Rietveld Code Review Tool
Help | Bug tracker | Discussion group | Source code | Sign in
(1543)

Issue 8670044: code review 8670044: net: implement netpoll for windows (Closed)

Can't Edit
Can't Publish+Mail
Start Review
Created:
10 years, 11 months ago by brainman
Modified:
10 years, 8 months ago
Reviewers:
dvyukov
CC:
dvyukov, golang-dev
Visibility:
Public.

Description

net: implement netpoll for windows Moves the network poller from net package into runtime. benchmark old ns/op new ns/op delta BenchmarkTCP4OneShot 316386 287061 -9.27% BenchmarkTCP4OneShot-2 339822 313424 -7.77% BenchmarkTCP4OneShot-3 330057 306589 -7.11% BenchmarkTCP4OneShotTimeout 341775 287061 -16.01% BenchmarkTCP4OneShotTimeout-2 380835 295849 -22.32% BenchmarkTCP4OneShotTimeout-3 398412 328070 -17.66% BenchmarkTCP4Persistent 40622 33392 -17.80% BenchmarkTCP4Persistent-2 44528 35736 -19.74% BenchmarkTCP4Persistent-3 44919 36907 -17.84% BenchmarkTCP4PersistentTimeout 45309 33588 -25.87% BenchmarkTCP4PersistentTimeout-2 50289 38079 -24.28% BenchmarkTCP4PersistentTimeout-3 51559 37103 -28.04% BenchmarkTCP6OneShot 361305 345645 -4.33% BenchmarkTCP6OneShot-2 361305 331976 -8.12% BenchmarkTCP6OneShot-3 376929 347598 -7.78% BenchmarkTCP6OneShotTimeout 361305 322212 -10.82% BenchmarkTCP6OneShotTimeout-2 378882 333928 -11.86% BenchmarkTCP6OneShotTimeout-3 388647 335881 -13.58% BenchmarkTCP6Persistent 47653 35345 -25.83% BenchmarkTCP6Persistent-2 49215 35736 -27.39% BenchmarkTCP6Persistent-3 38474 37493 -2.55% BenchmarkTCP6PersistentTimeout 56637 34369 -39.32% BenchmarkTCP6PersistentTimeout-2 42575 38079 -10.56% BenchmarkTCP6PersistentTimeout-3 44137 37689 -14.61%

