[RFC v4 0/7] Implement Data Access Monitoring-based Memory Operation Schemes

From: SeongJae Park
Date: Tue Mar 03 2020 - 07:14:53 EST

From: SeongJae Park <sjpark@xxxxxxxxx>

DAMON[1] can be used as a primitive for data access awared memory management
optimizations. That said, users who want such optimizations should run DAMON,
read the monitoring results, analyze it, plan a new memory management scheme,
and apply the new scheme by themselves. Such efforts will be inevitable for
some complicated optimizations.

However, in many other cases, the users could simply want the system to apply a
memory management action to a memory region of a specific size having a
specific access frequency for a specific time. For example, "page out a memory
region larger than 100 MiB keeping only rare accesses more than 2 minutes", or
"Do not use THP for a memory region larger than 2 MiB rarely accessed for more
than 1 seconds".

This RFC patchset makes DAMON to handle such data access monitoring-based
operation schemes. With this change, users can do the data access awared
optimizations by simply specifying their schemes to DAMON.


Evaluations
===========

Efficient THP
-------------

Transparent Huge Pages (THP) subsystem could waste memory space in some cases
because it aggressively promotes regular pages to huge pages. For the reason,
use of THP is prohivited by a number of memory intensive programs such as
Redis[1] and MongoDB[2].

Below two simple data access monitoring-based operation schemes might be
helpful for the problem:

# format: <min/max size> <min/max frequency (0-100)> <min/max age> <action>

# If a memory region larger than 2 MiB is showing access rate higher than
# 5%, apply MADV_HUGEPAGE to the region.
2M null 5 null null null hugepage

# If a memory region larger than 2 MiB is showing access rate lower than 5%
# for more than 1 second, apply MADV_NOHUGEPAGE to the region.
2M null null 5 1s null nohugepage

We can expect the schmes would reduce the memory space overhead but preserve
some of the performance benefit of THP. I call this schemes Efficient THP
(ETHP).

Please note that these schemes are neither highly tuned nor for general
usecases. These are made with my straightforward instinction for only a
demonstration of DAMOS.


Setup
-----

On my personal QEMU/KVM based virtual machine on an Intel i7 host machine
running Ubuntu 18.04, I measure runtime and consumed memory space of various
realistic workloads with several configurations. I use 13 and 12 workloads in
PARSEC3[3] and SPLASH-2X[4] benchmark suites, respectively. I personally use
another wrapper scripts[5] for setup and run of the workloads.

For the measurement of the amount of consumed memory in system global scope, I
drop caches before starting each of the workloads and monitor 'MemFree' in the
'/proc/meminfo' file.

The configurations I use are as below:

orig: Linux v5.5 with 'madvise' THP policy
thp: Linux v5.5 with 'always' THP policy
ethp: Linux v5.5 applying the above schemes

To minimize the measurement errors, I repeat the run 5 times and average
results. You can get stdev, min, and max of the numbers among the repeated
runs in appendix below.


[1] "Redis latency problems troubleshooting", https://redis.io/topics/latency
[2] "Disable Transparent Huge Pages (THP)",
https://docs.mongodb.com/manual/tutorial/transparent-huge-pages/
[3] "The PARSEC Becnhmark Suite", https://parsec.cs.princeton.edu/index.htm
[4] "SPLASH-2x", https://parsec.cs.princeton.edu/parsec3-doc.htm#splash2x
[5] "parsec3_on_ubuntu", https://github.com/sjp38/parsec3_on_ubuntu


Results
-------

TL;DR: 'ethp' removes 97.61% of 'thp' memory space overhead while preserving
25.40% (up to 88.36%) of 'thp' performance improvement in total.

Following sections show the results of the measurements with raw numbers and
'orig'-relative overheads (percent) of each configuration.


