Messages in this thread | | | Date | Mon, 29 Sep 2014 14:19:27 +0800 | From | Fengguang Wu <> | Subject | [xen/spinlock] e0fc17a9363: +24.6% aim9.dir_rtns_1.ops_per_sec |
| |
Hi Konrad,
We are glad to find that your patch increases the aim9 test performance by up to +24.6%!
e0fc17a936334c08b2729fff87168c03fdecf5b6 ("xen/spinlock: Don't enable them unconditionally.")
test case: brickland3/aim9/5s-all
brickland3 is an Ivy Bridge-EX with 512G memory.
c0914e61660fa7d e0fc17a936334c08b2729fff8 --------------- ------------------------- 10401445 ± 0% +24.6% 12958400 ± 0% TOTAL aim9.dir_rtns_1.ops_per_sec 494461 ± 5% +8.0% 534028 ± 1% TOTAL aim9.link_test.ops_per_sec 1840820 ± 0% +4.5% 1924528 ± 0% TOTAL aim9.fifo_test.ops_per_sec 1251672 ± 0% +5.0% 1314071 ± 0% TOTAL aim9.dgram_pipe.ops_per_sec 1097240 ± 1% +5.0% 1152480 ± 0% TOTAL aim9.signal_test.ops_per_sec 1407192 ± 0% +4.8% 1474424 ± 0% TOTAL aim9.stream_pipe.ops_per_sec 1898068 ± 1% +4.7% 1987232 ± 0% TOTAL aim9.pipe_cpy.ops_per_sec 1298388 ± 1% +3.5% 1343704 ± 0% TOTAL aim9.shared_memory.ops_per_sec 64202 ± 0% -3.2% 62178 ± 1% TOTAL aim9.misc_rtns_1.ops_per_sec -12622 ±-16% +32.4% -16710 ±-14% TOTAL sched_debug.cfs_rq[116]:/.spread0 -1977 ±-26% -59.3% -805 ±-42% TOTAL cpuidle.C3-IVB.usage -13457 ±-8% +19.2% -16045 ±-17% TOTAL sched_debug.cfs_rq[14]:/.spread0 -7867 ±-42% +61.3% -12692 ±-10% TOTAL sched_debug.cfs_rq[18]:/.spread0 -4 ±-27% -36.4% -2 ±-34% TOTAL sched_debug.cpu#19.nr_uninterruptible -12763 ±-8% +19.1% -15202 ±-10% TOTAL sched_debug.cfs_rq[24]:/.spread0 -13552 ±-8% +20.5% -16333 ±-14% TOTAL sched_debug.cfs_rq[97]:/.spread0 1 ± 0% -100.0% 0 ± 0% TOTAL sched_debug.cfs_rq[105]:/.nr_spread_over 1857 ±48% -69.7% 563 ±20% TOTAL sched_debug.cfs_rq[97]:/.min_vruntime 111 ±44% -73.0% 30 ±42% TOTAL sched_debug.cfs_rq[4]:/.blocked_load_avg 160 ±48% -60.5% 63 ±26% TOTAL sched_debug.cpu#104.sched_goidle 1288 ±37% -58.1% 540 ±37% TOTAL sched_debug.cfs_rq[67]:/.min_vruntime 988 ±36% -49.7% 497 ±46% TOTAL sched_debug.cpu#55.ttwu_count 114 ±20% -59.4% 46 ±24% TOTAL sched_debug.cfs_rq[48]:/.tg_load_contrib 101 ±16% -54.2% 46 ±24% TOTAL sched_debug.cfs_rq[48]:/.blocked_load_avg 32.50 ±41% -52.7% 15.37 ±40% TOTAL sched_debug.cfs_rq[66]:/.exec_clock 38 ±29% -39.3% 23 ±43% TOTAL sched_debug.cfs_rq[44]:/.avg->runnable_avg_sum 469 ±24% -40.5% 279 ±47% TOTAL sched_debug.cpu#26.ttwu_local 1754 ±34% +113.5% 3744 ±14% TOTAL sched_debug.cpu#40.sched_count 123 ±27% -49.9% 62 ±36% TOTAL sched_debug.cfs_rq[47]:/.blocked_load_avg 151 ±12% -40.8% 89 ±31% TOTAL sched_debug.cpu#93.ttwu_count 68 ±44% -49.6% 34 ±27% TOTAL sched_debug.cfs_rq[50]:/.avg->runnable_avg_sum 5584 ±20% +40.0% 7818 ±18% TOTAL sched_debug.cpu#84.nr_load_updates 411 ±30% -36.1% 262 ±24% TOTAL sched_debug.cpu#23.ttwu_local 1655 ±33% -45.4% 903 ±32% TOTAL sched_debug.cpu#2.sched_goidle 3341 ±33% -45.0% 1836 ±31% TOTAL sched_debug.cpu#2.nr_switches 8488 ±11% +24.1% 10538 ±13% TOTAL