5d4cf996cf1: -84.0% fileio.request_latency_max_ms
From: fengguang . wu
Date: Mon Dec 23 2013 - 00:14:39 EST
Hi Mel,
We are glad to reprort much improved fileio.request_latency_max_ms on commit
commit 5d4cf996cf134e8ddb4f906b8197feb9267c2b77
Author: Mel Gorman <mgorman@xxxxxxx>
Date: Tue Dec 17 09:21:25 2013 +0000
sched: Assign correct scheduling domain to 'sd_llc'
Commit 42eb088e (sched: Avoid NULL dereference on sd_busy) corrected a NULL
dereference on sd_busy but the fix also altered what scheduling domain it
used for the 'sd_llc' percpu variable.
One impact of this is that a task selecting a runqueue may consider
idle CPUs that are not cache siblings as candidates for running.
Tasks are then running on CPUs that are not cache hot.
This was found through bisection where ebizzy threads were not seeing equal
performance and it looked like a scheduling fairness issue. This patch
mitigates but does not completely fix the problem on all machines tested
implying there may be an additional bug or a common root cause. Here are
the average range of performance seen by individual ebizzy threads. It
was tested on top of candidate patches related to x86 TLB range flushing.
4-core machine
3.13.0-rc3 3.13.0-rc3
vanilla fixsd-v3r3
Mean 1 0.00 ( 0.00%) 0.00 ( 0.00%)
Mean 2 0.34 ( 0.00%) 0.10 ( 70.59%)
Mean 3 1.29 ( 0.00%) 0.93 ( 27.91%)
Mean 4 7.08 ( 0.00%) 0.77 ( 89.12%)
Mean 5 193.54 ( 0.00%) 2.14 ( 98.89%)
Mean 6 151.12 ( 0.00%) 2.06 ( 98.64%)
Mean 7 115.38 ( 0.00%) 2.04 ( 98.23%)
Mean 8 108.65 ( 0.00%) 1.92 ( 98.23%)
8-core machine
Mean 1 0.00 ( 0.00%) 0.00 ( 0.00%)
Mean 2 0.40 ( 0.00%) 0.21 ( 47.50%)
Mean 3 23.73 ( 0.00%) 0.89 ( 96.25%)
Mean 4 12.79 ( 0.00%) 1.04 ( 91.87%)
Mean 5 13.08 ( 0.00%) 2.42 ( 81.50%)
Mean 6 23.21 ( 0.00%) 69.46 (-199.27%)
Mean 7 15.85 ( 0.00%) 101.72 (-541.77%)
Mean 8 109.37 ( 0.00%) 19.13 ( 82.51%)
Mean 12 124.84 ( 0.00%) 28.62 ( 77.07%)
Mean 16 113.50 ( 0.00%) 24.16 ( 78.71%)
It's eliminated for one machine and reduced for another.
Signed-off-by: Mel Gorman <mgorman@xxxxxxx>
Signed-off-by: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
Cc: Alex Shi <alex.shi@xxxxxxxxxx>
Cc: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
Cc: Fengguang Wu <fengguang.wu@xxxxxxxxx>
Cc: H Peter Anvin <hpa@xxxxxxxxx>
Cc: Linus Torvalds <torvalds@xxxxxxxxxxxxxxxxxxxx>
Link: http://lkml.kernel.org/r/20131217092124.GV11295@xxxxxxx
Signed-off-by: Ingo Molnar <mingo@xxxxxxxxxx>
9dbdb155532395b 5d4cf996cf134e8ddb4f906b8
--------------- -------------------------
1898 ~110% -84.0% 303 ~28% snb-drag/sysbench/fileio/600s-100%-1HDD-btrfs-64G-1024-seqwr-sync
1898 -84.0% 303 TOTAL fileio.request_latency_max_ms
9dbdb155532395b 5d4cf996cf134e8ddb4f906b8
--------------- -------------------------
1712 ~ 3% +75.1% 2997 ~ 3% snb-drag/sysbench/fileio/600s-100%-1HDD-btrfs-64G-1024-seqwr-sync
1712 +75.1% 2997 TOTAL proc-vmstat.nr_tlb_remote_flush
9dbdb155532395b 5d4cf996cf134e8ddb4f906b8
--------------- -------------------------
1774 ~ 3% +74.3% 3093 ~ 3% snb-drag/sysbench/fileio/600s-100%-1HDD-btrfs-64G-1024-seqwr-sync
1774 +74.3% 3093 TOTAL proc-vmstat.nr_tlb_remote_flush_received
9dbdb155532395b 5d4cf996cf134e8ddb4f906b8
--------------- -------------------------
