summaryrefslogtreecommitdiffstats
path: root/sys/ia64
diff options
context:
space:
mode:
authorphk <phk@FreeBSD.org>2006-02-07 21:22:02 +0000
committerphk <phk@FreeBSD.org>2006-02-07 21:22:02 +0000
commitbb2f62f5367d29657fc52a4cfcd3573c39679b70 (patch)
tree95918de865b5c6e707ecd4ce458d058d287406e9 /sys/ia64
parentc7f3a565d2e7b4fed16d19bfe4b300ca8ef7ad27 (diff)
downloadFreeBSD-src-bb2f62f5367d29657fc52a4cfcd3573c39679b70.zip
FreeBSD-src-bb2f62f5367d29657fc52a4cfcd3573c39679b70.tar.gz
Modify the way we account for CPU time spent (step 1)
Keep track of time spent by the cpu in various contexts in units of "cputicks" and scale to real-world microsec^H^H^H^H^H^H^H^Hclock_t only when somebody wants to inspect the numbers. For now "cputicks" are still derived from the current timecounter and therefore things should by definition remain sensible also on SMP machines. (The main reason for this first milestone commit is to verify that hypothesis.) On slower machines, the avoided multiplications to normalize timestams at every context switch, comes out as a 5-7% better score on the unixbench/context1 microbenchmark. On more modern hardware no change in performance is seen.
Diffstat (limited to 'sys/ia64')
-rw-r--r--sys/ia64/ia64/mp_machdep.c2
1 files changed, 1 insertions, 1 deletions
diff --git a/sys/ia64/ia64/mp_machdep.c b/sys/ia64/ia64/mp_machdep.c
index dedc395..0b9e7dd 100644
--- a/sys/ia64/ia64/mp_machdep.c
+++ b/sys/ia64/ia64/mp_machdep.c
@@ -138,7 +138,7 @@ ia64_ap_startup(void)
spinlock_exit();
KASSERT(curthread->td_md.md_spinlock_count == 1, ("invalid count"));
- binuptime(PCPU_PTR(switchtime));
+ PCPU_SET(switchtime, cpu_ticks());
PCPU_SET(switchticks, ticks);
ia64_set_tpr(0);
OpenPOWER on IntegriCloud