Previously we cache the time origin for the milestones in the user
land, and refresh it at pre-execution. As result the time origin
gets serialized into the snapshot and is therefore not deterministic.
Now we store it in the milestone array as an internal value and
reset the milestones at serialization time instead of
deserialization time. This improves the determinism of the snapshot.
Drive-by: remove the unused MarkMilestone() binding.
PR-URL: https://github.com/nodejs/node/pull/48708
Refs: https://github.com/nodejs/build/issues/3043
Reviewed-By: Yagiz Nizipli <yagiz@nizipli.com>
use the perf_hooks to trace the time spent by net.connect, dns.lookup,
dns.lookupService, dns.resolvexxx.
PR-URL: https://github.com/nodejs/node/pull/42390
Reviewed-By: Matteo Collina <matteo.collina@gmail.com>
Reviewed-By: Paolo Insogna <paolo@cowtech.it>
Reviewed-By: Mohammed Keyvanzadeh <mohammadkeyvanzade94@gmail.com>
Only call into hrtime if there's an observer
Also, fix up some previously missed changes from the original refactor
Signed-off-by: James M Snell <jasnell@gmail.com>
Refs: https://github.com/nodejs/node/issues/37937
Refs: https://github.com/nodejs/node/pull/37136
PR-URL: https://github.com/nodejs/node/pull/38110
Reviewed-By: Matteo Collina <matteo.collina@gmail.com>
Reviewed-By: Luigi Pinca <luigipinca@gmail.com>
Reviewed-By: Richard Lau <rlau@redhat.com>
Reviewed-By: Ruben Bridgewater <ruben@bridgewater.de>
Reviewed-By: Benjamin Gruenbaum <benjamingr@gmail.com>
* Update the user timing implementation to conform to
User Timing Level 3.
* Reimplement user timing and timerify with pure JavaScript
implementations
* Simplify the C++ implementation for gc and http2 perf
* Runtime deprecate additional perf entry properties
in favor of the standard detail argument
* Disable the `buffered` option on PerformanceObserver,
all entries are queued and dispatched on setImmediate.
Only entries with active observers are buffered.
* This does remove the user timing and timerify
trace events. Because the trace_events are still
considered experimental, those can be removed without
a deprecation cycle. They are removed to improve
performance and reduce complexity.
Old: `perf_hooks/usertiming.js n=100000: 92,378.01249733355`
New: perf_hooks/usertiming.js n=100000: 270,393.5280638482`
PR-URL: https://github.com/nodejs/node/pull/37136
Refs: https://github.com/nodejs/diagnostics/issues/464
Reviewed-By: Antoine du Hamel <duhamelantoine1995@gmail.com>
Reviewed-By: Matteo Collina <matteo.collina@gmail.com>
Reviewed-By: Michaël Zasso <targos@protonmail.com>
This commit suggest renaming AlaisedBufferInfo to AlaisedBufferIndex to
make the code more readable.
The main motivation for this change is that I personally think that
the following code could be a little clearer:
context->GetDataFromSnapshotOnce<V8T>(*info_).ToLocalChecked();
Even knowing that GetDataFromSnapshotOnce takes a size_t I had to double
check the type of info_ to make sure.
PR-URL: https://github.com/nodejs/node/pull/36339
Reviewed-By: Zeyu Yang <himself65@outlook.com>
Reviewed-By: Rich Trott <rtrott@gmail.com>
Alliased buffer header is not imported in the `env.h` file due to which
the language server is giving an error.
Fixed it via importing the header file for "aliased_buffer.h"
PR-URL: https://github.com/nodejs/node/pull/35436
Reviewed-By: Gireesh Punathil <gpunathi@in.ibm.com>
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Joyee Cheung <joyeec9h3@gmail.com>
Reviewed-By: Rich Trott <rtrott@gmail.com>
This includes the initial Environment (without running bootstrap
scripts) into the builtin snapshot
PR-URL: https://github.com/nodejs/node/pull/32984
Reviewed-By: Anna Henningsen <anna@addaleax.net>
Reviewed-By: Daniel Bevenius <daniel.bevenius@gmail.com>
Class names are written in UpperCamelCase. Otherwise, this looks like
it’s a variable, not a class name.