Memory Space Overheads
~~~~~~~~~~~~~~~~~~~~~~

Below shows measured memory space overheads. Raw numbers are in KiB, and the
overheads in parentheses are in percent. For example, 'parsec3/blackscholes'
consumes about 1.819 GiB and 1.824 GiB with 'orig' and 'thp' configuration,
respectively. The overhead of 'thp' compared to 'orig' for the workload is
0.3%.

workloads orig thp (overhead) ethp (overhead)
parsec3/blackscholes 1819486.000 1824921.400 ( 0.30) 1829070.600 ( 0.53)
parsec3/bodytrack 1417885.800 1417077.600 ( -0.06) 1427560.800 ( 0.68)
parsec3/canneal 1043876.800 1039773.000 ( -0.39) 1048445.200 ( 0.44)
parsec3/dedup 2400000.400 2434625.600 ( 1.44) 2417374.400 ( 0.72)
parsec3/facesim 540206.400 542422.400 ( 0.41) 551485.400 ( 2.09)
parsec3/ferret 320480.200 320157.000 ( -0.10) 331470.400 ( 3.43)
parsec3/fluidanimate 573961.400 572329.600 ( -0.28) 581836.000 ( 1.37)
parsec3/freqmine 983981.200 994839.600 ( 1.10) 996124.600 ( 1.23)
parsec3/raytrace 1745175.200 1742756.400 ( -0.14) 1751706.000 ( 0.37)
parsec3/streamcluster 120558.800 120309.800 ( -0.21) 131997.800 ( 9.49)
parsec3/swaptions 14820.400 23388.800 ( 57.81) 24698.000 ( 66.65)
parsec3/vips 2956319.200 2955803.600 ( -0.02) 2977506.200 ( 0.72)
parsec3/x264 3187699.000 3184944.000 ( -0.09) 3198462.800 ( 0.34)
splash2x/barnes 1212774.800 1221892.400 ( 0.75) 1212100.800 ( -0.06)
splash2x/fft 9364725.000 9267074.000 ( -1.04) 8997901.200 ( -3.92)
splash2x/lu_cb 515242.400 519881.400 ( 0.90) 526621.600 ( 2.21)
splash2x/lu_ncb 517308.000 520396.400 ( 0.60) 521732.400 ( 0.86)
splash2x/ocean_cp 3348189.400 3380799.400 ( 0.97) 3328473.400 ( -0.59)
splash2x/ocean_ncp 3908599.800 7072076.800 ( 80.94) 4449410.400 ( 13.84)
splash2x/radiosity 1469087.800 1482244.400 ( 0.90) 1471781.000 ( 0.18)
splash2x/radix 1712487.400 1385972.800 (-19.07) 1420461.800 (-17.05)
splash2x/raytrace 45030.600 50946.600 ( 13.14) 58586.200 ( 30.10)
splash2x/volrend 151037.800 151188.000 ( 0.10) 163213.600 ( 8.06)
splash2x/water_nsquared 47442.400 47257.000 ( -0.39) 59285.800 ( 24.96)
splash2x/water_spatial 667355.200 666824.400 ( -0.08) 673274.400 ( 0.89)
total 40083800.000 42939900.000 ( 7.13) 40150600.000 ( 0.17)

In total, 'thp' shows 7.13% memory space overhead while 'ethp' shows only 0.17%
overhead. In other words, 'ethp' removed 97.61% of 'thp' memory space
overhead.

For almost every workload, 'ethp' constantly show about 10-15 MiB memory space
overhead, mainly due to its python wrapper I used for convenient test runs.
Using DAMON's raw interface would further remove this overhead.

In case of 'parsec3/swaptions' and 'splash2x/raytrace', 'ethp' shows even
higher memory space overhead. This is mainly due to the small size of the
workloads and the constant memory overhead of 'ethp', which came from the
python wrapper. The workloads consumes only about 14 MiB and 45 MiB each.
Because the constant memory consumption from the python wrapper of 'ethp'
(about 10-15 MiB) is relatively huge to the small working set, the relative
overhead becomes high. Nonetheless, such small workloads are not appropriate
target of the 'ethp' and the overhead can be removed by avoiding use of the
wrapper.


Runtime Overheads
~~~~~~~~~~~~~~~~~