1707 ~ 2% +64.7% 2812 ~ 2% snb-drag/sysbench/fileio/600s-100%-1HDD-btrfs-64G-1024-seqwr-sync
1707 +64.7% 2812 TOTAL proc-vmstat.kswapd_high_wmark_hit_quickly
9dbdb155532395b 5d4cf996cf134e8ddb4f906b8
--------------- -------------------------
13752 ~ 4% -71.5% 3916 ~ 1% snb-drag/sysbench/fileio/600s-100%-1HDD-btrfs-64G-1024-seqwr-sync
13752 -71.5% 3916 TOTAL time.involuntary_context_switches
9dbdb155532395b 5d4cf996cf134e8ddb4f906b8
--------------- -------------------------
2797211 ~ 0% +22.8% 3434219 ~ 0% snb-drag/sysbench/fileio/600s-100%-1HDD-btrfs-64G-1024-seqwr-sync
2797211 +22.8% 3434219 TOTAL time.voluntary_context_switches
9dbdb155532395b 5d4cf996cf134e8ddb4f906b8
--------------- -------------------------
9885 ~ 0% +22.4% 12102 ~ 0% snb-drag/sysbench/fileio/600s-100%-1HDD-btrfs-64G-1024-seqwr-sync
9885 +22.4% 12102 TOTAL vmstat.system.cs
9dbdb155532395b 5d4cf996cf134e8ddb4f906b8
--------------- -------------------------
6 ~ 0% +16.7% 7 ~ 0% snb-drag/sysbench/fileio/600s-100%-1HDD-btrfs-64G-1024-seqwr-sync
6 +16.7% 7 TOTAL time.percent_of_cpu_this_job_got
9dbdb155532395b 5d4cf996cf134e8ddb4f906b8
--------------- -------------------------
39.61 ~ 0% +14.9% 45.50 ~ 0% snb-drag/sysbench/fileio/600s-100%-1HDD-btrfs-64G-1024-seqwr-sync
39.61 +14.9% 45.50 TOTAL time.system_time
Here are the visualized comparison of all GOOD/BAD commits during the bisect:
fileio.request_latency_max_ms
8000 ++------------------------------------------------------------------+
| * |
7000 *+ * : * |
6000 ++ : * : : |
| : : : : |
5000 ++ :: : :: : : |
|: :: :: :: : : |
4000 ++ :: :: : : : : |
|:: : :: : : : : |
3000 ++: : : : : : : : |
2000 ++: : : : : : : : |
| : : : : : : : : |
1000 ++: : : :.* .*.* : : .*. : * |
| * **.*.* * *.** *.*.** O*.*.**.** **.*.**.**.*.** *.**.*
0 O+OO-OO-O-OO-O-OO-OO-O-O----------O-OO-OO-O-OO-O-O--OO-O--O-O-OO----+
time.voluntary_context_switches
3.5e+06 ++---------------------------------------------------------------+
O OO OO OO OO OO O OO OO OO OO O OO OO O O O OO O OO OO |
3.4e+06 ++ O O O O O |
| |
3.3e+06 ++ |
3.2e+06 ++ |
| |
3.1e+06 ++ |
| |
3e+06 ++ |
2.9e+06 ++ |
| * *. *. |
2.8e+06 *+* .* + *.* .**. : * **.*.**. *. *.* .**. *.* .* .* .* .*
| * * *.**.* * * * * *.* * * * * |
2.7e+06 ++---------------------------------------------------------------+
time.involuntary_context_switches
16000 ++-----------------------------------------------------------------+
| * |
14000 ++ :+ .**.* .**. * .* .* .**. .*.* *.*.* *. *.**.*
*.**.* * *.* **. + * *.* * ** *.* *.* *.* |
12000 ++ * |
| |
10000 ++ |
| |
8000 ++ |
| |
6000 ++ |
| |
4000 O+OO OO O OO OO O OO OO O OO OO O OO OO OO O OO OO O OO OO O OO |
| |
2000 ++-----------------------------------------------------------------+
vmstat.system.cs
12500 ++-----------------------------------------------------------------+
| O O O |
12000 O+OO OO O OO OO O OO OO O OO OO O OO OO OO O OO OO OO O O |
| O |
| |
11500 ++ |
| |
11000 ++ |
| |
10500 ++ |
| |
*. *. * *. *. .* |
10000 ++* .* + * + *.* + * **.*.**.**.** + *.**.*.**. *. .* .* .*
| * * **.*.* * * * * * * |
9500 ++-----------------------------------------------------------------+
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/