Patch Set 1 #

Patch Set 2 : diff -r 0c9561ddf631 https://go.googlecode.com/hg/ #

Total comments: 17

Patch Set 3 : diff -r fab6ba2a2d10 https://go.googlecode.com/hg/ #

Patch Set 4 : diff -r 2d4825868d95 https://go.googlecode.com/hg/ #

Patch Set 5 : diff -r f809e1e845c4 https://go.googlecode.com/hg/ #

Total comments: 14

Patch Set 6 : diff -r 657760a60531 https://go.googlecode.com/hg/ #

Total comments: 2

Patch Set 7 : diff -r 5b2b5f8d89b0 https://go.googlecode.com/hg/ #

Patch Set 8 : diff -r 8de6d4c58935 https://go.googlecode.com/hg/ #

Total comments: 21

Patch Set 9 : diff -r d3f99f091748 https://go.googlecode.com/hg/ #

Total comments: 4

Patch Set 10 : diff -r afb7b199f074 https://go.googlecode.com/hg/ #

Patch Set 11 : diff -r fccd815ed3bb https://go.googlecode.com/hg/ #

Patch Set 12 : diff -r 2bbca155a87f https://go.googlecode.com/hg/ #

Patch Set 13 : diff -r 5ee81a14cdfe https://go.googlecode.com/hg/ #

Unified diffs Side-by-side diffs Delta from patch set Stats (+228 lines, -172 lines) Patch
M src/pkg/net/fd_poll_runtime.go View 1 2 3 4 5 6 7 1 chunk +1 line, -1 line 0 comments Download
M src/pkg/net/fd_windows.go View 1 2 3 4 5 6 7 8 14 chunks +88 lines, -147 lines 0 comments Download
M src/pkg/net/sendfile_windows.go View 1 1 chunk +1 line, -1 line 0 comments Download
M src/pkg/net/sockopt_windows.go View 1 2 chunks +0 lines, -19 lines 0 comments Download
M src/pkg/runtime/defs_windows.go View 1 3 chunks +6 lines, -2 lines 0 comments Download
M src/pkg/runtime/defs_windows_386.h View 1 2 3 4 3 chunks +10 lines, -0 lines 0 comments Download
M src/pkg/runtime/defs_windows_amd64.h View 1 2 3 4 3 chunks +10 lines, -0 lines 0 comments Download
M src/pkg/runtime/netpoll.goc View 1 2 3 4 5 6 7 8 1 chunk +1 line, -1 line 0 comments Download
M src/pkg/runtime/netpoll_stub.c View 1 1 chunk +1 line, -1 line 0 comments Download
A src/pkg/runtime/netpoll_windows.c View 1 2 3 4 5 6 7 8 9 10 11 1 chunk +110 lines, -0 lines 0 comments Download