Below shows measured runtime in similar way. The raw numbers are in seconds
and the overheads are in percent. Minus runtime overheads mean speedup.

runtime orig thp (overhead) ethp (overhead)
parsec3/blackscholes 107.003 106.468 ( -0.50) 107.260 ( 0.24)
parsec3/bodytrack 78.854 78.757 ( -0.12) 79.261 ( 0.52)
parsec3/canneal 137.520 120.854 (-12.12) 132.427 ( -3.70)
parsec3/dedup 11.873 11.665 ( -1.76) 11.883 ( 0.09)
parsec3/facesim 207.895 204.215 ( -1.77) 206.170 ( -0.83)
parsec3/ferret 190.507 189.972 ( -0.28) 190.818 ( 0.16)
parsec3/fluidanimate 211.064 208.862 ( -1.04) 211.874 ( 0.38)
parsec3/freqmine 290.157 288.831 ( -0.46) 292.495 ( 0.81)
parsec3/raytrace 118.460 118.741 ( 0.24) 119.808 ( 1.14)
parsec3/streamcluster 324.524 283.709 (-12.58) 307.209 ( -5.34)
parsec3/swaptions 154.458 154.894 ( 0.28) 155.307 ( 0.55)
parsec3/vips 58.588 58.622 ( 0.06) 59.037 ( 0.77)
parsec3/x264 66.493 66.604 ( 0.17) 67.051 ( 0.84)
splash2x/barnes 79.769 73.886 ( -7.38) 78.737 ( -1.29)
splash2x/fft 32.857 22.960 (-30.12) 25.808 (-21.45)
splash2x/lu_cb 85.113 84.939 ( -0.20) 85.344 ( 0.27)
splash2x/lu_ncb 92.408 90.103 ( -2.49) 93.585 ( 1.27)
splash2x/ocean_cp 44.374 42.876 ( -3.37) 43.613 ( -1.71)
splash2x/ocean_ncp 80.710 51.831 (-35.78) 71.498 (-11.41)
splash2x/radiosity 90.626 90.398 ( -0.25) 91.238 ( 0.68)
splash2x/radix 30.875 25.226 (-18.30) 25.882 (-16.17)
splash2x/raytrace 84.114 82.602 ( -1.80) 85.124 ( 1.20)
splash2x/volrend 86.796 86.347 ( -0.52) 88.223 ( 1.64)
splash2x/water_nsquared 230.781 220.667 ( -4.38) 232.664 ( 0.82)
splash2x/water_spatial 88.719 90.187 ( 1.65) 89.228 ( 0.57)
total 2984.530 2854.220 ( -4.37) 2951.540 ( -1.11)

In total, 'thp' shows 4.37% speedup while 'ethp' shows 1.11% speedup. In other
words, 'ethp' preserves about 25.40% of THP performance benefit.

In the best case (splash2x/raytrace), 'ethp' preserves 88.36% of the benefit.

If we narrow down to workloads showing high THP performance benefits
(splash2x/fft, splash2x/ocean_ncp, and splash2x/radix), 'thp' and 'ethp' shows
30.75% and 14.71% speedup in total, respectively. In other words, 'ethp'
preserves about 47.83% of the benefit.

Even in the worst case (splash2x/volrend), 'ethp' incurs only 1.64% runtime
overhead, which is similar to that of 'thp' (1.65% for
'splash2x/water_spatial').


Sequence Of Patches
===================

The patches are based on the v5.5 plus v5 DAMON patchset[1] and Minchan's
``madvise()`` factor-out patch[2]. Minchan's patch was necessary for reuse of
``madvise()`` code in DAMON. You can also clone the complete git tree:

$ git clone git://github.com/sjp38/linux -b damos/rfc/v4

The web is also available:
https://github.com/sjp38/linux/releases/tag/damos/rfc/v4


[1] https://lore.kernel.org/linux-mm/20200217103110.30817-1-sjpark@xxxxxxxxxx/
[2] https://lore.kernel.org/linux-mm/20200128001641.5086-2-minchan@xxxxxxxxxx/

The first patch allows DAMON to reuse ``madvise()`` code for the actions. The
second patch accounts age of each region. The third patch implements the
handling of the schemes in DAMON and exports a kernel space programming
interface for it. The fourth patch implements a debugfs interface for
privileged people and programs. The fifth and sixth patches each adds
kunittests and selftests for these changes, and finally the seventhe patch
modifies the user space tool for DAMON to support description and applying of
schemes in human freiendly way.


