node: improve nextTick performance #5092
Closed
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
I'm not sure what you mean. Commit 924cc6c didn't touch error handling, only how exception objects are decorated. |
|
@bnoordhuis if you run this code without 924cc6c it'll pass. But with, it'll fail. |
|
Hm, I'm sorry I'm wrong. I'm a bit puzzled as all tests passed until I rebased to the current master but I might have only run test-parallel instead of test. |
|
I speculate what you're seeing is the result of removing the |
|
Reading the error messages properly helps a lot! ;-) While testing I did not use the |
added 2 commits
February 5, 2016 00:35
Prevent deoptimization of process.nextTick by removing the try finally block. This is not necessary as the next tick queue will be reset anyway, no matter if the callback throws or not.
Using a predefined array size prevents resizing the array and is therefor faster.
| nextTickCallbackWithManyArgs(callback, args); | ||
| } | ||
| } | ||
| _combinedTickCallback(args, callback); |
There was a problem hiding this comment.
To remove a line in the call stack, think we could just move the if/else back into here? Open to why it should be kept in another function.
There was a problem hiding this comment.
For me I had two reasons to split it up:
- It's less duplicated code
- The functions are smaller and the compiler might optimize them faster but I did not check that
I do not have any strong feelings about that though
There was a problem hiding this comment.
oop. missed that it's called in multiple locations. sounds good.
|
CI: https://ci.nodejs.org/job/node-test-pull-request/1555/ (on lock down ATM, will report back when complete) |
|
CI is all green except for a couple of flaky tests on ARM. |
|
LGTM @bnoordhuis Have an opinion on the change? |
|
LGTM |
1 similar comment
|
LGTM |
trevnorris
pushed a commit
that referenced
this pull request
Feb 9, 2016
Prevent deoptimization of process.nextTick by removing the try finally block. This is not necessary as the next tick queue will be reset anyway, no matter if the callback throws or not. Use a predefined array size prevents resizing the array and is therefor faster. PR-URL: #5092 Reviewed-By: Trevor Norris <[email protected]> Reviewed-By: Ben Noordhuis <[email protected]> Reviewed-By: James M Snell <[email protected]>
|
Thanks much. Landed in 8830797. |
rvagg
pushed a commit
that referenced
this pull request
Feb 10, 2016
Prevent deoptimization of process.nextTick by removing the try finally block. This is not necessary as the next tick queue will be reset anyway, no matter if the callback throws or not. Use a predefined array size prevents resizing the array and is therefor faster. PR-URL: #5092 Reviewed-By: Trevor Norris <[email protected]> Reviewed-By: Ben Noordhuis <[email protected]> Reviewed-By: James M Snell <[email protected]>
|
Marking this for LTS but we should let it sit in master and v5 for a bit before landing |
|
I think I'm leaning -1 on this given that it changes the call stack. If you're relying on it for some reason (say in a debugging tool of some kind) then you'd be in trouble. I'd even be tempted to push for this as |
|
@rvagg ... I'm leaning the same way. We need to let these kinds of changes sit in stable for a while before we even consider them eligible for pulling back to LTS. |
|
This is a micro-performance optimization. If there's a question as to whether it should be backported, the answer is most likely no. |
|
I'm marking this as don't land for now. Please feel free to change |
rvagg
added a commit
that referenced
this pull request
Feb 21, 2016
* buffer:
- You can now supply an encoding argument when filling a
Buffer Buffer#fill(string[, start[, end]][, encoding]), supplying
an existing Buffer will also work with
Buffer#fill(buffer[, start[, end]]). See the API documentation for
details on how this works. (Trevor Norris) #4935
- Buffer#indexOf() no longer requires a byteOffset argument if you
also wish to specify an encoding:
Buffer#indexOf(val[, byteOffset][, encoding]).
(Trevor Norris) #4803
* child_process: spawn() and spawnSync() now support a 'shell' option
to allow for optional execution of the given command inside a shell.
If set to true, cmd.exe will be used on Windows and /bin/sh
elsewhere. A path to a custom shell can also be passed to override
these defaults. On Windows, this option allows .bat. and .cmd files
to be executed with spawn() and spawnSync(). (Colin Ihrig) #4598
* http_parser: Update to http-parser 2.6.2 to fix an unintentionally
strict limitation of allowable header characters.
(James M Snell) #5237
* dgram: socket.send() now supports accepts an array of Buffers or
Strings as the first argument. See the API docs for details on how
this works. (Matteo Collina) #4374
* http: Fix a bug where handling headers will mistakenly trigger an
'upgrade' event where the server is just advertising its protocols.
This bug can prevent HTTP clients from communicating with HTTP/2
enabled servers. (Fedor Indutny) #4337
* net: Added a listening Boolean property to net and http servers to
indicate whether the server is listening for connections.
