AArch64: Backport memcpy improvements

Hello Michael, or anyone else affected,

Accepted glibc into focal-proposed. The package will build now and be available at https://launchpad.net/ubuntu/+source/glibc/2.31-0ubuntu9.4 in a few hours, and then in the -proposed repository.

Please help us by testing this new package. See https://wiki.ubuntu.com/Testing/EnableProposed for documentation on how to enable and use -proposed. Your feedback will aid us getting this update out to other Ubuntu users.

If this package fixes the bug for you, please add a comment to this bug, mentioning the version of the package you tested, what testing has been performed on the package and change the tag from verification-needed-focal to verification-done-focal. If it does not fix the bug for you, please add a comment stating that, and change the tag to verification-failed-focal. In either case, without details of your testing we will not be able to proceed.

Further information regarding the verification process can be found at https://wiki.ubuntu.com/QATeam/PerformingSRUVerification . Thank you in advance for helping!

N.B. The updated package will be released to -updates after the bug(s) fixed by this package have been verified and the package has been in -proposed for a minimum of 7 days.

All autopkgtests for the newly accepted glibc (2.31-0ubuntu9.4) for focal have finished running.
The following regressions have been reported in tests triggered by the package:

snapd-glib/1.58-0ubuntu0.20.04.0 (armhf)
apt/2.0.6 (armhf)
libmath-mpfr-perl/4.13-1 (armhf)
art-nextgen-simulation-tools/20160605+dfsg-4 (armhf)
ruby-nokogiri/1.10.7+dfsg1-2build1 (armhf)
r-cran-rgdal/1.4-8-1build2 (armhf)
arrayfire/3.3.2+dfsg1-4ubuntu4 (armhf)
libpango-perl/1.227-3build1 (armhf)
libimage-sane-perl/5-1 (s390x)
ruby-bootsnap/1.4.6-1 (arm64)
mle/1.4.3-1 (ppc64el, arm64)
libsyntax-keyword-try-perl/0.11-1build1 (armhf)
awesome/4.3-4 (armhf)
cysignals/1.10.2+ds-4 (arm64)
gvfs/1.44.1-1ubuntu1 (ppc64el)
libuv1/1.34.2-1ubuntu1.3 (amd64)
bali-phy/3.4.1+dfsg-2build1 (arm64, s390x)
g10k/0.5.7-1 (armhf)
litl/0.1.9-7 (amd64)
ruby-libxml/3.1.0-2 (armhf)
ffmpeg/7:4.2.4-1ubuntu0.1 (armhf)
yorick/2.2.04+dfsg1-10 (ppc64el, s390x)
linux-ibm/5.4.0-1010.11 (amd64)
liborcus/0.15.3-3build2 (armhf)
node-nodedbi/1.0.13+dfsg-1build1 (amd64)
r-bioc-delayedarray/0.12.2+dfsg-1 (armhf)
postgresql-unit/7.2-2 (armhf)
python-freecontact/1.1-5build2 (armhf)
r-cran-rwave/2.4-8-2 (armhf)
libproc-fastspawn-perl/1.2-1build2 (armhf)
linux-hwe-5.11/5.11.0-44.48~20.04.2 (armhf)
foo2zjs/20171202dfsg0-4 (armhf)
r-cran-erm/1.0-0-1 (armhf)
libsys-cpuload-perl/0.03-8build5 (armhf)
libhttp-parser-xs-perl/0.17-1build5 (armhf)
php-luasandbox/3.0.3-2build2 (armhf)
pynfft/1.3.2-3build1 (armhf)
r-cran-processx/3.4.2-1 (ppc64el)
r-bioc-multtest/2.42.0-1 (armhf)
linux-hwe-5.13/5.13.0-23.23~20.04.2 (armhf)
python-blosc/1.7.0+ds1-2ubuntu2 (armhf)
gyoto/1.4.4-3 (armhf)
r-cran-sem/3.1.9-2build1 (armhf)
libtext-reflow-perl/1.17-1build3 (armhf)
python3.9/3.9.5-3ubuntu0~20.04.1 (armhf)
r-cran-samr/3.0-1 (armhf)
r-cran-dplyr/0.8.4-1 (armhf)
python3.8/3.8.10-0ubuntu1~20.04.2 (armhf)
findent/3.1.1-1build1 (armhf)
mercurial/5.3.1-1ubuntu1 (amd64)
libmemcached-libmemcached-perl/1.001801+dfsg-2build4 (armhf)
openbabel/3.0.0+dfsg-3ubuntu3 (armhf)

