| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Date: Wed, 10 Jun 2015 12:33:36 +0530
From: Srikar Dronamraju <srikar@...ux.vnet.ibm.com>
To: Ingo Molnar <mingo@...nel.org>,
Peter Zijlstra <peterz@...radead.org>
Cc: linux-kernel@...r.kernel.org, srikar@...ux.vnet.ibm.com,
Rik van Riel <riel@...hat.com>, Mel Gorman <mgorman@...e.de>
Subject: Performance numbers with prefer numa hotness over cache hotness
Ran 5 runs of autonuma-benchmark
(https://github.com/pholasek/autonuma-benchmark)
KernelVersion: 4.1.0-rc6
Testcase: Min Max Avg StdDev
elapsed_numa01: 533.73 789.12 696.77 86.86
elapsed_numa02: 22.59 28.15 25.87 1.99
Testcase: Min Max Avg StdDev
system_numa01: 342.78 1655.01 1128.26 479.49
system_numa02: 36.12 99.35 62.93 20.95
Testcase: Min Max Avg StdDev
user_numa01: 43724.45 62948.12 55832.97 6486.37
user_numa02: 1693.81 1945.30 1817.80 101.85
Testcase: Min Max Avg StdDev
total_numa01: 44067.20 64603.10 56961.22 6941.77
total_numa02: 1729.93 2044.65 1880.73 118.16
KernelVersion: 4.1.0-rc6+patch
Testcase: Min Max Avg StdDev %Change
elapsed_numa01: 484.08 726.45 648.46 85.14 6.65%
elapsed_numa02: 20.38 29.56 24.99 2.93 2.98%
Testcase: Min Max Avg StdDev %Change
system_numa01: 325.60 1111.46 837.76 290.98 26.14%
system_numa02: 47.65 73.59 60.59 10.97 3.18%
Testcase: Min Max Avg StdDev %Change
user_numa01: 39256.04 53387.31 48009.88 4717.79 14.65%
user_numa02: 1498.92 2089.51 1762.66 193.30 2.64%
Testcase: Min Max Avg StdDev %Change
total_numa01: 39581.60 54483.60 48847.66 4987.34 14.89%
total_numa02: 1546.57 2161.73 1823.26 202.28 2.66%
Performance counter stats for 'system wide': (for 1st run).
numa01
(Before patch)
5,81,911 cs [100.00%]
1,07,380 migrations [100.00%]
5,15,465 faults
2,99,19,83,196 cache-misses
724.330728328 seconds time elapsed
(After patch)
3,04,022 cs [100.00%]
47,539 migrations [100.00%]
3,14,508 faults
1,91,03,93,197 cache-misses
484.088557526 seconds time elapsed
numa02
(Before patch)
26,078 cs [100.00%]
4,932 migrations [100.00%]
28,025 faults
11,75,46,393 cache-misses
28.156065587 seconds time elapsed
(After patch)
18,972 cs [100.00%]
3,650 migrations [100.00%]
24,620 fault
9,19,64,503 cache-misses
20.382626292 seconds time elapsed
# numactl -H
available: 4 nodes (0-3)
node 0 cpus: 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
node 0 size: 32425 MB
node 0 free: 25493 MB
node 1 cpus: 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47
node 1 size: 31711 MB
node 1 free: 30933 MB
node 2 cpus: 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71
node 2 size: 30431 MB
node 2 free: 29577 MB
node 3 cpus: 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95
node 3 size: 32219 MB
node 3 free: 31455 MB
node distances:
node 0 1 2 3
0: 10 20 40 40
1: 20 10 40 40
2: 40 40 10 20
3: 40 40 20 10
Srikar Dronamraju (1):
sched:Prefer numa hotness over cache hotness
kernel/sched/fair.c | 96 ++++++++++++++++++-------------------------------
kernel/sched/features.h | 18 +++-------
2 files changed, 40 insertions(+), 74 deletions(-)
--
1.8.3.1
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists