code review 6814103: runtime: use vDSO clock_gettime for time.now & runtime....

runtime: use vDSO clock_gettime for time.now & runtime.nanotime on Linux/amd64
Performance improvement aside, time.Now() now gets real nanosecond resolution
on supported systems.

Benchmark done on Core i7-2600 @ 3.40GHz with kernel 3.5.2-gentoo.
original vDSO gettimeofday:
BenchmarkNow    100000000               27.4 ns/op
new vDSO gettimeofday fallback:
BenchmarkNow    100000000               27.6 ns/op
new vDSO clock_gettime:
BenchmarkNow    100000000               24.4 ns/op

[minux1, 2012-11-08 10:47:23]

Hello golang-dev@googlegroups.com (cc: golang-dev@googlegroups.com),

I'd like you to review this change to
https://code.google.com/p/go

[minux1, 2012-11-08 10:54:49]

this CL lets runtime.nanotime call time.now.

if necessary, i can duplicate time.now code for runtime.nanotime to speed
the
latter up a little (as did before).

[minux1, 2012-11-08 13:23:07]

FYI, on linux/386, using clock_gettime actually slows time.now() down.

110 ns/op vs. 118 ns/op.

This huge performance difference is due to the lack of vDSO
clock_gettime and gettimeofday on linux/386.

[bradfitz, 2012-11-08 16:23:52]

My personal opinion is that a 8 ns loss on a platform where people don't
run servers is far outweighed by any speed-up or increase in functionality
(nanosecond resolution) elsewhere.

On Thu, Nov 8, 2012 at 7:23 AM, <minux.ma@gmail.com> wrote:

> FYI, on linux/386, using clock_gettime actually slows time.now() down.
>
> 110 ns/op vs. 118 ns/op.
>
> This huge performance difference is due to the lack of vDSO
> clock_gettime and gettimeofday on linux/386.
>
>
https://codereview.appspot.**com/6814103/<https://codereview.appspot.com/6814...
>

[iant, 2012-11-08 16:43:10]

LGTM

[minux1, 2012-11-08 16:50:22]

what do you think of using clock_gettime on all
unix systems where ns resolution is available?

should i add 386 code in this cl or in a subsequent cl?

[iant2, 2012-11-08 17:14:25]

On Thu, Nov 8, 2012 at 8:50 AM, minux <minux.ma@gmail.com> wrote:
> what do you think of using clock_gettime on all
> unix systems where ns resolution is available?

Sounds like a good idea to me, unless it is slower.

> should i add 386 code in this cl or in a subsequent cl?

Subsequent CL, I think.  This CL is pretty amd64 specific, and the 386
code will be quite different.

Ian

[minux1, 2012-11-09 06:19:19]

*** Submitted as http://code.google.com/p/go/source/detail?r=42c8d3aadc40 ***

runtime: use vDSO clock_gettime for time.now & runtime.nanotime on Linux/amd64
Performance improvement aside, time.Now() now gets real nanosecond resolution
on supported systems.

Benchmark done on Core i7-2600 @ 3.40GHz with kernel 3.5.2-gentoo.
original vDSO gettimeofday:
BenchmarkNow    100000000               27.4 ns/op
new vDSO gettimeofday fallback:
BenchmarkNow    100000000               27.6 ns/op
new vDSO clock_gettime:
BenchmarkNow    100000000               24.4 ns/op

R=golang-dev, bradfitz, iant, iant
CC=golang-dev
http://codereview.appspot.com/6814103