view tests/test-issue3084.t @ 51681:522b4d729e89

mmap: populate the mapping by default

Without pre-population, accessing all data through a mmap can result in many
pagefault, reducing performance significantly. If the mmap is prepopulated, the
performance can no longer get slower than a full read.
(See benchmark number below)

In some cases were very few data is read, prepopulating can be overkill and
slower than populating on access (through page fault). So that behavior can be
controlled when the caller can pre-determine the best behavior.
(See benchmark number below)

In addition, testing with populating in a secondary thread yield great result
combining the best of each approach. This might be implemented in later
changesets.

In all cases, using mmap has a great effect on memory usage when many processes
run in parallel on the same machine.

### Benchmarks

# What did I run

A couple of month back I ran a large benchmark campaign to assess the impact of
various approach for using mmap with the revlog (and other files), it
highlighted a few benchmarks that capture the impact of the changes well. So to
validate this change I checked the following:

- log command displaying various revisions
  (read the changelog index)
- log command displaying the patch of listed revisions
  (read the changelog index, the manifest index and a few files indexes)
- unbundling a few revisions
  (read and write changelog, manifest and few files indexes, and walk the graph
  to update some cache)
- pushing a few revisions
  (read and write changelog, manifest and few files indexes, walk the graph to
  update some cache, performs various accesses locally and remotely during
  discovery)

Benchmarks were run using the default module policy (c+py) and the rust one. No
significant difference were found between the two implementation, so we will
present result using the default policy (unless otherwise specified).

I ran them on a few repositories :

- mercurial: a "public changeset only" copy of mercurial from 2018-08-01 using
  zstd compression and sparse-revlog

- pypy: a copy of pypy from 2018-08-01 using zstd compression and sparse-revlog

- netbeans: a copy of netbeans from 2018-08-01 using zstd compression and
  sparse-revlog

- mozilla-try: a copy of mozilla-try from 2019-02-18 using zstd compression and
  sparse-revlog

- mozilla-try persistent-nodemap: Same as the above but with a persistent
  nodemap. Used for the log --patch benchmark only

# Results

For the smaller repositories (mercurial, pypy), the impact of mmap is almost
imperceptible, other cost dominating the operation. The impact of prepopulating
is undiscernible in the benchmark we ran.

For larger repositories the benchmark support explanation given above:

On netbeans, the log can be about 1% faster without repopulation (for a
difference < 100ms) but unbundle becomes a bit slower, even when small.

### data-env-vars.name                             = netbeans-2018-08-01-zstd-sparse-revlog
  # benchmark.name                                 = hg.command.unbundle
  # benchmark.variants.issue6528                   = disabled
  # benchmark.variants.reuse-external-delta-parent = yes
  # benchmark.variants.revs                        = any-1-extra-rev
  # benchmark.variants.source                      = unbundle
  # benchmark.variants.verbosity                   = quiet
with-populate: 0.240157
no-populate:   0.265087  (+10.38%, +0.02)
  # benchmark.variants.revs                        = any-100-extra-rev
with-populate: 1.459518
no-populate:   1.481290   (+1.49%, +0.02)
 ## benchmark.name                                 = hg.command.push
  # benchmark.variants.explicit-rev                = none
  # benchmark.variants.issue6528                   = disabled
  # benchmark.variants.protocol                    = ssh
  # benchmark.variants.reuse-external-delta-parent = yes
  # benchmark.variants.revs                        = any-1-extra-rev
with-populate: 0.771919
no-populate:   0.792025   (+2.60%, +0.02)
  # benchmark.variants.revs                        = any-100-extra-rev
with-populate: 1.459518
no-populate:   1.481290   (+1.49%, +0.02)


For mozilla-try, the "slow down" from pre-populate for small `hg log` is more
visible, but still small in absolute time. (using rust value for the persistent
nodemap value to be relevant).

