Testing of Xen vPMU filter options.
Three tests are executed on a simple program (noploop, which runs NOP in a loop). The first two show Linux "perf stat" summaries, and the third tests specific counters: the IPC ones, an architectural one, and a few extras. Which of these will work depends on the filter mode.
I'll begin by showing off and on, the current modes, and then the new modes: ipc and arch.
root@lgud-bgregg:~> xl dmesg | grep vpmu
root@lgud-bgregg:~>
root@vm0hvm:~# perf stat ./noploop
Performance counter stats for './noploop':
1499.745527 task-clock (msec) # 0.982 CPUs utilized
54 context-switches # 0.036 K/sec
0 cpu-migrations # 0.000 K/sec
113 page-faults # 0.075 K/sec
<not supported> cycles
<not supported> stalled-cycles-frontend
<not supported> stalled-cycles-backend
<not supported> instructions
<not supported> branches
<not supported> branch-misses
1.526790639 seconds time elapsed
root@vm0hvm:~# perf stat -d ./noploop
Performance counter stats for './noploop':
1516.745765 task-clock (msec) # 0.999 CPUs utilized
50 context-switches # 0.033 K/sec
0 cpu-migrations # 0.000 K/sec
113 page-faults # 0.075 K/sec
<not supported> cycles
<not supported> stalled-cycles-frontend
<not supported> stalled-cycles-backend
<not supported> instructions
<not supported> branches
<not supported> branch-misses
<not supported> L1-dcache-loads
<not supported> L1-dcache-load-misses
<not supported> LLC-loads
<not supported> LLC-load-misses
1.517752197 seconds time elapsed
root@vm0hvm:~# perf stat -e r003c -e r013c -e r00c0 -e r4f2e -e r0180 -e r01a2 -e r010e ./noploop
Performance counter stats for './noploop':
<not supported> r003c
<not supported> r013c
<not supported> r00c0
<not supported> r4f2e
<not supported> r0180
<not supported> r01a2
<not supported> r010e
1.514596689 seconds time elapsed
root@lgud-bgregg:~> xl dmesg | grep vpmu
(XEN) Command line: placeholder vpmu=on loglvl=all guest_loglvl=all
root@vm0hvm:~# perf stat ./noploop
Performance counter stats for './noploop':
1507.440800 task-clock (msec) # 1.000 CPUs utilized
3 context-switches # 0.002 K/sec
0 cpu-migrations # 0.000 K/sec
113 page-faults # 0.075 K/sec
5,030,128,010 cycles # 3.337 GHz
7,213,421 stalled-cycles-frontend # 0.14% frontend cycles idle
4,965,684,648 stalled-cycles-backend # 98.72% backend cycles idle
20,043,873,623 instructions # 3.98 insns per cycle
# 0.25 stalled cycles per insn
10,691,703 branches # 7.093 M/sec
11,897 branch-misses # 0.11% of all branches
1.507885894 seconds time elapsed
root@vm0hvm:~# perf stat -d ./noploop
Performance counter stats for './noploop':
1507.844761 task-clock (msec) # 0.982 CPUs utilized
41 context-switches # 0.027 K/sec
0 cpu-migrations # 0.000 K/sec
113 page-faults # 0.075 K/sec
5,030,954,063 cycles # 3.337 GHz
7,740,438 stalled-cycles-frontend # 0.15% frontend cycles idle
4,966,031,978 stalled-cycles-backend # 98.71% backend cycles idle
20,044,349,973 instructions # 3.98 insns per cycle
# 0.25 stalled cycles per insn
10,785,189 branches # 7.153 M/sec
16,720 branch-misses # 0.16% of all branches
0 L1-dcache-loads # 0.000 K/sec
64,164 L1-dcache-load-misses # 0.00% of all L1-dcache hits
0 LLC-loads # 0.000 K/sec
<not supported> LLC-load-misses:HG
1.535188852 seconds time elapsed
root@vm0hvm:~# perf stat -e r003c -e r013c -e r00c0 -e r4f2e -e r0180 -e r01a2 -e r010e ./noploop
Performance counter stats for './noploop':
5,029,189,849 r003c
149,506,375 r013c
20,043,947,740 r00c0