Patch History
=============

Changes from RFC v3
(https://lore.kernel.org/linux-mm/20200225102300.23895-1-sjpark@xxxxxxxxxx/)
- Add Reviewed-by from Brendan Higgins
- Code cleanup: Modularize madvise() call
- Fix a trivial bug in the wrapper python script
- Add more stable and detailed evaluation results with updated ETHP scheme

Changes from RFC v2
(https://lore.kernel.org/linux-mm/20200218085309.18346-1-sjpark@xxxxxxxxxx/)
- Fix aging mechanism for more better 'old region' selection
- Add more kunittests and kselftests for this patchset
- Support more human friedly description and application of 'schemes'

Changes from RFC v1
(https://lore.kernel.org/linux-mm/20200210150921.32482-1-sjpark@xxxxxxxxxx/)
- Properly adjust age accounting related properties after splitting, merging,
and action applying

SeongJae Park (7):
mm/madvise: Export madvise_common() to mm internal code
mm/damon: Account age of target regions
mm/damon: Implement data access monitoring-based operation schemes
mm/damon/schemes: Implement a debugfs interface
mm/damon-test: Add kunit test case for regions age accounting
mm/damon/selftests: Add 'schemes' debugfs tests
damon/tools: Support more human friendly 'schemes' control

include/linux/damon.h | 29 ++
mm/damon-test.h | 5 +
mm/damon.c | 391 +++++++++++++++++-
mm/internal.h | 4 +
mm/madvise.c | 3 +-
tools/damon/_convert_damos.py | 125 ++++++
tools/damon/_damon.py | 143 +++++++
tools/damon/damo | 7 +
tools/damon/record.py | 135 +-----
tools/damon/schemes.py | 105 +++++
.../testing/selftests/damon/debugfs_attrs.sh | 29 ++
11 files changed, 845 insertions(+), 131 deletions(-)
create mode 100755 tools/damon/_convert_damos.py
create mode 100644 tools/damon/_damon.py
create mode 100644 tools/damon/schemes.py

--
2.17.1

==================================== >8 =======================================

Appendix: Stdev / min / max numbers among the repeated runs
===========================================================

Below are stdev/min/max of each number in the 5 repeated runs.

runtime_stdev orig thp ethp
parsec3/blackscholes 0.884 0.932 0.693
parsec3/bodytrack 0.672 0.501 0.470
parsec3/canneal 3.434 1.278 4.112
parsec3/dedup 0.074 0.032 0.070
parsec3/facesim 1.079 0.572 0.688
parsec3/ferret 1.674 0.498 0.801
parsec3/fluidanimate 1.422 1.804 1.273
parsec3/freqmine 2.285 2.735 3.852
parsec3/raytrace 1.240 0.821 1.407
parsec3/streamcluster 2.226 2.221 2.778
parsec3/swaptions 1.760 2.164 1.650
parsec3/vips 0.071 0.113 0.433
parsec3/x264 4.972 4.732 5.464
splash2x/barnes 0.149 0.434 0.944
splash2x/fft 0.186 0.074 2.053
splash2x/lu_cb 0.358 0.674 0.054
splash2x/lu_ncb 0.694 0.586 0.301
splash2x/ocean_cp 0.214 0.181 0.163
splash2x/ocean_ncp 0.738 0.574 5.860
splash2x/radiosity 0.447 0.786 0.493
splash2x/radix 0.183 0.195 0.250
splash2x/raytrace 0.869 1.248 1.071
splash2x/volrend 0.896 0.801 0.759
splash2x/water_nsquared 3.050 3.032 1.750
splash2x/water_spatial 0.497 1.607 0.665