PR-URL: https://github.com/nodejs/node/pull/32539
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: James M Snell <jasnell@gmail.com>
Instead of allowing the callers to instantiate the template
with any numeric types (such as aliasing a Uint8Array to double[]),
predefine types that make sense and use those instead.
PR-URL: https://github.com/nodejs/node/pull/27334
Reviewed-By: Anna Henningsen <anna@addaleax.net>
Reviewed-By: James M Snell <jasnell@gmail.com>
This patch joins `per_process::prog_start_time` (a double)
and `performance::performance_node_start` (a uint64_t) into a
`per_process::node_start_time` (a uint64_t) which gets initialized
in `node::Start()`.
PR-URL: https://github.com/nodejs/node/pull/26016
Reviewed-By: Anna Henningsen <anna@addaleax.net>
PR-URL: https://github.com/nodejs/node/pull/25507
Reviewed-By: Daniel Bevenius <daniel.bevenius@gmail.com>
Reviewed-By: Ben Noordhuis <info@bnoordhuis.nl>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: Anna Henningsen <anna@addaleax.net>
Reviewed-By: James M Snell <jasnell@gmail.com>
Otherwise this is global state that may be subject to race
conditions e.g. when running `perf_hooks` inside of Worker threads.
Tracking the GC type is removed entirely since the variable was unused.
PR-URL: https://github.com/nodejs/node/pull/25053
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: James M Snell <jasnell@gmail.com>
We generally add these to all headers that are considered
internal to Node.
These aren’t distributed as part of the headers tarball,
so I think this does not have to be semver-major
(and we have been changing the APIs in these headers
freely anyway).
PR-URL: https://github.com/nodejs/node/pull/22514
Reviewed-By: James M Snell <jasnell@gmail.com>
While `perf_hooks` is still in experimental, remove less useful
bootstrap milestones.
PR-URL: https://github.com/nodejs/node/pull/21247
Reviewed-By: Anna Henningsen <anna@addaleax.net>
Reviewed-By: Tiancheng "Timothy" Gu <timothygu99@gmail.com>
Update performance_state to use AliasedBuffer and update usage sites.
PR-URL: https://github.com/nodejs/node/pull/18300
Reviewed-By: Anna Henningsen <anna@addaleax.net>
Reviewed-By: James M Snell <jasnell@gmail.com>
The `double` fields in `performance_state` could previously have
been aligned at 4-byte instead of 8-byte boundaries, which would
have made creating an Float64Array them as a array buffer view
for an ArrayBuffer extending over the entire struct an invalid
operation.
Ref: 67269fd7f3
Comments out related flaky failure
PR-URL: https://github.com/nodejs/node/pull/14996
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Refael Ackermann <refack@gmail.com>
An initial implementation of the Performance Timing API for Node.js.
This is the same Performance Timing API implemented by modern browsers
with a number of Node.js specific properties. The User Timing mark()
and measure() APIs are implemented, garbage collection timing, and
node startup milestone timing.
```js
const { performance } = require('perf_hooks');
performance.mark('A');
setTimeout(() => {
performance.mark('B');
performance.measure('A to B', 'A', 'B');
const entry = performance.getEntriesByName('A to B', 'measure')[0];
console.log(entry.duration);
}, 10000);
```
The implementation is at the native layer and makes use of uv_hrtime().
This should enable *eventual* integration with things like Tracing
and Inspection.
The implementation is extensible and should allow us to add new
performance entry types as we go (e.g. for measuring i/o perf,
etc).
Documentation and a test are provided.
PR-URL: https://github.com/nodejs/node/pull/14680
Reviewed-By: Matteo Collina <matteo.collina@gmail.com>