### data-env-vars.name            = mozilla-try-2019-02-18-ds2-pnm
  # benchmark.name                = hg.command.log
  # bin-env-vars.hg.flavor        = rust
  # benchmark.variants.patch      = yes
  # benchmark.variants.limit-rev  = 1
with-populate: 0.237813
no-populate:   0.229452   (-3.52%, -0.01)
  # benchmark.variants.limit-rev  = 10
  # benchmark.variants.patch      = yes
with-populate: 1.213578
no-populate:   1.205189
### data-env-vars.name            = mozilla-try-2019-02-18-zstd-sparse-revlog
  # benchmark.variants.limit-rev  = 1000
  # benchmark.variants.patch      = no
  # benchmark.variants.rev        = tip
with-populate: 0.198607
no-populate:   0.195038   (-1.80%, -0.00)

However pre-populating provide a significant boost on more complex operations
like unbundle or push:

### data-env-vars.name                             = mozilla-try-2019-02-18-zstd-sparse-revlog
  # benchmark.name                                 = hg.command.push
  # benchmark.variants.explicit-rev                = none
  # benchmark.variants.issue6528                   = disabled
  # benchmark.variants.protocol                    = ssh
  # benchmark.variants.reuse-external-delta-parent = yes
  # benchmark.variants.revs                        = any-1-extra-rev
with-populate: 4.798632
no-populate:   4.953295   (+3.22%, +0.15)
  # benchmark.variants.revs                        = any-100-extra-rev
with-populate: 4.903618
no-populate:   5.014963   (+2.27%, +0.11)
 ## benchmark.name                                 = hg.command.unbundle
  # benchmark.variants.revs                        = any-1-extra-rev
with-populate: 1.423411
no-populate:   1.585365  (+11.38%, +0.16)
  # benchmark.variants.revs                        = any-100-extra-rev
with-populate: 1.537909
no-populate:   1.688489   (+9.79%, +0.15)

  $ echo "[extensions]" >> $HGRCPATH
  $ echo "largefiles =" >> $HGRCPATH

Create the repository outside $HOME since largefiles write to
$HOME/.cache/largefiles.

  $ hg init test
  $ cd test
  $ echo "root" > root
  $ hg add root
  $ hg commit -m "Root commit" --config extensions.largefiles=!

Ensure that .hg/largefiles isn't created before largefiles are added
#if unix-permissions
  $ chmod 555 .hg
#endif
  $ hg status
#if unix-permissions
  $ chmod 755 .hg
#endif

  $ test -f .hg/largefiles
  [1]

  $ echo "large" > foo
  $ hg add --large foo
  $ hg commit -m "Add foo as a largefile"

  $ hg update -r 0
  getting changed largefiles
  0 largefiles updated, 1 removed
  0 files updated, 0 files merged, 1 files removed, 0 files unresolved

  $ echo "normal" > foo
  $ hg add foo
  $ hg commit -m "Add foo as normal file"
  created new head

Normal file in the working copy, keeping the normal version:

  $ echo "n" | hg merge --config ui.interactive=Yes
  remote turned local normal file foo into a largefile
  use (l)argefile or keep (n)ormal file? n
  0 files updated, 0 files merged, 0 files removed, 0 files unresolved
  (branch merge, don't forget to commit)

  $ hg status
  $ cat foo
  normal

Normal file in the working copy, keeping the largefile version:

  $ hg update -q -C
  $ echo "l" | hg merge --config ui.interactive=Yes
  remote turned local normal file foo into a largefile
  use (l)argefile or keep (n)ormal file? l
  getting changed largefiles
  1 largefiles updated, 0 removed
  1 files updated, 0 files merged, 1 files removed, 0 files unresolved
  (branch merge, don't forget to commit)

  $ hg status
  M foo

  $ hg diff --nodates
  diff -r fa129ab6b5a7 .hglf/foo
  --- /dev/null
  +++ b/.hglf/foo
  @@ -0,0 +1,1 @@
  +7f7097b041ccf68cc5561e9600da4655d21c6d18
  diff -r fa129ab6b5a7 foo
  --- a/foo
  +++ /dev/null
  @@ -1,1 +0,0 @@
  -normal

  $ cat foo
  large

Largefile in the working copy, keeping the normal version:

  $ hg update -q -C -r 1
  $ echo "n" | hg merge --config ui.interactive=Yes
  remote turned local largefile foo into a normal file
  keep (l)argefile or use (n)ormal file? n
  getting changed largefiles
  0 largefiles updated, 0 removed
  1 files updated, 0 files merged, 1 files removed, 0 files unresolved
  (branch merge, don't forget to commit)

  $ hg status
  M foo

  $ hg diff --nodates
  diff -r ff521236428a .hglf/foo
  --- a/.hglf/foo
  +++ /dev/null
  @@ -1,1 +0,0 @@
  -7f7097b041ccf68cc5561e9600da4655d21c6d18
  diff -r ff521236428a foo
  --- /dev/null
  +++ b/foo
  @@ -0,0 +1,1 @@
  +normal

