buffer: Fix dataview-set benchmark.

[RReverser]

Checklist

• tests and code linting passes
• a test and/or benchmark is included
• the commit message follows commit guidelines

Affected core subsystem(s)

buffer

Description of change

Just fixes a benchmark code itself to provide proper measurement.

Improves numbers up to 4x by avoiding repetitive dynamic method lookup.

Before the change:

buffers\dataview-set.js type=Uint8 millions=1: 4.52066 buffers\dataview-set.js type=Uint16LE millions=1: 5.03784 buffers\dataview-set.js type=Uint16BE millions=1: 5.36833 buffers\dataview-set.js type=Uint32LE millions=1: 4.49819 buffers\dataview-set.js type=Uint32BE millions=1: 4.46667 buffers\dataview-set.js type=Int8 millions=1: 5.52648 buffers\dataview-set.js type=Int16LE millions=1: 5.54321 buffers\dataview-set.js type=Int16BE millions=1: 5.48971 buffers\dataview-set.js type=Int32LE millions=1: 5.61572 buffers\dataview-set.js type=Int32BE millions=1: 5.48682 buffers\dataview-set.js type=Float32LE millions=1: 5.43213 buffers\dataview-set.js type=Float32BE millions=1: 5.49020 buffers\dataview-set.js type=Float64LE millions=1: 5.41688 buffers\dataview-set.js type=Float64BE millions=1: 5.15072 

After the change:

buffers\dataview-set.js type=Uint8 millions=1: 19.16623 buffers\dataview-set.js type=Uint16LE millions=1: 17.47599 buffers\dataview-set.js type=Uint16BE millions=1: 17.82669 buffers\dataview-set.js type=Uint32LE millions=1: 11.06974 buffers\dataview-set.js type=Uint32BE millions=1: 11.06411 buffers\dataview-set.js type=Int8 millions=1: 19.36784 buffers\dataview-set.js type=Int16LE millions=1: 13.45155 buffers\dataview-set.js type=Int16BE millions=1: 11.97407 buffers\dataview-set.js type=Int32LE millions=1: 18.88772 buffers\dataview-set.js type=Int32BE millions=1: 18.31484 buffers\dataview-set.js type=Float32LE millions=1: 18.33773 buffers\dataview-set.js type=Float32BE millions=1: 15.50081 buffers\dataview-set.js type=Float64LE millions=1: 16.86952 buffers\dataview-set.js type=Float64BE millions=1: 17.21089 

(Just noticed this when going through #6893 and couldn't pass by.)

[addaleax]

const?

[RReverser]

Um... well, maybe, but in that case makes sense to change other places in this file as well. And, generally, unless already consistent with a file styling, I'm personally not a big fan of using block-scoped variables in benchmark code at least until V8 fixes their deopts (it does affect numbers quite a bit at the moment), but I can change if you really wish.

[addaleax]

If it affects the numbers (I know it can when it comes to let), feel free to leave it, but it’s become pretty standard around here to always use const when it makes sense… personally, I don’t care much, though. 😄

[RReverser]

Well, if changing only those to const and not touching var i -> let i, then it looks fine, within the limtis.

[RReverser]

Updated.

[addaleax]

Btw, if you want to refer to github issues/PRs/etc. in the commit message, you can do that, but it’s strongly preferred to use the full URL… I’d maybe drop that line anyway. :)

[RReverser]

Oh ok. Wasn't aware of that.

[addaleax]

LGTM

[Fishrock123]

cc @nodejs/buffer

[trevnorris]

nits: git message title, don't cap "Fix" and don't end with period. In git message body want to place URL's in Ref:. e.g.

Improves numbers up to 4x by avoiding repetitive dynamic method lookup. Ref: https://github.com/nodejs/node/pull/6893 

Though as @addaleax mentioned, probably best to just drop that line since the ref isn't directly applicable to this change.

Change itself LGTM

[RReverser]

Changed the commit message.

[trevnorris]

Great. LGTM.

[jasnell]

LGTM
@mscdex might have some thoughts tho.

[mscdex]

lint passes: https://ci.nodejs.org/job/node-test-linter/2672/

LGTM

[addaleax]

Landed in 4a56e89 … Thanks!

[MylesBorins]

@RReverser lts?

[RReverser]

@thealphanerd Can do. Does it require backporting PR?

[MylesBorins]

nope landed cleanly