Please visit the excuses page listed below and investigate the failures, proceeding afterwards as per the StableReleaseUpdates policy regarding autopkgtest regressions [1].

https://people.canonical.com/~ubuntu-archive/proposed-migration/focal/update_excuses.html#glibc

[1] https://wiki.ubuntu.com/StableReleaseUpdates#Autopkgtest_Regressions

Thank you!

I attempted to regression test this on a set of arm64 server SoCs, but:

$ gcc -g -O3 test_memcpy.c -o test_memcpy64
test_memcpy.c:6:10: fatal error: mm_malloc.h: No such file or directory
    6 | #include <mm_malloc.h>
      | ^~~~~~~~~~~~~
compilation terminated.

Here's a version that compiles on arm64. Do note it expects a command line argument, the size of the buffer to use in megabytes.

Bumping up the buffer sizes mentioned in [Test Case] from 32 -> 1024. In my testing, 32 results are inconsistent, perhaps due to caching effects. 1024 results look far more reliable.

Here's a set of regression tests I ran across various ARM Server SoCs:
https://docs.google.com/spreadsheets/d/1hSdI5XKgXXw2iKV1Ceab0w3kqaO8wcO_cwYKQn0cdwA/edit#gid=0

This does show a negative performance impact for all but 1, the worst one being just over 2% (Altra).

Hello Michael, or anyone else affected,

Accepted glibc into focal-proposed. The package will build now and be available at https://launchpad.net/ubuntu/+source/glibc/2.31-0ubuntu9.5 in a few hours, and then in the -proposed repository.

Please help us by testing this new package. See https://wiki.ubuntu.com/Testing/EnableProposed for documentation on how to enable and use -proposed. Your feedback will aid us getting this update out to other Ubuntu users.

If this package fixes the bug for you, please add a comment to this bug, mentioning the version of the package you tested, what testing has been performed on the package and change the tag from verification-needed-focal to verification-done-focal. If it does not fix the bug for you, please add a comment stating that, and change the tag to verification-failed-focal. In either case, without details of your testing we will not be able to proceed.

Further information regarding the verification process can be found at https://wiki.ubuntu.com/QATeam/PerformingSRUVerification . Thank you in advance for helping!

N.B. The updated package will be released to -updates after the bug(s) fixed by this package have been verified and the package has been in -proposed for a minimum of 7 days.

All autopkgtests for the newly accepted glibc (2.31-0ubuntu9.5) for focal have finished running.
The following regressions have been reported in tests triggered by the package:

garli/2.1-3build1 (armhf)
fpc/3.0.4+dfsg-23 (armhf)
lazarus/2.0.6+dfsg-3 (armhf)
libuv1/1.34.2-1ubuntu1.3 (i386)
ikiwiki-hosting/0.20180719-2 (armhf)
rtags/2.37-1 (amd64)
mercurial/5.3.1-1ubuntu1 (amd64, ppc64el)
foo2zjs/20171202dfsg0-4 (armhf)
frameworkintegration/5.68.0-0ubuntu1 (armhf)
hilive/2.0a-3build2 (arm64)
ruby-libxml/3.1.0-2 (s390x)
plasma-framework/5.68.0-0ubuntu1 (armhf)
feersum/1.407-2 (s390x)
r-bioc-delayedarray/0.12.2+dfsg-1 (armhf)
php-luasandbox/3.0.3-2build2 (ppc64el)
snapd-glib/1.58-0ubuntu0.20.04.0 (armhf)
bolt/0.8-4ubuntu1 (ppc64el)
threadweaver/5.68.0-0ubuntu1 (armhf)
python3.8/3.8.10-0ubuntu1~20.04.2 (amd64)

Please visit the excuses page listed below and investigate the failures, proceeding afterwards as per the StableReleaseUpdates policy regarding autopkgtest regressions [1].

https://people.canonical.com/~ubuntu-archive/proposed-migration/focal/update_excuses.html#glibc

[1] https://wiki.ubuntu.com/StableReleaseUpdates#Autopkgtest_Regressions

Thank you!

