Patchwork [4,of,4] changelog: disable delta chains

# HG changeset patch
# User Gregory Szorc <gregory.szorc@gmail.com>
# Date 1474745226 25200
#      Sat Sep 24 12:27:06 2016 -0700
# Node ID db0073124b876cc8b8e1201ba1b867b56c8a0448
# Parent  712281c3ab4b6e6e8ae109fa0673d4d6321d82fd
changelog: disable delta chains

This patch disables delta chains on changelogs. After this patch, new
entries on changelogs - including existing changelogs - will be stored
as the fulltext of that data (likely compressed). No delta computation
will be performed.

An overview of delta chains and data justifying this change follows.

Revlogs try to store entries as a delta against a previous entry (either
a parent revision in the case of generaldelta or the previous physical
revision when not using generaldelta). Most of the time this is the
correct thing to do: it frequently results in less CPU usage and smaller
storage.

Delta chains are most effective when the base revision being deltad
against is similar to the current data. This tends to occur naturally
for manifests and file data, since only small parts of each tend to
change with each revision. Changelogs, however, are a different story.

Changelog entries represent changesets/commits. And unless commits in a
repository are homogonous (same author, changing same files, similar
commit messages, etc), a delta from one entry to the next tends to be
relatively large compared to the size of the entry. This means that
delta chains tend to be short. How short? Here is the full vs delta
revision breakdown on some real world repos:

Repo             % Full    % Delta   Max Length
hg                45.8       54.2        6
mozilla-central   42.4       57.6        8
mozilla-unified   42.5       57.5       17
pypy              46.1       53.9        6
python-zstandard  46.1       53.9        3

(I threw in python-zstandard as an example of a repo that is homogonous.
It contains a small Python project with changes all from the same
author.)

Contrast this with the manifest revlog for these repos, where 99+% of
revisions are deltas and delta chains run into the thousands.

So delta chains aren't as useful on changelogs. But even a short delta
chain may provide benefits. Let's measure that.

Delta chains may require less CPU to read revisions if the CPU time
spent reading smaller deltas is less than the CPU time used to
decompress larger individual entries. We can measure this via
`hg perfrevlog -c -d 1` to iterate a revlog to resolve each revision's
fulltext. Here are the results of that command on a repo using delta
chains in its changelog and on a repo without delta chains:

hg (forward)
! wall 0.407008 comb 0.410000 user 0.410000 sys 0.000000 (best of 25)
! wall 0.390061 comb 0.390000 user 0.390000 sys 0.000000 (best of 26)

hg (reverse)
! wall 0.515221 comb 0.520000 user 0.520000 sys 0.000000 (best of 19)
! wall 0.400018 comb 0.400000 user 0.390000 sys 0.010000 (best of 25)

mozilla-central (forward)
! wall 4.508296 comb 4.490000 user 4.490000 sys 0.000000 (best of 3)
! wall 4.370222 comb 4.370000 user 4.350000 sys 0.020000 (best of 3)

mozilla-central (reverse)
! wall 5.758995 comb 5.760000 user 5.720000 sys 0.040000 (best of 3)
! wall 4.346503 comb 4.340000 user 4.320000 sys 0.020000 (best of 3)

mozilla-unified (forward)
! wall 4.957088 comb 4.950000 user 4.940000 sys 0.010000 (best of 3)
! wall 4.660528 comb 4.650000 user 4.630000 sys 0.020000 (best of 3)

mozilla-unified (reverse)
! wall 6.119827 comb 6.110000 user 6.090000 sys 0.020000 (best of 3)
! wall 4.675136 comb 4.670000 user 4.670000 sys 0.000000 (best of 3)

pypy (forward)
! wall 1.231122 comb 1.240000 user 1.230000 sys 0.010000 (best of 8)
! wall 1.164896 comb 1.160000 user 1.160000 sys 0.000000 (best of 9)

pypy (reverse)
! wall 1.467049 comb 1.460000 user 1.460000 sys 0.000000 (best of 7)
! wall 1.160200 comb 1.170000 user 1.160000 sys 0.010000 (best of 9)

The data clearly shows that it takes less wall and CPU time to resolve
revisions when there are no delta chains in the changelogs, regardless
of the direction of traversal. Furthermore, not using a delta chain
means that fulltext resolution in reverse is as fast as iterating
forward. So not using delta chains on the changelog is a clear CPU win
for reading operations.