  $ cat foo
  normal

Largefile in the working copy, keeping the largefile version:

  $ hg update -q -C -r 1
  $ echo "l" | hg merge --config ui.interactive=Yes
  remote turned local largefile foo into a normal file
  keep (l)argefile or use (n)ormal file? l
  0 files updated, 0 files merged, 0 files removed, 0 files unresolved
  (branch merge, don't forget to commit)

  $ hg status

  $ cat foo
  large

Whatever ... commit something so we can invoke merge when updating

  $ hg commit -m '3: Merge'

Updating from largefile to normal - no reason to prompt

  $ hg up -r 2
  getting changed largefiles
  0 largefiles updated, 0 removed
  1 files updated, 0 files merged, 1 files removed, 0 files unresolved
  $ cat foo
  normal

(the update above used to leave the working dir in a very weird state - clean it
  $ hg up -qr null
  $ hg up -qr 2
)

Updating from normal to largefile - no reason to prompt

  $ hg up -r 3
  getting changed largefiles
  1 largefiles updated, 0 removed
  1 files updated, 0 files merged, 1 files removed, 0 files unresolved
  $ cat foo
  large

  $ cd ..


Systematic testing of merges involving largefiles:

Ancestor: normal  Parent: normal-id  Parent: large   result: large
Ancestor: normal  Parent: normal2    Parent: large   result: ?
Ancestor: large   Parent: large-id   Parent: normal  result: normal
Ancestor: large   Parent: large2     Parent: normal  result: ?

All cases should try merging both ways.

Prepare test repo:

  $ hg init merges
  $ cd merges

prepare cases with "normal" ancestor:

  $ hg up -qr null
  $ echo normal > f
  $ hg ci -Aqm "normal-ancestor"
  $ hg tag -l "normal-ancestor"
  $ touch f2
  $ hg ci -Aqm "normal-id"
  $ hg tag -l "normal-id"
  $ echo normal2 > f
  $ hg ci -m "normal2"
  $ hg tag -l "normal2"
  $ echo normal > f
  $ hg ci -Aqm "normal-same"
  $ hg tag -l "normal-same"
  $ hg up -qr "normal-ancestor"
  $ hg rm f
  $ echo large > f
  $ hg add --large f
  $ hg ci -qm "large"
  $ hg tag -l "large"

prepare cases with "large" ancestor:

  $ hg up -qr null
  $ echo large > f
  $ hg add --large f
  $ hg ci -qm "large-ancestor"
  $ hg tag -l "large-ancestor"
  $ touch f2
  $ hg ci -Aqm "large-id"
  $ hg tag -l "large-id"
  $ echo large2 > f
  $ hg ci -m "large2"
  $ hg tag -l "large2"
  $ echo large > f
  $ hg ci -Aqm "large-same"
  $ hg tag -l "large-same"
  $ hg up -qr "large-ancestor"
  $ hg rm f
  $ echo normal > f
  $ hg ci -qAm "normal"
  $ hg tag -l "normal"

  $ hg log -GT '{tags}'
  @  normal tip
  |
  | o  large-same
  | |
  | o  large2
  | |
  | o  large-id
  |/
  o  large-ancestor
  
  o  large
  |
  | o  normal-same
  | |
  | o  normal2
  | |
  | o  normal-id
  |/
  o  normal-ancestor
  


Ancestor: normal  Parent: normal-id  Parent: large   result: large

  $ hg up -Cqr normal-id
  $ hg merge -r large
  getting changed largefiles
  1 largefiles updated, 0 removed
  1 files updated, 0 files merged, 1 files removed, 0 files unresolved
  (branch merge, don't forget to commit)
  $ cat f
  large

swap

  $ hg up -Cqr large
  $ hg merge -r normal-id
  1 files updated, 0 files merged, 0 files removed, 0 files unresolved
  (branch merge, don't forget to commit)
  $ cat f
  large

Ancestor: normal  Parent: normal-same  Parent: large   result: large

  $ hg up -Cqr normal-same
  $ hg merge -r large
  getting changed largefiles
  1 largefiles updated, 0 removed
  1 files updated, 0 files merged, 1 files removed, 0 files unresolved
  (branch merge, don't forget to commit)
  $ cat f
  large

swap

  $ hg up -Cqr large
  $ hg merge -r normal-same
  1 files updated, 0 files merged, 0 files removed, 0 files unresolved
  (branch merge, don't forget to commit)
  $ cat f
  large