It was pointed out to me that my compare.py script printed the times the wrong way around, so I fixed that and also changed it to print output in MiB/s -- always easier to reason about a benchmark when "bigger is better"!

= ThunderX2 =
   length | before (MiB/s) | after (MiB/s) | delta
----------|----------------|----------------|----------
    32768 | 126.67 | 115.21 | -9.05%
    65536 | 267.91 | 244.92 | -8.58%
   131072 | 510.48 | 473.60 | -7.22%
   262144 | 985.45 | 928.21 | -5.81%
   524288 | 1893.52 | 1799.80 | -4.95%
  1048576 | 3755.69 | 3576.35 | -4.78%

= Altra =
   length | before (MiB/s) | after (MiB/s) | delta
----------|----------------|----------------|----------
    32768 | 271.34 | 287.24 | 5.86%
    65536 | 520.02 | 548.40 | 5.46%
   131072 | 998.58 | 1047.18 | 4.87%
   262144 | 1890.99 | 1970.43 | 4.20%
   524288 | 2571.28 | 2731.96 | 6.25%
  1048576 | 3873.37 | 4134.21 | 6.73%

= X-Gene =
   length | before (MiB/s) | after (MiB/s) | delta
----------|----------------|----------------|----------
    32768 | 122.15 | 125.95 | 3.11%
    65536 | 212.34 | 218.16 | 2.74%
   131072 | 321.07 | 317.36 | -1.15%
   262144 | 513.98 | 486.59 | -5.33%
   524288 | 935.09 | 866.97 | -7.28%
  1048576 | 1785.51 | 1647.05 | -7.75%

= Hi1620 =
   length | before (MiB/s) | after (MiB/s) | delta
----------|----------------|----------------|----------
    32768 | 211.63 | 213.69 | 0.97%
    65536 | 408.22 | 407.80 | -0.11%
   131072 | 753.09 | 764.44 | 1.51%
   262144 | 1396.94 | 1414.23 | 1.24%
   524288 | 2360.75 | 2375.61 | 0.63%
  1048576 | 4063.19 | 4085.62 | 0.55%

= ThunderX =
   length | before (MiB/s) | after (MiB/s) | delta
----------|----------------|----------------|----------
    32768 | 62.84 | 62.87 | 0.06%
    65536 | 96.08 | 96.10 | 0.02%
   131072 | 146.16 | 146.16 | -0.00%
   262144 | 261.62 | 261.64 | 0.01%
   524288 | 476.86 | 476.87 | 0.00%
  1048576 | 642.25 | 642.53 | 0.04%

= Hi1616 =
   length | before (MiB/s) | after (MiB/s) | delta
----------|----------------|----------------|----------
    32768 | 132.80 | 135.33 | 1.90%
    65536 | 245.60 | 259.11 | 5.50%
   131072 | 465.64 | 493.09 | 5.90%
   262144 | 917.01 | 967.84 | 5.54%
   524288 | 1607.14 | 1658.73 | 3.21%
  1048576 | 2898.84 | 2893.28 | -0.19%

= Raspberry Pi 4B (rev 1.1) 4GB =

   length | before (MiB/s) | after (MiB/s) | delta
----------|----------------|----------------|----------
    32768 | 83.88 | 84.21 | 0.39%
    65536 | 156.15 | 158.85 | 1.73%
   131072 | 292.38 | 298.58 | 2.12%
   262144 | 551.15 | 543.42 | -1.40%
   524288 | 606.33 | 599.74 | -1.09%
  1048576 | 651.02 | 654.38 | 0.52%

(will test some more models when I'm back home this afternoon)

Some more results:

= Raspberry Pi 3B 1GB =

   length | before (MiB/s) | after (MiB/s) | delta
----------|----------------|----------------|----------
    32768 | 48.24 | 46.19 | -4.26%
    65536 | 85.99 | 79.96 | -7.02%
   131072 | 154.00 | 139.68 | -9.30%
   262144 | 178.72 | 164.12 | -8.17%
   524288 | 163.56 | 156.55 | -4.28%
  1048576 | 246.15 | 234.32 | -4.81%