An example of a user-visible operation showing this speed-up is revset
evaluation. Here are results for
`hg perfrevset 'author(gps) or author(mpm)'`:

hg
! wall 1.655506 comb 1.660000 user 1.650000 sys 0.010000 (best of 6)
! wall 1.612723 comb 1.610000 user 1.600000 sys 0.010000 (best of 7)

mozilla-central
! wall 17.629826 comb 17.640000 user 17.600000 sys 0.040000 (best of 3)
! wall 17.311033 comb 17.300000 user 17.260000 sys 0.040000 (best of 3)

What about 00changelog.i size?

Repo                Delta Chains     No Delta Chains
hg                    7,033,250         6,976,771
mozilla-central      82,978,748        81,574,623
mozilla-unified      88,112,349        86,702,162
pypy                 20,740,699        20,659,741

The data shows that removing delta chains from the changelog makes the
changelog smaller.

Delta chains are also used during changegroup generation. This
operation essentially converts a series of revisions to one large
delta chain. And changegroup generation is smart: if the delta in
the revlog matches what the changegroup is emitting, it will reuse
the delta instead of recalculating it. We can measure the impact
removing changelog delta chains has on changegroup generation via
`hg perfchangegroupchangelog`:

hg
! wall 1.589245 comb 1.590000 user 1.590000 sys 0.000000 (best of 7)
! wall 1.788060 comb 1.790000 user 1.790000 sys 0.000000 (best of 6)

mozilla-central
! wall 17.382585 comb 17.380000 user 17.340000 sys 0.040000 (best of 3)
! wall 20.161357 comb 20.160000 user 20.120000 sys 0.040000 (best of 3)

mozilla-unified
! wall 18.722839 comb 18.720000 user 18.680000 sys 0.040000 (best of 3)
! wall 21.168075 comb 21.170000 user 21.130000 sys 0.040000 (best of 3)

pypy
! wall 4.828317 comb 4.830000 user 4.820000 sys 0.010000 (best of 3)
! wall 5.415455 comb 5.420000 user 5.410000 sys 0.010000 (best of 3)

The data shows eliminating delta chains makes the changelog part of
changegroup generation slower. This is expected since we now have to
compute deltas for revisions where we could recycle the delta before.

It is worth putting this regression into context of overall changegroup
times. Here is the rough total CPU time spent in changegroup generation
for various repos while using delta chains on the changelog:

Repo              CPU Time (s)    CPU Time w/ compression
hg                  4.50              7.05
mozilla-central   111.1             222.0
pypy               28.68             75.5

Before compression, removing delta chains from the changegroup adds
~4.4% overhead to hg changegroup generation, 1.3% to mozilla-central,
and 2.0% to pypy. When you factor in zlib compression, these percentages
are roughly divided by 2.

While the increased CPU usage for changegroup generation is unfortunate,
I think it is acceptable because the percentage is small, server
operators (those likely impacted most by this) have other mechanisms
to mitigate CPU consumption (namely reducing zlib compression level and
pre-generated clone bundles), and because there is room to optimize this
in the future. For example, we could introduce a new changegroup format
that allowed changelog entries to be represented differently so we don't
have to compute deltas.

Removing delta chains from the changelog also opens the door to future
optimizations.

I just mentioned the possibility of a new changegroup representation for
changelog data.

Storing changelog entries as independent entities also enables more
advanced storage options. We could refactor the changelog API to
distinguish between a supported data access API and the internal
mechanism for storing that data (much like the manifest API refactor
currently under way). This would make it easier to swap in alternative
storage backends for the changelog, such as SQLite or some such. This
in turn would make it easier to support features like shallow clone,
where changesets may arrive to local storage out of DAG order.