Ancestor: normal  Parent: normal2  Parent: large   result: ?
(annoying extra prompt ... but it do not do any serious harm)

  $ hg up -Cqr normal2
  $ hg merge -r large
  remote turned local normal file f into a largefile
  use (l)argefile or keep (n)ormal file? l
  getting changed largefiles
  1 largefiles updated, 0 removed
  1 files updated, 0 files merged, 1 files removed, 0 files unresolved
  (branch merge, don't forget to commit)
  $ cat f
  large

  $ hg up -Cqr normal2
  $ echo n | hg merge -r large --config ui.interactive=Yes
  remote turned local normal file f into a largefile
  use (l)argefile or keep (n)ormal file? n
  0 files updated, 0 files merged, 0 files removed, 0 files unresolved
  (branch merge, don't forget to commit)
  $ cat f
  normal2

swap

  $ hg up -Cqr large
  $ hg merge -r normal2
  remote turned local largefile f into a normal file
  keep (l)argefile or use (n)ormal file? l
  1 files updated, 0 files merged, 0 files removed, 0 files unresolved
  (branch merge, don't forget to commit)
  $ cat f
  large

  $ hg up -Cqr large
  $ echo n | hg merge -r normal2 --config ui.interactive=Yes
  remote turned local largefile f into a normal file
  keep (l)argefile or use (n)ormal file? n
  getting changed largefiles
  0 largefiles updated, 0 removed
  2 files updated, 0 files merged, 1 files removed, 0 files unresolved
  (branch merge, don't forget to commit)
  $ cat f
  normal2

Ancestor: large   Parent: large-id   Parent: normal  result: normal

  $ hg up -Cqr large-id
  $ hg merge -r normal
  getting changed largefiles
  0 largefiles updated, 0 removed
  1 files updated, 0 files merged, 1 files removed, 0 files unresolved
  (branch merge, don't forget to commit)
  $ cat f
  normal

swap

  $ hg up -Cqr normal
  $ hg merge -r large-id
  1 files updated, 0 files merged, 0 files removed, 0 files unresolved
  (branch merge, don't forget to commit)
  $ cat f
  normal

Ancestor: large   Parent: large-same   Parent: normal  result: normal

  $ hg up -Cqr large-same
  $ hg merge -r normal
  getting changed largefiles
  0 largefiles updated, 0 removed
  1 files updated, 0 files merged, 1 files removed, 0 files unresolved
  (branch merge, don't forget to commit)
  $ cat f
  normal

swap

  $ hg up -Cqr normal
  $ hg merge -r large-same
  1 files updated, 0 files merged, 0 files removed, 0 files unresolved
  (branch merge, don't forget to commit)
  $ cat f
  normal

Ancestor: large   Parent: large2   Parent: normal  result: ?
(annoying extra prompt ... but it do not do any serious harm)

  $ hg up -Cqr large2
  $ hg merge -r normal
  remote turned local largefile f into a normal file
  keep (l)argefile or use (n)ormal file? l
  0 files updated, 0 files merged, 0 files removed, 0 files unresolved
  (branch merge, don't forget to commit)
  $ cat f
  large2

  $ hg up -Cqr large2
  $ echo n | hg merge -r normal --config ui.interactive=Yes
  remote turned local largefile f into a normal file
  keep (l)argefile or use (n)ormal file? n
  getting changed largefiles
  0 largefiles updated, 0 removed
  1 files updated, 0 files merged, 1 files removed, 0 files unresolved
  (branch merge, don't forget to commit)
  $ cat f
  normal

swap

  $ hg up -Cqr normal
  $ hg merge -r large2
  remote turned local normal file f into a largefile
  use (l)argefile or keep (n)ormal file? l
  getting changed largefiles
  1 largefiles updated, 0 removed
  2 files updated, 0 files merged, 1 files removed, 0 files unresolved
  (branch merge, don't forget to commit)
  $ cat f
  large2

  $ hg up -Cqr normal
  $ echo n | hg merge -r large2 --config ui.interactive=Yes
  remote turned local normal file f into a largefile
  use (l)argefile or keep (n)ormal file? n
  1 files updated, 0 files merged, 0 files removed, 0 files unresolved
  (branch merge, don't forget to commit)
  $ cat f
  normal

  $ cd ..