can
only make use of that instruction in kernel mode?
Thanks a lot!
testprefetch.c:
#include <stdlib.h>
#include <stdio.h>
#include <sys/time.h>
#define MEGA 1024*1024
#define TIMES 1
#define LASTITEM 100*MEGA-1
int main(){
struct timeval begin,end;
register int i,j;
int foo, bar;
int *data=(int *)malloc(100*sizeof(int)*MEGA); //400MB in
order to avoid being cached in L2 cache, which is 3M
register int *buf, *lastone;
lastone=&data[LASTITEM];
gettimeofday(&begin,NULL);
for(i=0;i<TIMES;i++){
buf=data;
while(buf<lastone){
__asm__ __volatile__("prefetch\t[%0],
#n_writes"::"r"(buf));
__asm__ __volatile__("prefetch\t[%0],
#n_writes"::"r"(buf+16));
__asm__ __volatile__("prefetch\t[%0],
#n_writes"::"r"(buf+32));
__asm__ __volatile__("prefetch\t[%0],
#n_writes"::"r"(buf+48));
for(j=0;j<100;j++) foo+=j; //do some trivial
things to pass time
bar=foo;
*buf=i;
*(buf+16)=i;
*(buf+32)=i;
*(buf+48)=i;
buf+=64;
}
}
gettimeofday(&end,NULL);
printf("time in usec is
%d\n",((end.tv_sec*1000000+end.tv_usec)-(begin.tv_sec*1000000+begin.tv_usec)));
}
--
Best regards,
Jiaqi
-
To unsubscribe from this list: send the line "unsubscribe sparclinux" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at http://vger.kernel.org/majordomo-info.html