24,034 r4f2e
1,271,411,108 r0180
411,934 r01a2
20,058,593,068 r010e
1.503972994 seconds time elapsed
root@lgud-bgregg:~> xl dmesg | grep vpmu
(XEN) Command line: placeholder vpmu=ipc loglvl=all guest_loglvl=all
root@vm0hvm:~# perf stat ./noploop
Performance counter stats for './noploop':
1507.341723 task-clock (msec) # 0.999 CPUs utilized
46 context-switches # 0.031 K/sec
0 cpu-migrations # 0.000 K/sec
113 page-faults # 0.075 K/sec
5,028,003,852 cycles # 3.336 GHz
0 stalled-cycles-frontend # 0.00% frontend cycles idle
0 stalled-cycles-backend # 0.00% backend cycles idle
20,043,375,516 instructions # 3.99 insns per cycle
0 branches # 0.000 K/sec
0 branch-misses # 0.00% of all branches
1.509467250 seconds time elapsed
root@vm0hvm:~# perf stat -d ./noploop
Performance counter stats for './noploop':
1511.326375 task-clock (msec) # 0.999 CPUs utilized
24 context-switches # 0.016 K/sec
0 cpu-migrations # 0.000 K/sec
113 page-faults # 0.075 K/sec
5,028,638,883 cycles # 3.327 GHz
0 stalled-cycles-frontend # 0.00% frontend cycles idle
0 stalled-cycles-backend # 0.00% backend cycles idle
20,043,427,933 instructions # 3.99 insns per cycle
0 branches # 0.000 K/sec
0 branch-misses # 0.00% of all branches
0 L1-dcache-loads # 0.000 K/sec
0 L1-dcache-load-misses # 0.00% of all L1-dcache hits
0 LLC-loads # 0.000 K/sec
<not supported> LLC-load-misses:HG
1.513292010 seconds time elapsed
root@vm0hvm:~# perf stat -e r003c -e r013c -e r00c0 -e r4f2e -e r0180 -e r01a2 -e r010e ./noploop
Performance counter stats for './noploop':
5,028,696,563 r003c
149,839,807 r013c
20,043,585,995 r00c0
0 r4f2e
0 r0180
0 r01a2
0 r010e
1.510042623 seconds time elapsed
root@lgud-bgregg:~> xl dmesg | grep vpmu
(XEN) Command line: placeholder vpmu=arch loglvl=all guest_loglvl=all
root@vm0hvm:~# perf stat ./noploop
Performance counter stats for './noploop':
1520.094185 task-clock (msec) # 0.998 CPUs utilized
51 context-switches # 0.034 K/sec
0 cpu-migrations # 0.000 K/sec
113 page-faults # 0.074 K/sec
5,032,584,693 cycles # 3.311 GHz
0 stalled-cycles-frontend # 0.00% frontend cycles idle
0 stalled-cycles-backend # 0.00% backend cycles idle
20,045,352,756 instructions # 3.98 insns per cycle
10,958,173 branches # 7.209 M/sec
16,925 branch-misses # 0.15% of all branches
1.522474349 seconds time elapsed
root@vm0hvm:~# perf stat -d ./noploop
Performance counter stats for './noploop':
1510.478919 task-clock (msec) # 0.998 CPUs utilized
48 context-switches # 0.032 K/sec
0 cpu-migrations # 0.000 K/sec
113 page-faults # 0.075 K/sec
5,030,765,239 cycles # 3.331 GHz
0 stalled-cycles-frontend # 0.00% frontend cycles idle
0 stalled-cycles-backend # 0.00% backend cycles idle
20,044,553,437 instructions # 3.98 insns per cycle
10,818,602 branches # 7.162 M/sec
15,169 branch-misses # 0.14% of all branches
0 L1-dcache-loads # 0.000 K/sec
0 L1-dcache-load-misses # 0.00% of all L1-dcache hits
0 LLC-loads # 0.000 K/sec
<not supported> LLC-load-misses:HG
1.513579213 seconds time elapsed
root@vm0hvm:~# perf stat -e r003c -e r013c -e r00c0 -e r4f2e -e r0180 -e r01a2 -e r010e ./noploop
Performance counter stats for './noploop':
5,030,582,397 r003c
150,300,076 r013c
20,044,024,689 r00c0
30,887 r4f2e
0 r0180
0 r01a2
0 r010e
1.516622998 seconds time elapsed
/proc/cpuinfo:
model name : Intel(R) Core(TM) i5-2400 CPU @ 3.10GHz
Hi, It's really interesting results, and I want to do the same experiment on my machine. Here are my questions:
Thanks.