Another possibility is enabling dictionary compression on revlogs.
Currently, each entry is compressed in isolation. Now that we have a
compressed representation of each revision and each revision has similar
structure, we would likely see benefits by using dictionary compression.
We can't do this with zlib. But it does enable easier experimentation
with compression formats that do (fun fact: experimenting with zstd
led me down a path that resulted in this patch).

There is potentially room to add a revlog flag indicating whether not
to store delta chains. This could potentially lead to more read path
optimizations to skip delta chain computation. But I'm not convinced
it is worthwhile and worth wasting a revlog flag bit on.

On 09/24/2016 09:37 PM, Gregory Szorc wrote:
> # HG changeset patch
> # User Gregory Szorc <gregory.szorc@gmail.com>
> # Date 1474745226 25200
> #      Sat Sep 24 12:27:06 2016 -0700
> # Node ID db0073124b876cc8b8e1201ba1b867b56c8a0448
> # Parent  712281c3ab4b6e6e8ae109fa0673d4d6321d82fd
> changelog: disable delta chains

This patch seems overall surprising but great. I've some 
question/feedback regarding the impact on changegroup generation, see 
below for details.

> […]

> Delta chains are also used during changegroup generation. This
> operation essentially converts a series of revisions to one large
> delta chain. And changegroup generation is smart: if the delta in
> the revlog matches what the changegroup is emitting, it will reuse
> the delta instead of recalculating it. We can measure the impact
> removing changelog delta chains has on changegroup generation via
> `hg perfchangegroupchangelog`:

But changegroup version 2 and 3 are a bit smarter than just a plain 
deltra chain as they support general delta. Therefore we can have 
changegroup emit a non-delta chain for changelog (using 'nullid' as the 
delta base for all entry).

> hg
> ! wall 1.589245 comb 1.590000 user 1.590000 sys 0.000000 (best of 7)
> ! wall 1.788060 comb 1.790000 user 1.790000 sys 0.000000 (best of 6)
>
> mozilla-central
> ! wall 17.382585 comb 17.380000 user 17.340000 sys 0.040000 (best of 3)
> ! wall 20.161357 comb 20.160000 user 20.120000 sys 0.040000 (best of 3)
>
> mozilla-unified
> ! wall 18.722839 comb 18.720000 user 18.680000 sys 0.040000 (best of 3)
> ! wall 21.168075 comb 21.170000 user 21.130000 sys 0.040000 (best of 3)
>
> pypy
> ! wall 4.828317 comb 4.830000 user 4.820000 sys 0.010000 (best of 3)
> ! wall 5.415455 comb 5.420000 user 5.410000 sys 0.010000 (best of 3)
>
> The data shows eliminating delta chains makes the changelog part of
> changegroup generation slower. This is expected since we now have to
> compute deltas for revisions where we could recycle the delta before.
>
> It is worth putting this regression into context of overall changegroup
> times. Here is the rough total CPU time spent in changegroup generation
> for various repos while using delta chains on the changelog:
>
> Repo              CPU Time (s)    CPU Time w/ compression
> hg                  4.50              7.05
> mozilla-central   111.1             222.0
> pypy               28.68             75.5
>
> Before compression, removing delta chains from the changegroup adds
> ~4.4% overhead to hg changegroup generation, 1.3% to mozilla-central,
> and 2.0% to pypy. When you factor in zlib compression, these percentages
> are roughly divided by 2.

Even if the impact is small, it in the range of the performance gain we 
win on local operation, so still kind of relevant (otherwise that whole 
change would not be relevant). Can you experiment with a no-delta 
changelog chain the changegroup and tell us how it looks like ?

> While the increased CPU usage for changegroup generation is unfortunate,
> I think it is acceptable because the percentage is small, server
> operators (those likely impacted most by this) have other mechanisms
> to mitigate CPU consumption (namely reducing zlib compression level and
> pre-generated clone bundles), and because there is room to optimize this
> in the future. For example, we could introduce a new changegroup format
> that allowed changelog entries to be represented differently so we don't
> have to compute deltas.