Messages

Total messages: 82
dvyukov
LGTM (but this is not for submission, right) We can also experiment with GetQueueCompletionStatusEx() and ...
10 years, 10 months ago (2013-05-19 15:15:48 UTC) #1
dvyukov
https://codereview.appspot.com/8670044/diff/2001/src/pkg/net/fd_windows.go File src/pkg/net/fd_windows.go (right): https://codereview.appspot.com/8670044/diff/2001/src/pkg/net/fd_windows.go#newcode216 src/pkg/net/fd_windows.go:216: o.fd.pd.WaitCanceled(int(o.mode)) wait. you need the result from WaitCanceled. at ...
10 years, 10 months ago (2013-05-19 15:20:31 UTC) #2
brainman
Hello dvyukov@google.com (cc: golang-dev@googlegroups.com), I'd like you to review this change to https://go.googlecode.com/hg/
10 years, 10 months ago (2013-05-20 02:12:23 UTC) #3
brainman
Thank you for review. Alex https://codereview.appspot.com/8670044/diff/2001/src/pkg/net/fd_poll_runtime.go File src/pkg/net/fd_poll_runtime.go (right): https://codereview.appspot.com/8670044/diff/2001/src/pkg/net/fd_poll_runtime.go#newcode107 src/pkg/net/fd_poll_runtime.go:107: println("unreachable: ", res) I ...
10 years, 10 months ago (2013-05-20 02:12:45 UTC) #4
dvyukov
https://codereview.appspot.com/8670044/diff/2001/src/pkg/net/fd_windows.go File src/pkg/net/fd_windows.go (right): https://codereview.appspot.com/8670044/diff/2001/src/pkg/net/fd_windows.go#newcode216 src/pkg/net/fd_windows.go:216: o.fd.pd.WaitCanceled(int(o.mode)) On 2013/05/20 02:12:45, brainman wrote: > I don't ...
10 years, 10 months ago (2013-05-20 04:02:53 UTC) #5
brainman
On 2013/05/20 04:02:53, dvyukov wrote: > > ... you still need to wait for overlapped ...
10 years, 10 months ago (2013-05-20 05:12:08 UTC) #6
dvyukov
On 2013/05/20 05:12:08, brainman wrote: > On 2013/05/20 04:02:53, dvyukov wrote: > > > > ...
10 years, 10 months ago (2013-05-20 06:50:05 UTC) #7
brainman
On 2013/05/20 06:50:05, dvyukov wrote: > > I am pretty sure it must work that ...
10 years, 10 months ago (2013-05-20 07:45:08 UTC) #8
dvyukov
On 2013/05/20 07:45:08, brainman wrote: > On 2013/05/20 06:50:05, dvyukov wrote: > > > > ...
10 years, 10 months ago (2013-05-20 10:09:55 UTC) #9
dvyukov
On 2013/05/20 10:09:55, dvyukov wrote: > On 2013/05/20 07:45:08, brainman wrote: > > On 2013/05/20 ...
10 years, 10 months ago (2013-05-20 10:11:32 UTC) #10
brainman
On 2013/05/20 10:11:32, dvyukov wrote: > > But how do you want to handle old ...
10 years, 10 months ago (2013-05-20 11:23:15 UTC) #11
brainman
On 2013/05/20 04:02:53, dvyukov wrote: > > ... Even if timeout has happened and you ...
10 years, 10 months ago (2013-05-21 04:24:20 UTC) #12
brainman
Hello dvyukov@google.com (cc: golang-dev@googlegroups.com), Please take another look.
10 years, 10 months ago (2013-05-21 04:24:40 UTC) #13
dvyukov
On Tue, May 21, 2013 at 8:24 AM, <alex.brainman@gmail.com> wrote: > On 2013/05/20 04:02:53, dvyukov ...
10 years, 10 months ago (2013-05-21 06:47:30 UTC) #14
brainman
On 2013/05/21 06:47:30, dvyukov wrote: > > Is there any chance you do 0-length read/write? ...
10 years, 10 months ago (2013-05-21 07:02:11 UTC) #15
dvyukov
On Tue, May 21, 2013 at 11:02 AM, <alex.brainman@gmail.com> wrote: > On 2013/05/21 06:47:30, dvyukov ...
10 years, 10 months ago (2013-05-21 07:26:18 UTC) #16
brainman
I have been busy lately. And it is more complicated then I expected. But I ...
10 years, 10 months ago (2013-05-27 10:35:39 UTC) #17
brainman
Dmitriy, The way I see things is this: - ready to read, so net issues ...
10 years, 10 months ago (2013-06-03 00:18:26 UTC) #18
dvyukov
On 2013/06/03 00:18:26, brainman wrote: > Dmitriy, > > The way I see things is ...
10 years, 10 months ago (2013-06-03 04:07:22 UTC) #19
brainman
Like that: diff -r 2d4825868d95 src/pkg/runtime/netpoll.goc --- a/src/pkg/runtime/netpoll.goc Mon May 20 15:23:45 2013 +1000 +++ ...