= Raspberry Pi 3A+ 512MB =

   length | before (MiB/s) | after (MiB/s) | delta
----------|----------------|----------------|----------
    32768 | 57.11 | 54.22 | -5.06%
    65536 | 101.16 | 94.53 | -6.56%
   131072 | 186.94 | 168.37 | -9.94%
   262144 | 200.16 | 181.37 | -9.39%
   524288 | 175.91 | 168.93 | -3.97%
  1048576 | 261.19 | 250.62 | -4.04%

= Raspberry Pi Zero 2 =

   length | before (MiB/s) | after (MiB/s) | delta
----------|----------------|----------------|----------
    32768 | 40.58 | 38.75 | -4.51%
    65536 | 72.51 | 67.57 | -6.81%
   131072 | 132.02 | 121.20 | -8.20%
   262144 | 165.26 | 149.13 | -9.76%
   524288 | 160.46 | 153.15 | -4.55%
  1048576 | 241.92 | 230.87 | -4.57%

Worth noting that the Pi 4 uses the 2711 SoC, while these (the 3B, 3A+, and Zero 2) all use the older 2837 SoC. In other words, while the new memcpy seems "okay" on the 2711, it's got "some" performance regression on the 2837.

Results for the 2.31-0ubuntu9.8~ppa2 build:

= X-Gene =
   length | before (MiB/s) | after (MiB/s) | delta
----------|----------------|----------------|----------
    32768 | 122.20 | 122.20 | 0.00%
    65536 | 215.42 | 213.49 | -0.89%
   131072 | 321.09 | 321.35 | 0.08%
   262144 | 513.74 | 514.24 | 0.10%
   524288 | 933.63 | 934.49 | 0.09%
  1048576 | 1782.39 | 1782.30 | -0.00%

= ThunderX =
   length | before (MiB/s) | after (MiB/s) | delta
----------|----------------|----------------|----------
    32768 | 64.26 | 64.32 | 0.09%
    65536 | 96.20 | 96.22 | 0.03%
   131072 | 146.15 | 146.19 | 0.03%
   262144 | 261.63 | 261.69 | 0.02%
   524288 | 476.88 | 476.90 | 0.00%
  1048576 | 642.50 | 642.45 | -0.01%

= ThunderX2 =
   length | before (MiB/s) | after (MiB/s) | delta
----------|----------------|----------------|----------
    32768 | 126.34 | 127.45 | 0.88%
    65536 | 267.46 | 267.44 | -0.01%
   131072 | 510.13 | 508.72 | -0.28%
   262144 | 985.29 | 982.52 | -0.28%
   524288 | 1898.92 | 1892.56 | -0.34%
  1048576 | 3758.78 | 3737.30 | -0.57%

= Hi1616 =
   length | before (MiB/s) | after (MiB/s) | delta
----------|----------------|----------------|----------
    32768 | 134.48 | 134.65 | 0.13%
    65536 | 249.25 | 250.88 | 0.66%
   131072 | 471.34 | 473.74 | 0.51%
   262144 | 932.67 | 930.73 | -0.21%
   524288 | 1616.07 | 1611.49 | -0.28%
  1048576 | 2915.93 | 2928.78 | 0.44

= Hi1620 =
   length | before (MiB/s) | after (MiB/s) | delta
----------|----------------|----------------|----------
    32768 | 211.49 | 212.56 | 0.50%
    65536 | 407.57 | 408.56 | 0.24%
   131072 | 753.83 | 755.84 | 0.27%
   262144 | 1401.03 | 1399.78 | -0.09%
   524288 | 2371.48 | 2364.44 | -0.30%
  1048576 | 4075.81 | 4041.50 | -0.84%

= Altra =
   length | before (MiB/s) | after (MiB/s) | delta
----------|----------------|----------------|----------
    32768 | 270.18 | 270.38 | 0.08%
    65536 | 520.13 | 519.56 | -0.11%
   131072 | 990.88 | 1000.81 | 1.00%
   262144 | 1890.44 | 1914.64 | 1.28%
   524288 | 2711.94 | 2707.39 | -0.17%
  1048576 | 4051.21 | 4057.54 | 0.16%

