Hi Jin Yao
This is because every time the clk counter overflows, dtrace cpc resets
the rma counter to the initial value (MAX-100) and hence there are very
few occasions when the rma counter actually gets to overflow.
You should be able to observe this happening through the following
dtrace script.
==========
#!/usr/sbin/dtrace -s
sdt:pcbe.GenuineIntel.6.*:core_pcbe_*:wrmsr,
sdt:pcbe.GenuineIntel.6.*:core_pcbe_*:rdmsr
{
printf("%llX %lu(0x%llX)", arg0, (unsigned long)arg1, arg1);
}
sdt:pcbe.GenuineIntel.6.*:core_pcbe_sample:core-pcbe-sample
{
printf("%llX %lu(0x%lX) %lu(0x%lX) %lu(0x%lX)", arg0, (unsigned
long)arg1, arg1, (unsigned long)arg2, arg2, (unsigned long)arg3, arg3);
}
==========
collect(1), the other user of overflow profiling, samples the counters
on overflow and on the subsequent program, writes the last sampled value
instead of the initial value to the counter that did not overflow.
Jonathan, is the current behaviour what is expected of dtrace cpc or
should I file a bug?
/kuriakose
On 06/16/10 23:29, Jin Yao wrote:
I write 2 dtrace scripts in order to measure "clk" and "rma" for sytem workload
on nhm-ex (32 cores) when specjbb2005 runs.
r...@shz-os:~# ./test_clk_rma.d
......................
clk
309547
rma
32
kcpc_int
309579
......................
clk
309931
rma
57
kcpc_int
309985
......................
clk
310158
rma
25
kcpc_int
310183
^C
r...@shz-os:~# ./test_rma.d
............
rma
1531
kcpc_int
1531
............
rma
1645
kcpc_int
1645
............
rma
1537
kcpc_int
1537
I find the "rma" in test_clk_rma.d output is smaller than the "rma"
from test_rma.d. I guess some overflow interrupts lost when the dtrace
script test_clk_rma.d runs. But if it's true, why most of time are in
user space not in system space (from the output of vmstat and mpstat)?
The script sources are bellow and I also changed the setting to
"dcpc-min-overflow=100;" in "/kernel/drv/dcpc.conf" before tests.
r...@shz-os:~# cat test_clk_rma.d
#!/usr/sbin/dtrace -s
#pragma D option quiet
cpc:::cpu_clk_unhalted.ref-all-1000000
{
@clk = count();
}
cpc:::mem_uncore_retired.remote_dram-all-100
{
@rma = count();
}
kcpc_hw_overflow_intr:entry
{
@kcpc_int = count();
}
tick-5s
{
printf("......................\n");
printf("clk");
printa(@clk);
trunc(@clk);
printf("rma");
printa(@rma);
trunc(@rma);
printf("kcpc_int");
printa(@kcpc_int);
trunc(@kcpc_int);
}
r...@shz-os:~# cat test_rma.d
#!/usr/sbin/dtrace -s
#pragma D option quiet
cpc:::mem_uncore_retired.remote_dram-all-100
{
@rma = count();
}
kcpc_hw_overflow_intr:entry
{
@kcpc_int = count();
}
tick-5s
{
printf("............\n");
printf("rma");
printa(@rma);
trunc(@rma);
printf("kcpc_int");
printa(@kcpc_int);
trunc(@kcpc_int);
}
Can anybody give me some suggestions?
Thanks
Jin Yao
_______________________________________________
perf-discuss mailing list
perf-discuss@opensolaris.org