10 years, 10 months ago (2013-06-03 04:57:05 UTC) #20
dvyukov
On 2013/06/03 04:57:05, brainman wrote: > Like that: > > diff -r 2d4825868d95 src/pkg/runtime/netpoll.goc > ...
10 years, 10 months ago (2013-06-03 06:44:24 UTC) #21
brainman
> > netpollunblock() is executed under the descriptor lock. But, I think, it is possible ...
10 years, 10 months ago (2013-06-03 07:19:41 UTC) #22
dvyukov
On Mon, Jun 3, 2013 at 11:19 AM, <alex.brainman@gmail.com> wrote: > >> netpollunblock() is executed ...
10 years, 10 months ago (2013-06-03 08:17:15 UTC) #23
brainman
On 2013/06/03 08:17:15, dvyukov wrote: > > I do not understand the exact failure scenario. ...
10 years, 10 months ago (2013-06-03 09:26:06 UTC) #24
dvyukov
On 2013/06/03 09:26:06, brainman wrote: > On 2013/06/03 08:17:15, dvyukov wrote: > > > > ...
10 years, 10 months ago (2013-06-03 10:37:19 UTC) #25
brainman
> I see the problem. > The problem is that wait is preempted by IO ...
10 years, 10 months ago (2013-06-03 10:51:12 UTC) #26
brainman
On 2013/06/03 10:37:19, dvyukov wrote: > ... > I think you need to make ...
10 years, 10 months ago (2013-06-04 04:30:25 UTC) #27
brainman
Hello dvyukov@google.com (cc: golang-dev@googlegroups.com), Please take another look.
10 years, 10 months ago (2013-06-04 04:30:54 UTC) #28
dvyukov
On 2013/06/04 04:30:25, brainman wrote: > On 2013/06/03 10:37:19, dvyukov wrote: > > ... > ...
10 years, 10 months ago (2013-06-04 09:31:10 UTC) #29
dvyukov
https://codereview.appspot.com/8670044/diff/44001/src/pkg/net/fd_windows.go File src/pkg/net/fd_windows.go (right): https://codereview.appspot.com/8670044/diff/44001/src/pkg/net/fd_windows.go#newcode224 src/pkg/net/fd_windows.go:224: return int(o.qty), nil add a comment explaining that we've ...
10 years, 10 months ago (2013-06-04 09:47:33 UTC) #30
dvyukov
The high-level algorithm looks good. I actually like the way you changed netpoll.goc. I think ...
10 years, 10 months ago (2013-06-04 09:53:49 UTC) #31
dvyukov
https://codereview.appspot.com/8670044/diff/44001/src/pkg/runtime/netpoll_windows.c File src/pkg/runtime/netpoll_windows.c (right): https://codereview.appspot.com/8670044/diff/44001/src/pkg/runtime/netpoll_windows.c#newcode85 src/pkg/runtime/netpoll_windows.c:85: // TODO(brainman): Call GetQueuedCompletionStatusEx() here when possible. This is ...
10 years, 10 months ago (2013-06-04 09:56:42 UTC) #32
dvyukov
And later you may also look at removing memory allocation from Read/Write (by embedding some ...
10 years, 10 months ago (2013-06-04 10:08:10 UTC) #33
brainman
On 2013/06/04 09:56:42, dvyukov wrote: > This is unrelated to this change, but I am ...
10 years, 9 months ago (2013-06-05 07:37:15 UTC) #34
brainman
On 2013/06/04 10:08:10, dvyukov wrote: > And later you may also look at removing memory ...
10 years, 9 months ago (2013-06-05 07:39:22 UTC) #35
brainman
Hello dvyukov@google.com (cc: golang-dev@googlegroups.com), Please take another look.
10 years, 9 months ago (2013-06-06 05:37:22 UTC) #36
brainman
Thank you for review. Alex https://codereview.appspot.com/8670044/diff/44001/src/pkg/net/fd_windows.go File src/pkg/net/fd_windows.go (right): https://codereview.appspot.com/8670044/diff/44001/src/pkg/net/fd_windows.go#newcode224 src/pkg/net/fd_windows.go:224: return int(o.qty), nil On ...
10 years, 9 months ago (2013-06-06 05:37:56 UTC) #37
brainman
ping
10 years, 9 months ago (2013-06-20 23:04:55 UTC) #38
dvyukov
On 2013/06/20 23:04:55, brainman wrote: > ping I thought we are waiting for you: > ...
10 years, 9 months ago (2013-06-21 15:09:42 UTC) #39
brainman
On 2013/06/21 15:09:42, dvyukov wrote: > On 2013/06/20 23:04:55, brainman wrote: > > ping > ...
10 years, 9 months ago (2013-06-23 10:34:07 UTC) #40
dvyukov
On 2013/06/23 10:34:07, brainman wrote: > On 2013/06/21 15:09:42, dvyukov wrote: > > On 2013/06/20 ...