(José Moreira) #4743
* node: The C++ node::MakeCallback() API is now reentrant and calling
it from inside another MakeCallback() call no longer causes the
nextTick queue or Promises microtask queue to be processed out of
order. (Trevor Norris) #4507
* tls: Add a new tlsSocket.getProtocol() method to get the negotiated
TLS protocol version of the current connection. (Brian White) #4995
* vm: Introduce new 'produceCachedData' and 'cachedData' options to
new vm.Script() to interact with V8's code cache. When a new
vm.Script object is created with the 'produceCachedData' set to true
a Buffer with V8's code cache data will be produced and stored in
cachedData property of the returned object. This data in turn may be
supplied back to another vm.Script() object with a 'cachedData'
option if the supplied source is the same. Successfully executing a
script from cached data can speed up instantiation time. See the API
docs for details. (Fedor Indutny) #4777
* performance: Improvements in:
- process.nextTick() (Ruben Bridgewater) #5092
- path module (Brian White) #5123
- querystring module (Brian White) #5012
- streams module when processing small chunks (Matteo Collina) #4354
rvagg
added a commit
that referenced
this pull request
Feb 21, 2016
* buffer:
- You can now supply an encoding argument when filling a
Buffer Buffer#fill(string[, start[, end]][, encoding]), supplying
an existing Buffer will also work with
Buffer#fill(buffer[, start[, end]]). See the API documentation for
details on how this works. (Trevor Norris) #4935
- Buffer#indexOf() no longer requires a byteOffset argument if you
also wish to specify an encoding:
Buffer#indexOf(val[, byteOffset][, encoding]).
(Trevor Norris) #4803
* child_process: spawn() and spawnSync() now support a 'shell' option
to allow for optional execution of the given command inside a shell.
If set to true, cmd.exe will be used on Windows and /bin/sh
elsewhere. A path to a custom shell can also be passed to override
these defaults. On Windows, this option allows .bat. and .cmd files
to be executed with spawn() and spawnSync(). (Colin Ihrig) #4598
* http_parser: Update to http-parser 2.6.2 to fix an unintentionally
strict limitation of allowable header characters.
(James M Snell) #5237
* dgram: socket.send() now supports accepts an array of Buffers or
Strings as the first argument. See the API docs for details on how
this works. (Matteo Collina) #4374
* http: Fix a bug where handling headers will mistakenly trigger an
'upgrade' event where the server is just advertising its protocols.
This bug can prevent HTTP clients from communicating with HTTP/2
enabled servers. (Fedor Indutny) #4337
* net: Added a listening Boolean property to net and http servers to
indicate whether the server is listening for connections.
(José Moreira) #4743
* node: The C++ node::MakeCallback() API is now reentrant and calling
it from inside another MakeCallback() call no longer causes the
nextTick queue or Promises microtask queue to be processed out of
order. (Trevor Norris) #4507
* tls: Add a new tlsSocket.getProtocol() method to get the negotiated
TLS protocol version of the current connection. (Brian White) #4995
* vm: Introduce new 'produceCachedData' and 'cachedData' options to
new vm.Script() to interact with V8's code cache. When a new
vm.Script object is created with the 'produceCachedData' set to true
a Buffer with V8's code cache data will be produced and stored in
cachedData property of the returned object. This data in turn may be
supplied back to another vm.Script() object with a 'cachedData'
option if the supplied source is the same. Successfully executing a
script from cached data can speed up instantiation time. See the API
docs for details. (Fedor Indutny) #4777
* performance: Improvements in:
- process.nextTick() (Ruben Bridgewater) #5092
- path module (Brian White) #5123
- querystring module (Brian White) #5012
- streams module when processing small chunks (Matteo Collina) #4354
rvagg
added a commit
that referenced
this pull request
Feb 23, 2016
* buffer:
- You can now supply an encoding argument when filling a
Buffer Buffer#fill(string[, start[, end]][, encoding]), supplying
an existing Buffer will also work with
Buffer#fill(buffer[, start[, end]]). See the API documentation for
details on how this works. (Trevor Norris) #4935
- Buffer#indexOf() no longer requires a byteOffset argument if you
also wish to specify an encoding:
Buffer#indexOf(val[, byteOffset][, encoding]).
(Trevor Norris) #4803
* child_process: spawn() and spawnSync() now support a 'shell' option
to allow for optional execution of the given command inside a shell.
If set to true, cmd.exe will be used on Windows and /bin/sh
elsewhere. A path to a custom shell can also be passed to override
these defaults. On Windows, this option allows .bat. and .cmd files
to be executed with spawn() and spawnSync(). (Colin Ihrig) #4598
* http_parser: Update to http-parser 2.6.2 to fix an unintentionally
strict limitation of allowable header characters.