memused.avg_stdev orig thp ethp
parsec3/blackscholes 6837.158 4942.183 5531.310
parsec3/bodytrack 5591.783 5771.259 3959.415
parsec3/canneal 4034.250 5205.223 3294.782
parsec3/dedup 56582.594 12462.196 49390.950
parsec3/facesim 1879.070 3572.512 2407.374
parsec3/ferret 1686.811 4110.648 3050.263
parsec3/fluidanimate 5252.273 3550.694 3577.428
parsec3/freqmine 2634.481 12225.383 2220.963
parsec3/raytrace 5652.660 5615.677 4645.947
parsec3/streamcluster 2296.864 1906.081 2189.578
parsec3/swaptions 1100.155 18202.456 1689.923
parsec3/vips 5260.607 9104.494 2508.632
parsec3/x264 14892.433 18097.263 16853.532
splash2x/barnes 3055.563 2552.379 3749.773
splash2x/fft 115636.847 18058.645 193864.925
splash2x/lu_cb 2266.989 2495.620 9615.377
splash2x/lu_ncb 4816.990 3106.290 3406.873
splash2x/ocean_cp 5597.264 2189.592 40420.686
splash2x/ocean_ncp 6962.524 5038.039 352254.041
splash2x/radiosity 6151.433 1561.840 6976.647
splash2x/radix 12938.174 4141.470 64272.890
splash2x/raytrace 912.177 1473.169 1812.460
splash2x/volrend 1866.708 1527.107 1881.400
splash2x/water_nsquared 2126.581 4481.707 2471.129
splash2x/water_spatial 1495.886 3564.505 3182.864


runtime_min orig thp ethp
parsec3/blackscholes 106.073 105.724 106.799
parsec3/bodytrack 78.361 78.327 78.994
parsec3/canneal 130.735 118.456 125.902
parsec3/dedup 11.816 11.631 11.781
parsec3/facesim 206.358 203.462 205.526
parsec3/ferret 189.118 189.461 190.130
parsec3/fluidanimate 209.879 207.381 210.656
parsec3/freqmine 287.349 285.988 288.519
parsec3/raytrace 117.320 118.014 118.021
parsec3/streamcluster 322.404 280.907 304.489
parsec3/swaptions 153.017 153.133 154.307
parsec3/vips 58.480 58.518 58.496
parsec3/x264 61.569 61.987 62.333
splash2x/barnes 79.595 73.170 77.782
splash2x/fft 32.588 22.838 24.391
splash2x/lu_cb 84.897 84.229 85.300
splash2x/lu_ncb 91.640 89.480 93.192
splash2x/ocean_cp 44.216 42.661 43.403
splash2x/ocean_ncp 79.912 50.717 63.298
splash2x/radiosity 90.332 89.911 90.786
splash2x/radix 30.617 25.012 25.569
splash2x/raytrace 82.972 81.291 83.608
splash2x/volrend 86.205 85.414 86.772
splash2x/water_nsquared 228.749 216.488 230.019
splash2x/water_spatial 88.326 88.636 88.469