Results for the 2.31-0ubuntu9.8~ppa3 build:

= X-Gene =
   length | before (MiB/s) | after (MiB/s) | delta
----------|----------------|----------------|----------
    32768 | 122.23 | 122.24 | 0.01%
    65536 | 215.55 | 215.23 | -0.15%
   131072 | 321.10 | 323.41 | 0.72%
   262144 | 514.52 | 514.71 | 0.04%
   524288 | 934.18 | 934.98 | 0.09%
  1048576 | 1783.41 | 1781.47 | -0.11%

= ThunderX =
   length | before (MiB/s) | after (MiB/s) | delta
----------|----------------|----------------|----------
    32768 | 122.23 | 122.24 | 0.01%
    65536 | 215.55 | 215.23 | -0.15%
   131072 | 321.10 | 323.41 | 0.72%
   262144 | 514.52 | 514.71 | 0.04%
   524288 | 934.18 | 934.98 | 0.09%
  1048576 | 1783.41 | 1781.47 | -0.11%

= ThunderX2 =
   length | before (MiB/s) | after (MiB/s) | delta
----------|----------------|----------------|----------
    32768 | 127.67 | 127.31 | -0.28%
    65536 | 267.54 | 267.83 | 0.11%
   131072 | 511.12 | 511.25 | 0.03%
   262144 | 984.24 | 984.64 | 0.04%
   524288 | 1894.43 | 1893.61 | -0.04%
  1048576 | 3750.56 | 3750.02 | -0.01%

= Hi1616 =
   length | before (MiB/s) | after (MiB/s) | delta
----------|----------------|----------------|----------
    32768 | 135.17 | 134.92 | -0.19%
    65536 | 251.17 | 249.38 | -0.71%
   131072 | 473.54 | 471.53 | -0.42%
   262144 | 932.73 | 929.82 | -0.31%
   524288 | 1586.08 | 1596.86 | 0.68%
  1048576 | 2837.49 | 2853.36 | 0.56%

= Hi1620 =
   length | before (MiB/s) | after (MiB/s) | delta
----------|----------------|----------------|----------
    32768 | 210.90 | 213.21 | 1.10%
    65536 | 407.92 | 408.19 | 0.07%
   131072 | 756.54 | 761.39 | 0.64%
   262144 | 1405.13 | 1420.10 | 1.07%
   524288 | 2376.22 | 2390.99 | 0.62%
  1048576 | 4094.42 | 4062.20 | -0.79%

= Altra =
   length | before (MiB/s) | after (MiB/s) | delta
----------|----------------|----------------|----------
    32768 | 270.03 | 286.89 | 6.25%
    65536 | 520.20 | 549.39 | 5.61%
   131072 | 999.18 | 1048.87 | 4.97%
   262144 | 1899.90 | 1931.89 | 1.68%
   524288 | 2694.23 | 2874.09 | 6.68%
  1048576 | 4102.38 | 4356.14 | 6.19%

On the performance regression on the Pi 0/3 models above, bear in mind that the models showing the regression are all the models with <=1GB of RAM. Assuming there's no regression on the armhf side of things (I haven't tested this, but I got the impression this was an arm64 only change?), we wouldn't be affecting users of an architecture which is, in several regards, preferable on the smaller platforms.

Dave, if you have time to test 2.31-0ubuntu9.8~ppa4 vs focal-update that would be interesting -- I think given Dann's results though they will be pretty neutral though so I wouldn't worry over much about it.

Hello Michael, or anyone else affected,

Accepted glibc into focal-proposed. The package will build now and be available at https://launchpad.net/ubuntu/+source/glibc/2.31-0ubuntu9.8 in a few hours, and then in the -proposed repository.

Please help us by testing this new package. See https://wiki.ubuntu.com/Testing/EnableProposed for documentation on how to enable and use -proposed. Your feedback will aid us getting this update out to other Ubuntu users.

If this package fixes the bug for you, please add a comment to this bug, mentioning the version of the package you tested, what testing has been performed on the package and change the tag from verification-needed-focal to verification-done-focal. If it does not fix the bug for you, please add a comment stating that, and change the tag to verification-failed-focal. In either case, without details of your testing we will not be able to proceed.