(James M Snell) #5237
* dgram: socket.send() now supports accepts an array of Buffers or
Strings as the first argument. See the API docs for details on how
this works. (Matteo Collina) #4374
* http: Fix a bug where handling headers will mistakenly trigger an
'upgrade' event where the server is just advertising its protocols.
This bug can prevent HTTP clients from communicating with HTTP/2
enabled servers. (Fedor Indutny) #4337
* net: Added a listening Boolean property to net and http servers to
indicate whether the server is listening for connections.
(José Moreira) #4743
* node: The C++ node::MakeCallback() API is now reentrant and calling
it from inside another MakeCallback() call no longer causes the
nextTick queue or Promises microtask queue to be processed out of
order. (Trevor Norris) #4507
* tls: Add a new tlsSocket.getProtocol() method to get the negotiated
TLS protocol version of the current connection. (Brian White) #4995
* vm: Introduce new 'produceCachedData' and 'cachedData' options to
new vm.Script() to interact with V8's code cache. When a new
vm.Script object is created with the 'produceCachedData' set to true
a Buffer with V8's code cache data will be produced and stored in
cachedData property of the returned object. This data in turn may be
supplied back to another vm.Script() object with a 'cachedData'
option if the supplied source is the same. Successfully executing a
script from cached data can speed up instantiation time. See the API
docs for details. (Fedor Indutny) #4777
* performance: Improvements in:
- process.nextTick() (Ruben Bridgewater) #5092
- path module (Brian White) #5123
- querystring module (Brian White) #5012
- streams module when processing small chunks (Matteo Collina) #4354
PR-URL: #5295
rvagg
added a commit
that referenced
this pull request
Feb 24, 2016
* buffer:
- You can now supply an encoding argument when filling a
Buffer Buffer#fill(string[, start[, end]][, encoding]), supplying
an existing Buffer will also work with
Buffer#fill(buffer[, start[, end]]). See the API documentation for
details on how this works. (Trevor Norris) #4935
- Buffer#indexOf() no longer requires a byteOffset argument if you
also wish to specify an encoding:
Buffer#indexOf(val[, byteOffset][, encoding]).
(Trevor Norris) #4803
* child_process: spawn() and spawnSync() now support a 'shell' option
to allow for optional execution of the given command inside a shell.
If set to true, cmd.exe will be used on Windows and /bin/sh
elsewhere. A path to a custom shell can also be passed to override
these defaults. On Windows, this option allows .bat. and .cmd files
to be executed with spawn() and spawnSync(). (Colin Ihrig) #4598
* http_parser: Update to http-parser 2.6.2 to fix an unintentionally
strict limitation of allowable header characters.
(James M Snell) #5237
* dgram: socket.send() now supports accepts an array of Buffers or
Strings as the first argument. See the API docs for details on how
this works. (Matteo Collina) #4374
* http: Fix a bug where handling headers will mistakenly trigger an
'upgrade' event where the server is just advertising its protocols.
This bug can prevent HTTP clients from communicating with HTTP/2
enabled servers. (Fedor Indutny) #4337
* net: Added a listening Boolean property to net and http servers to
indicate whether the server is listening for connections.
(José Moreira) #4743
* node: The C++ node::MakeCallback() API is now reentrant and calling
it from inside another MakeCallback() call no longer causes the
nextTick queue or Promises microtask queue to be processed out of
order. (Trevor Norris) #4507
* tls: Add a new tlsSocket.getProtocol() method to get the negotiated
TLS protocol version of the current connection. (Brian White) #4995
* vm: Introduce new 'produceCachedData' and 'cachedData' options to
new vm.Script() to interact with V8's code cache. When a new
vm.Script object is created with the 'produceCachedData' set to true
a Buffer with V8's code cache data will be produced and stored in
cachedData property of the returned object. This data in turn may be
supplied back to another vm.Script() object with a 'cachedData'
option if the supplied source is the same. Successfully executing a
script from cached data can speed up instantiation time. See the API
docs for details. (Fedor Indutny) #4777
* performance: Improvements in:
- process.nextTick() (Ruben Bridgewater) #5092
- path module (Brian White) #5123
- querystring module (Brian White) #5012
- streams module when processing small chunks (Matteo Collina) #4354
PR-URL: #5295
ChALkeR
added
the
performance
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
The performance gain is between 5-10% looking at the already existing nextTick benchmarks.
@trevnorris pointed out that this is might be depending on #4507. I guess this could be handled independently though.
@bnoordhuis the new v8::Private api seems to change the error handling in C and after that commit the nextTick callback errors are not handled in C anymore. I'm not sure this was a intended change?