(So as pointed above, I believe the current format already supports what 
you are looking for).

> Removing delta chains from the changelog also opens the door to future
> optimizations.
>
> I just mentioned the possibility of a new changegroup representation for
> changelog data.
>
> Storing changelog entries as independent entities also enables more
> advanced storage options. We could refactor the changelog API to
> distinguish between a supported data access API and the internal
> mechanism for storing that data (much like the manifest API refactor
> currently under way). This would make it easier to swap in alternative
> storage backends for the changelog, such as SQLite or some such. This
> in turn would make it easier to support features like shallow clone,
> where changesets may arrive to local storage out of DAG order.

I remember Durham presenting related plan during the latest sprint in 
Mozilla SF. I'm not too sure what the progress is as there is no wiki 
page about it but I assume we'll know more about it at the coming sprint.

That said, this part of the description seems a bit too disconnect to 
the actual change. I would advise for starting a mailing list thread 
about that after that path gets in.

> Another possibility is enabling dictionary compression on revlogs.
> Currently, each entry is compressed in isolation. Now that we have a
> compressed representation of each revision and each revision has similar
> structure, we would likely see benefits by using dictionary compression.
> We can't do this with zlib. But it does enable easier experimentation
> with compression formats that do (fun fact: experimenting with zstd
> led me down a path that resulted in this patch).
>
> There is potentially room to add a revlog flag indicating whether not
> to store delta chains. This could potentially lead to more read path
> optimizations to skip delta chain computation. But I'm not convinced
> it is worthwhile and worth wasting a revlog flag bit on.

Same feedback, "interesting but should probably be on the list instead 
of in that commit message.

Cheers,

On Tue, Sep 27, 2016 at 7:52 AM, Pierre-Yves David <
pierre-yves.david@ens-lyon.org> wrote:

>
>
> On 09/24/2016 09:37 PM, Gregory Szorc wrote:
>
>> # HG changeset patch
>> # User Gregory Szorc <gregory.szorc@gmail.com>
>> # Date 1474745226 25200
>> #      Sat Sep 24 12:27:06 2016 -0700
>> # Node ID db0073124b876cc8b8e1201ba1b867b56c8a0448
>> # Parent  712281c3ab4b6e6e8ae109fa0673d4d6321d82fd
>> changelog: disable delta chains
>>
>
> This patch seems overall surprising but great. I've some question/feedback
> regarding the impact on changegroup generation, see below for details.
>
> […]
>>
>
> Delta chains are also used during changegroup generation. This
>> operation essentially converts a series of revisions to one large
>> delta chain. And changegroup generation is smart: if the delta in
>> the revlog matches what the changegroup is emitting, it will reuse
>> the delta instead of recalculating it. We can measure the impact
>> removing changelog delta chains has on changegroup generation via
>> `hg perfchangegroupchangelog`:
>>
>
> But changegroup version 2 and 3 are a bit smarter than just a plain deltra
> chain as they support general delta. Therefore we can have changegroup emit
> a non-delta chain for changelog (using 'nullid' as the delta base for all
> entry).


Correct - we could do this as a followup. A mode where the server "streams"
revlog entries instead of calculating an infinite delta chain could result
in significant CPU reductions on the server (at the cost of more bandwidth).