Further information regarding the verification process can be found at https://wiki.ubuntu.com/QATeam/PerformingSRUVerification . Thank you in advance for helping!

N.B. The updated package will be released to -updates after the bug(s) fixed by this package have been verified and the package has been in -proposed for a minimum of 7 days.

All autopkgtests for the newly accepted glibc (2.31-0ubuntu9.8) for focal have finished running.
The following regressions have been reported in tests triggered by the package:

ruby-stackprof/0.2.15-2 (arm64)
sphinxbase/0.8+5prealpha+1-8 (armhf)
r-cran-ps/1.3.2-2 (s390x)
linux-hwe-5.13/5.13.0-37.42~20.04.1 (armhf)
mercurial/5.3.1-1ubuntu1 (armhf, ppc64el)
linux-hwe-5.11/5.11.0-61.61 (armhf)
mbedtls/2.16.4-1ubuntu2 (s390x)
libreoffice/1:6.4.7-0ubuntu0.20.04.4 (armhf)
ruby-ferret/0.11.8.7-2 (amd64)
cross-toolchain-base/43ubuntu3.1 (ppc64el)

Please visit the excuses page listed below and investigate the failures, proceeding afterwards as per the StableReleaseUpdates policy regarding autopkgtest regressions [1].

https://people.canonical.com/~ubuntu-archive/proposed-migration/focal/update_excuses.html#glibc

[1] https://wiki.ubuntu.com/StableReleaseUpdates#Autopkgtest_Regressions

Thank you!

Hello Michael, or anyone else affected,

Accepted glibc into focal-proposed. The package will build now and be available at https://launchpad.net/ubuntu/+source/glibc/2.31-0ubuntu9.9 in a few hours, and then in the -proposed repository.

Please help us by testing this new package. See https://wiki.ubuntu.com/Testing/EnableProposed for documentation on how to enable and use -proposed. Your feedback will aid us getting this update out to other Ubuntu users.

If this package fixes the bug for you, please add a comment to this bug, mentioning the version of the package you tested, what testing has been performed on the package and change the tag from verification-needed-focal to verification-done-focal. If it does not fix the bug for you, please add a comment stating that, and change the tag to verification-failed-focal. In either case, without details of your testing we will not be able to proceed.

Further information regarding the verification process can be found at https://wiki.ubuntu.com/QATeam/PerformingSRUVerification . Thank you in advance for helping!

N.B. The updated package will be released to -updates after the bug(s) fixed by this package have been verified and the package has been in -proposed for a minimum of 7 days.

All autopkgtests for the newly accepted glibc (2.31-0ubuntu9.9) for focal have finished running.
The following regressions have been reported in tests triggered by the package:

hilive/2.0a-3build2 (arm64)
tomb/2.7+dfsg2-1 (arm64)
linux-hwe-5.13/5.13.0-40.45~20.04.1 (armhf)
bali-phy/3.4.1+dfsg-2build1 (s390x, arm64)
smalt/0.7.6-9 (ppc64el)
mariadb-10.3/1:10.3.34-0ubuntu0.20.04.1 (armhf)
feersum/1.407-2 (s390x)
kopanocore/8.7.0-7ubuntu1 (amd64)
r-cran-ps/1.3.2-2 (s390x, ppc64el)
libreoffice/1:6.4.7-0ubuntu0.20.04.4 (amd64)
imagemagick/8:6.9.10.23+dfsg-2.1ubuntu11.4 (armhf)
ruby-stackprof/0.2.15-2 (amd64)
gnome-photos/3.34.1-1 (ppc64el)
linux-azure-5.11/5.11.0-1029.32~20.04.2 (amd64)
linux-intel-5.13/5.13.0-1010.10 (amd64)
php-luasandbox/3.0.3-2build2 (armhf, arm64)
mbedtls/2.16.4-1ubuntu2 (amd64, ppc64el)
cross-toolchain-base/43ubuntu3.1 (ppc64el)
rtags/2.37-1 (amd64)
gvfs/1.44.1-1ubuntu1 (arm64, ppc64el)
linux-oem-5.14/5.14.0-1033.36 (amd64)
linux-azure-cvm/5.4.0-1076.79+cvm1 (amd64)
mercurial/5.3.1-1ubuntu1 (armhf)
r-cran-satellite/1.0.2-1build1 (armhf)
s3ql/3.3.2+dfsg-1ubuntu1 (armhf)
snapd/2.54.3+20.04.1ubuntu0.2 (s390x, arm64, amd64, ppc64el)
sphinxbase/0.8+5prealpha+1-8 (armhf)
gemma/0.98.1+dfsg-1 (armhf)

