node: improve performance of process.hrtime()

[evanlucas]

Move argument validation out of C++ and into JS. Improves performance
by about 15-20%.

Benchmark comparisons between current master and this branch are below:

$ node benchmark/compare.js ./node ./node_before -- misc bench-hrtime running ./node misc/bench-hrtime.js running ./node_before misc/bench-hrtime.js misc/bench-hrtime.js n=1000000: ./node: 13037000 ./node_before: 10677000 . 22.11% $ node benchmark/compare.js ./node ./node_before -- misc bench-hrtime running ./node misc/bench-hrtime.js running ./node_before misc/bench-hrtime.js misc/bench-hrtime.js n=1000000: ./node: 12954000 ./node_before: 10642000 . 21.73% $ node benchmark/compare.js ./node ./node_before -- misc bench-hrtime running ./node misc/bench-hrtime.js running ./node_before misc/bench-hrtime.js misc/bench-hrtime.js n=1000000: ./node: 13322000 ./node_before: 10770000 . 23.69% 

R= @trevnorris

[evanlucas]

CI: https://ci.nodejs.org/job/node-test-pull-request/1113/

[trevnorris]

good catch. LGTM.

[jasnell]

CI failures are unrelated. LGTM

[evanlucas]

Landed in 78fd435. Thanks!

[MylesBorins]

@evanlucas this is not merging into v4.x-staging without conflicts. Would you be able to take a look and perhaps send a new PR against that branch?

[evanlucas]

Sure I'll try do get that done tonight.

[evanlucas]

@thealphanerd this one depends on 36e8a2c, which doesn't appear to have landed in v4.x-staging, so we should probably avoid backporting this one unless that one gets backported as well?

[MylesBorins]

it looks like 36e8a2c is part of a series of six commits that were not landing cleanly. Work done by @trevnorris

Do you think that it is worth the effort to attempt to back port all of this?

[jasnell]

I would say let's hold off for now.
On Jan 14, 2016 4:21 AM, "Myles Borins" [email protected] wrote:

it looks like 36e8a2c
36e8a2c
is part of a series of six commits that were not landing cleanly. Work done
by @trevnorrishttps://github.com/trevnorris

Do you think that it is worth the effort to attempt to back port all of
this?

—
Reply to this email directly or view it on GitHub
#4484 (comment).

[MylesBorins]

As #3780 is now dont-land-on-v4.x I'm going to mark this one similarly