>
>
> hg
>> ! wall 1.589245 comb 1.590000 user 1.590000 sys 0.000000 (best of 7)
>> ! wall 1.788060 comb 1.790000 user 1.790000 sys 0.000000 (best of 6)
>>
>> mozilla-central
>> ! wall 17.382585 comb 17.380000 user 17.340000 sys 0.040000 (best of 3)
>> ! wall 20.161357 comb 20.160000 user 20.120000 sys 0.040000 (best of 3)
>>
>> mozilla-unified
>> ! wall 18.722839 comb 18.720000 user 18.680000 sys 0.040000 (best of 3)
>> ! wall 21.168075 comb 21.170000 user 21.130000 sys 0.040000 (best of 3)
>>
>> pypy
>> ! wall 4.828317 comb 4.830000 user 4.820000 sys 0.010000 (best of 3)
>> ! wall 5.415455 comb 5.420000 user 5.410000 sys 0.010000 (best of 3)
>>
>> The data shows eliminating delta chains makes the changelog part of
>> changegroup generation slower. This is expected since we now have to
>> compute deltas for revisions where we could recycle the delta before.
>>
>> It is worth putting this regression into context of overall changegroup
>> times. Here is the rough total CPU time spent in changegroup generation
>> for various repos while using delta chains on the changelog:
>>
>> Repo              CPU Time (s)    CPU Time w/ compression
>> hg                  4.50              7.05
>> mozilla-central   111.1             222.0
>> pypy               28.68             75.5
>>
>> Before compression, removing delta chains from the changegroup adds
>> ~4.4% overhead to hg changegroup generation, 1.3% to mozilla-central,
>> and 2.0% to pypy. When you factor in zlib compression, these percentages
>> are roughly divided by 2.
>>
>
> Even if the impact is small, it in the range of the performance gain we
> win on local operation, so still kind of relevant (otherwise that whole
> change would not be relevant). Can you experiment with a no-delta changelog
> chain the changegroup and tell us how it looks like ?
>

Yes, I can do that. I would prefer it not block this landing as I think the
existing trade-off is worthwhile. Furthermore, I've contributed several
optimizations to bundle transfer already and it's worth noting that this
regression is extremely minor in the larger context of the gains we've
already seen and will continue to see.


>
> While the increased CPU usage for changegroup generation is unfortunate,
>> I think it is acceptable because the percentage is small, server
>> operators (those likely impacted most by this) have other mechanisms
>> to mitigate CPU consumption (namely reducing zlib compression level and
>> pre-generated clone bundles), and because there is room to optimize this
>> in the future. For example, we could introduce a new changegroup format
>> that allowed changelog entries to be represented differently so we don't
>> have to compute deltas.
>>
>
> (So as pointed above, I believe the current format already supports what
> you are looking for).
>
> Removing delta chains from the changelog also opens the door to future
>> optimizations.
>>
>> I just mentioned the possibility of a new changegroup representation for
>> changelog data.
>>
>> Storing changelog entries as independent entities also enables more
>> advanced storage options. We could refactor the changelog API to
>> distinguish between a supported data access API and the internal
>> mechanism for storing that data (much like the manifest API refactor
>> currently under way). This would make it easier to swap in alternative
>> storage backends for the changelog, such as SQLite or some such. This
>> in turn would make it easier to support features like shallow clone,
>> where changesets may arrive to local storage out of DAG order.
>>
>
> I remember Durham presenting related plan during the latest sprint in
> Mozilla SF. I'm not too sure what the progress is as there is no wiki page
> about it but I assume we'll know more about it at the coming sprint.
>
> That said, this part of the description seems a bit too disconnect to the
> actual change. I would advise for starting a mailing list thread about that
> after that path gets in.
>
> Another possibility is enabling dictionary compression on revlogs.
>> Currently, each entry is compressed in isolation. Now that we have a
>> compressed representation of each revision and each revision has similar
>> structure, we would likely see benefits by using dictionary compression.
>> We can't do this with zlib. But it does enable easier experimentation
>> with compression formats that do (fun fact: experimenting with zstd
>> led me down a path that resulted in this patch).
>>
>> There is potentially room to add a revlog flag indicating whether not
>> to store delta chains. This could potentially lead to more read path
>> optimizations to skip delta chain computation. But I'm not convinced
>> it is worthwhile and worth wasting a revlog flag bit on.
>>
>
> Same feedback, "interesting but should probably be on the list instead of
> in that commit message.
>
> Cheers,
>
> --
> Pierre-Yves David
>