Please visit the excuses page listed below and investigate the failures, proceeding afterwards as per the StableReleaseUpdates policy regarding autopkgtest regressions [1].

https://people.canonical.com/~ubuntu-archive/proposed-migration/focal/update_excuses.html#glibc

[1] https://wiki.ubuntu.com/StableReleaseUpdates#Autopkgtest_Regressions

Thank you!

I ran the tests on a graviton2 instance and saw the improvement as expected/desired:

   length | before (MiB/s) | after (MiB/s) | delta
----------|----------------|----------------|----------
    32768 | 233.19 | 247.61 | 6.18%
    65536 | 443.44 | 468.68 | 5.69%
   131072 | 852.52 | 895.00 | 4.98%
   262144 | 1630.48 | 1704.48 | 4.54%
   524288 | 2480.18 | 2601.47 | 4.89%
  1048576 | 3900.73 | 4112.08 | 5.42%

I'll try to get some people to test some other arm64 systems before calling this verification-done though.

Ran the tests against -updates and -proposed on a Pi 4, Pi 3B+, and Pi Zero 2 with the following results:

= Raspberry Pi 4 =

   length | before (MiB/s) | after (MiB/s) | delta
----------|----------------|----------------|----------
    32768 | 83.58 | 83.03 | -0.66%
    65536 | 156.08 | 155.69 | -0.25%
   131072 | 292.88 | 291.93 | -0.33%
   262144 | 543.00 | 550.95 | 1.47%
   524288 | 596.54 | 612.14 | 2.62%
  1048576 | 640.72 | 643.76 | 0.47%

= Raspberry Pi 3B+ =

   length | before (MiB/s) | after (MiB/s) | delta
----------|----------------|----------------|----------
    32768 | 55.55 | 55.38 | -0.31%
    65536 | 98.39 | 98.36 | -0.03%
   131072 | 173.72 | 169.80 | -2.26%
   262144 | 165.05 | 161.24 | -2.31%
   524288 | 146.80 | 147.16 | 0.25%
  1048576 | 221.23 | 221.32 | 0.04%

= Raspberry Pi Zero 2 =

   length | before (MiB/s) | after (MiB/s) | delta
----------|----------------|----------------|----------
    32768 | 39.09 | 39.32 | 0.58%
    65536 | 70.09 | 70.33 | 0.34%
   131072 | 125.40 | 122.88 | -2.01%
   262144 | 133.79 | 133.76 | -0.02%
   524288 | 134.59 | 133.84 | -0.56%
  1048576 | 203.99 | 203.89 | -0.04%

Looking much better than before; some mild improvement (and not too much cost) on the Pi 4 (the only platform where arm64 typically makes sense in the Pi series), and only minor degradation in the rest.

2.31-0ubuntu9.7 vs 2.31-0ubuntu9.9

= X-Gene =
   length | before (MiB/s) | after (MiB/s) | delta
----------|----------------|----------------|----------
    32768 | 122.27 | 122.26 | -0.00%
    65536 | 216.55 | 215.81 | -0.34%
   131072 | 321.97 | 322.25 | 0.09%
   262144 | 515.08 | 515.29 | 0.04%
   524288 | 934.42 | 934.58 | 0.02%
  1048576 | 1781.58 | 1783.57 | 0.11%

= ThunderX =
   length | before (MiB/s) | after (MiB/s) | delta
----------|----------------|----------------|----------
    32768 | 56.52 | 56.57 | 0.09%
    65536 | 86.45 | 86.54 | 0.10%
   131072 | 131.51 | 131.58 | 0.06%
   262144 | 235.38 | 235.50 | 0.05%
   524288 | 428.95 | 429.21 | 0.06%
  1048576 | 578.01 | 578.32 | 0.05%

