Conversation
RafaelGSS
left a comment
There was a problem hiding this comment.
Instead of creating a function inside globalThis, I'd add all the context into the beforeClockTemplate. Also, I wouldn't perform memory translations during the collect process, instead, you can do it textReporter. Just check if bench.plugins includes this new plugin.
| getReport() { | ||
| process._rawDebug('grp',arguments); | ||
|
|
||
| } |
There was a problem hiding this comment.
This shouldn't be here, right?
There was a problem hiding this comment.
I wanted to brainstorm what to produce as the report - or is producing results enough? If we get it right, the report might be nice tho.
There was a problem hiding this comment.
This function is called by the reporters, and it should return a string in case you want to show it there. See on textReporter
|
|
||
| ### Class: `RecordMemorySpikePlugin` | ||
|
|
||
| A plugin to record memory allocation spikes in your benchmark's run. It should help you understand the speed vs memory tradeoffs you're making. |
There was a problem hiding this comment.
We should mention folks that are going to use it will need to pass --expose-gc
There was a problem hiding this comment.
I'd also mention a possible performance inconsistency when this plugin is enabled due to GC blocking op.
There was a problem hiding this comment.
I'm only calling gc in oncomplete, which I assumed is not part of the span that is being measured. is that incorrect?
Co-authored-by: Rafael Gonzaga <rafael.nunu@hotmail.com>
|
|
@naugtur would mind if I take over this PR? |
|
I wouldn't mind that. I'm going on vacation next week so might have some time to put into it again, but trat time might as well go into reviewing and testing out your changes. |
it took going on holiday again, but here it is
open questions:
__recordMemorySpikegoing to cause trouble in workers?