memused.avg_min orig thp ethp
parsec3/blackscholes 1809578.000 1815893.000 1821555.000
parsec3/bodytrack 1407270.000 1408774.000 1422950.000
parsec3/canneal 1037996.000 1029491.000 1042278.000
parsec3/dedup 2290578.000 2419128.000 2322004.000
parsec3/facesim 536908.000 539368.000 548194.000
parsec3/ferret 317173.000 313275.000 325452.000
parsec3/fluidanimate 566148.000 566925.000 578031.000
parsec3/freqmine 979565.000 985279.000 992844.000
parsec3/raytrace 1737270.000 1735498.000 1745751.000
parsec3/streamcluster 117213.000 118264.000 127825.000
parsec3/swaptions 13012.000 10753.000 21858.000
parsec3/vips 2946474.000 2941690.000 2975157.000
parsec3/x264 3171581.000 3170872.000 3184577.000
splash2x/barnes 1208476.000 1218535.000 1205510.000
splash2x/fft 9160132.000 9250818.000 8835513.000
splash2x/lu_cb 511850.000 515668.000 519205.000
splash2x/lu_ncb 512127.000 514471.000 518500.000
splash2x/ocean_cp 3342506.000 3377932.000 3290066.000
splash2x/ocean_ncp 3901749.000 7063386.000 3962171.000
splash2x/radiosity 1457419.000 1479232.000 1467156.000
splash2x/radix 1690840.000 1380921.000 1344838.000
splash2x/raytrace 43518.000 48571.000 55468.000
splash2x/volrend 147356.000 148650.000 159562.000
splash2x/water_nsquared 43685.000 38495.000 54409.000
splash2x/water_spatial 665912.000 660742.000 669843.000


runtime_max orig thp ethp
parsec3/blackscholes 108.322 108.141 108.641
parsec3/bodytrack 80.166 79.687 80.200
parsec3/canneal 140.219 122.073 137.615
parsec3/dedup 12.014 11.723 12.000
parsec3/facesim 209.291 205.234 207.192
parsec3/ferret 193.589 190.830 192.235
parsec3/fluidanimate 213.730 212.390 213.867
parsec3/freqmine 293.634 292.283 299.323
parsec3/raytrace 120.096 120.346 121.437
parsec3/streamcluster 327.827 287.094 311.657
parsec3/swaptions 157.661 158.341 158.589
parsec3/vips 58.648 58.815 59.611
parsec3/x264 73.389 73.856 75.369
splash2x/barnes 79.975 74.413 80.244
splash2x/fft 33.168 23.043 29.852
splash2x/lu_cb 85.825 85.914 85.446
splash2x/lu_ncb 93.717 91.074 93.902
splash2x/ocean_cp 44.789 43.190 43.882
splash2x/ocean_ncp 81.981 52.296 80.782
splash2x/radiosity 91.509 91.966 92.180
splash2x/radix 31.130 25.546 26.299
splash2x/raytrace 85.347 84.163 86.881
splash2x/volrend 88.575 87.389 88.957
splash2x/water_nsquared 236.851 224.982 235.537
splash2x/water_spatial 89.689 92.978 90.276


memused.avg_max orig thp ethp
parsec3/blackscholes 1827350.000 1830922.000 1836584.000
parsec3/bodytrack 1423070.000 1422588.000 1434832.000
parsec3/canneal 1048155.000 1043151.000 1051713.000
parsec3/dedup 2446661.000 2452237.000 2459532.000
parsec3/facesim 542340.000 547457.000 554321.000
parsec3/ferret 321678.000 325083.000 333474.000
parsec3/fluidanimate 579067.000 576389.000 587029.000
parsec3/freqmine 986759.000 1018980.000 998800.000
parsec3/raytrace 1750980.000 1749291.000 1757761.000
parsec3/streamcluster 123761.000 122647.000 133602.000
parsec3/swaptions 16305.000 59605.000 26835.000
parsec3/vips 2961299.000 2964746.000 2982101.000
parsec3/x264 3209871.000 3219818.000 3230036.000
splash2x/barnes 1217047.000 1224832.000 1215995.000
splash2x/fft 9505048.000 9302095.000 9378025.000
splash2x/lu_cb 518393.000 522739.000 545540.000
splash2x/lu_ncb 526380.000 522996.000 528341.000
splash2x/ocean_cp 3358820.000 3384581.000 3383533.000
splash2x/ocean_ncp 3920669.000 7079011.000 4937246.000
splash2x/radiosity 1474991.000 1483739.000 1485635.000
splash2x/radix 1731625.000 1393183.000 1498907.000
splash2x/raytrace 46122.000 52292.000 61116.000
splash2x/volrend 152488.000 153180.000 164793.000
splash2x/water_nsquared 49449.000 50555.000 60859.000
splash2x/water_spatial 669943.000 669815.000 679012.000