= ThunderX2 =
   length | before (MiB/s) | after (MiB/s) | delta
----------|----------------|----------------|----------
    32768 | 124.61 | 121.43 | -2.54%
    65536 | 267.12 | 267.78 | 0.25%
   131072 | 508.33 | 509.59 | 0.25%
   262144 | 976.22 | 986.22 | 1.02%
   524288 | 1876.75 | 1894.43 | 0.94%
  1048576 | 3723.27 | 3747.04 | 0.64%

= Hi1616 =
   length | before (MiB/s) | after (MiB/s) | delta
----------|----------------|----------------|----------
    32768 | 135.41 | 132.37 | -2.25%
    65536 | 251.27 | 248.20 | -1.22%
   131072 | 473.71 | 471.16 | -0.54%
   262144 | 938.66 | 934.16 | -0.48%
   524288 | 1627.75 | 1611.05 | -1.03%
  1048576 | 2919.80 | 2912.29 | -0.26%

= Hi1620 =
   length | before (MiB/s) | after (MiB/s) | delta
----------|----------------|----------------|----------
    32768 | 213.00 | 213.93 | 0.44%
    65536 | 407.69 | 408.29 | 0.15%
   131072 | 755.45 | 761.75 | 0.83%
   262144 | 1409.21 | 1418.65 | 0.67%
   524288 | 2364.40 | 2388.89 | 1.04%
  1048576 | 4024.87 | 4080.98 | 1.39%

= Altra =
   length | before (MiB/s) | after (MiB/s) | delta
----------|----------------|----------------|----------
    32768 | 268.34 | 285.36 | 6.34%
    65536 | 519.58 | 549.29 | 5.72%
   131072 | 999.22 | 1049.03 | 4.98%
   262144 | 1853.42 | 1942.76 | 4.82%
   524288 | 2654.58 | 2650.23 | -0.16%
  1048576 | 4030.24 | 3955.51 | -1.85%

The verification of the Stable Release Update for glibc has completed successfully and the package is now being released to -updates. Subsequently, the Ubuntu Stable Release Updates Team is being unsubscribed and will not receive messages about this bug report. In the event that you encounter a regression using the package from -updates please report a new bug using ubuntu-bug and tag the bug report regression-update so we can easily find any regressions.

This bug was fixed in the package glibc - 2.31-0ubuntu9.9

---------------
glibc (2.31-0ubuntu9.9) focal; urgency=medium

  * Disable testsuite on riscv64. It is failing maths tests intermittently in
    ways that cannot be a glibc regression and is disabled in later series
    anyway.

glibc (2.31-0ubuntu9.8) focal; urgency=medium

  * Update for 20.04. (LP: #1951033)

  [ Balint Reczey ]
  * Cherry-pick upstream patch to fix building with -moutline-atomics
  * Prevent rare deadlock in pthread_cond_signal (LP: #1899800)

  [ Matthias Klose ]
  * Revert: Use DH_COMPAT=8 for dh_strip to fix debug sections for valgrind.
    Enables debugging ld.so related issues. (LP: #1918035)
  * Don't strip ld.so on armhf. (LP: #1927192)

  [ Gunnar Hjalmarsson ]
  * d/local/usr_sbin/update-locale: improve sanity checks. (LP: #1892825)

  [ Heitor Alves de Siqueira ]
  * d/p/u/git-lp1928508-reversing-calculation-of-__x86_shared_non_temporal.patch:
    - Fix memcpy() performance regression on x86 AMD systems (LP: #1928508)

  [ Aurelien Jarno ]
  * debian/debhelper.in/libc.preinst: drop the check for kernel release
    > 255 now that glibc and preinstall script are fixed. (LP: #1962225)

  [ Michael Hudson-Doyle ]
  * libc6 on arm64 is now built with -moutline-atomics so libc6-lse can now be
    an empty package that is safe to remove. (LP: #1912652)
  * d/patches/u/aarch64-memcpy-improvements.patch: Backport memcpy
    improvements. (LP: #1951032)
  * Add test-float64x-yn to xfails on riscv64.

 -- Michael Hudson-Doyle <email address hidden> Thu, 07 Apr 2022 13:24:41 +1200