sched_debug.cpu#10.nr_load_updates 19 ± 7% +25.0% 24 ±12% TOTAL sched_debug.cpu#114.ttwu_local 8752 ±13% +23.9% 10842 ±11% TOTAL sched_debug.cpu#8.nr_load_updates 24 ±10% -16.9% 20 ± 9% TOTAL sched_debug.cpu#68.ttwu_local 901 ±38% -46.5% 482 ±40% TOTAL sched_debug.cpu#62.sched_goidle 1862 ±37% -43.9% 1045 ±35% TOTAL sched_debug.cpu#62.nr_switches 14555 ± 7% +13.4% 16512 ± 8% TOTAL sched_debug.cpu#106.nr_load_updates 5433 ±10% +20.4% 6542 ± 5% TOTAL sched_debug.cpu#0.ttwu_local
aim9.signal_test.ops_per_sec
1.16e+06 O+-O----O----------------------O-------O----O--O-O--O--O----O-O--+ | O O O O O O O O O O | 1.14e+06 ++ O O O O | | | | 1.12e+06 ++ | | *..*. .*..*.. *..*..*.*..*.*..*..* | 1.1e+06 ++ : *..* *.*.. : : | | : : : | 1.08e+06 *+. *.. : * : | | + *.* : | | * : | 1.06e+06 ++ * | | | 1.04e+06 ++---------------------------------------------------------------+
aim9.dir_rtns_1.ops_per_sec
1.4e+07 ++---------------------------------------------------------------+ | O O O O O O O O O O O | 1.35e+07 ++ O | 1.3e+07 ++ O O O O O O | O O O O O O O O 1.25e+07 ++ | | | 1.2e+07 ++ | | | 1.15e+07 ++ | 1.1e+07 ++ | | | 1.05e+07 *+.*.*..*.*..*..*. .*.*..*.*..*.. .*.. | | *..*.*..*. *.*..*.*. *.* | 1e+07 ++---------------------------------------------------------------+
[*] bisect-good sample [O] bisect-bad sample
To reproduce:
apt-get install ruby ruby-oj git clone git://git.kernel.org/pub/scm/linux/kernel/git/wfg/lkp-tests.git cd lkp-tests bin/setup-local job.yaml # the job file attached in this email bin/run-local job.yaml
Disclaimer: Results have been estimated based on internal Intel analysis and are provided for informational purposes only. Any difference in system hardware or software design or configuration may affect actual performance.
Thanks, Fengguang --- testcase: aim9 default_monitors: watch-oom: wait: pre-test uptime: iostat: vmstat: numa-numastat: numa-vmstat: numa-meminfo: proc-vmstat: proc-stat: meminfo: slabinfo: interrupts: lock_stat: latency_stats: softirqs: bdi_dev_mapping: diskstats: energy: cpuidle: cpufreq: turbostat: sched_debug: interval: 10 pmeter: model: Brickland Ivy Bridge-EX nr_cpu: 120 memory: 512G hdd_partitions: swap_partitions: aim9: testtime: 5s test: - all branch: linus/master commit: 19583ca584d6f574384e17fe7613dfaeadcdc4a6 repeat_to: 3 enqueue_time: 2014-09-25 21:56:15.069539322 +08:00 testbox: brickland3 kconfig: x86_64-rhel kernel: "/kernel/x86_64-rhel/19583ca584d6f574384e17fe7613dfaeadcdc4a6/vmlinuz-3.16.0" user: lkp queue: wfg result_root: "/result/brickland3/aim9/5s-all/debian-x86_64.cgz/x86_64-rhel/19583ca584d6f574384e17fe7613dfaeadcdc4a6/0" job_file: "/lkp/scheduled/brickland3/wfg_aim9-5s-all-x86_64-rhel-19583ca584d6f574384e17fe7613dfaeadcdc4a6-2.yaml" dequeue_time: 2014-09-25 22:35:03.558121828 +08:00 history_time: 351.36 job_state: finished loadavg: 0.96 0.69 0.32 1/969 35857 start_time: '1411655785' end_time: '1411656085' version: "/lkp/lkp/.src-20140925-212910"
| |