10 years, 9 months ago (2013-06-23 11:05:51 UTC) #41
dvyukov
https://codereview.appspot.com/8670044/diff/57001/src/pkg/runtime/netpoll.goc File src/pkg/runtime/netpoll.goc (right): https://codereview.appspot.com/8670044/diff/57001/src/pkg/runtime/netpoll.goc#newcode127 src/pkg/runtime/netpoll.goc:127: netpollblock(pd, mode); we expect that this returns only ioready, ...
10 years, 9 months ago (2013-06-23 11:09:43 UTC) #42
brainman
I also sent new CL https://codereview.appspot.com/10485043/ as per your request. Alex https://codereview.appspot.com/8670044/diff/57001/src/pkg/runtime/netpoll.goc File src/pkg/runtime/netpoll.goc (right): ...
10 years, 9 months ago (2013-06-24 04:37:27 UTC) #43
brainman
Hello dvyukov@google.com (cc: golang-dev@googlegroups.com), Please take another look.
10 years, 9 months ago (2013-06-24 04:37:54 UTC) #44
dvyukov
On 2013/06/24 04:37:54, brainman wrote: > Hello mailto:dvyukov@google.com (cc: mailto:golang-dev@googlegroups.com), > > Please take another ...
10 years, 9 months ago (2013-06-24 13:02:08 UTC) #45
brainman
Hello dvyukov@google.com (cc: golang-dev@googlegroups.com), Please take another look.
10 years, 9 months ago (2013-06-25 04:41:58 UTC) #46
brainman
Updated to the tip. Waiting for instructions. Alex
10 years, 9 months ago (2013-06-25 04:43:00 UTC) #47
dvyukov
https://codereview.appspot.com/8670044/diff/81001/src/pkg/net/fd_windows.go File src/pkg/net/fd_windows.go (right): https://codereview.appspot.com/8670044/diff/81001/src/pkg/net/fd_windows.go#newcode96 src/pkg/net/fd_windows.go:96: var i int guard the following lines by !canCancelIO, ...
10 years, 9 months ago (2013-06-25 12:00:28 UTC) #48
dvyukov
On 2013/06/25 04:43:00, brainman wrote: > Updated to the tip. Waiting for instructions. I think ...
10 years, 9 months ago (2013-06-25 12:01:37 UTC) #49
brainman
Hello dvyukov@google.com (cc: golang-dev@googlegroups.com), Please take another look.
10 years, 9 months ago (2013-06-26 07:17:56 UTC) #50
brainman
https://codereview.appspot.com/8670044/diff/81001/src/pkg/net/fd_windows.go File src/pkg/net/fd_windows.go (right): https://codereview.appspot.com/8670044/diff/81001/src/pkg/net/fd_windows.go#newcode96 src/pkg/net/fd_windows.go:96: var i int On 2013/06/25 12:00:32, dvyukov wrote: > ...
10 years, 9 months ago (2013-06-26 07:18:04 UTC) #51
dvyukov
LGTM I think you can submit now (after addressing the comments), it is on review ...
10 years, 9 months ago (2013-06-26 10:23:22 UTC) #52
dvyukov
On 2013/06/26 10:23:22, dvyukov wrote: > LGTM > > I think you can submit now ...
10 years, 9 months ago (2013-06-26 10:24:59 UTC) #53
brainman
On 2013/06/26 10:24:59, dvyukov wrote: > > I think you need to extend CL description ...
10 years, 9 months ago (2013-06-26 23:09:01 UTC) #54
brainman
On 2013/06/26 10:23:22, dvyukov wrote: > > Hope you've done stress testing with GOARCH=386/amd64 GOMAXPROCS=1/2/4/8. ...
10 years, 9 months ago (2013-06-27 06:52:18 UTC) #55
brainman
Hello dvyukov@google.com (cc: golang-dev@googlegroups.com), Please take another look.
10 years, 9 months ago (2013-06-27 06:53:15 UTC) #56
dvyukov
On Thu, Jun 27, 2013 at 10:52 AM, <alex.brainman@gmail.com> wrote: > On 2013/06/26 10:23:22, dvyukov ...
10 years, 9 months ago (2013-06-27 11:25:45 UTC) #57
dvyukov
On Thu, Jun 27, 2013 at 3:25 PM, Dmitry Vyukov <dvyukov@google.com> wrote: > On Thu, ...
10 years, 9 months ago (2013-06-27 11:27:19 UTC) #58
brainman
On 2013/06/27 11:25:45, dvyukov wrote: > > Is it started happening when you set max ...
10 years, 9 months ago (2013-06-27 11:36:11 UTC) #59
dvyukov
On Thu, Jun 27, 2013 at 3:26 PM, Dmitry Vyukov <dvyukov@google.com> wrote: > On Thu, ...
10 years, 9 months ago (2013-06-27 11:37:15 UTC) #60
brainman
On 2013/06/27 11:27:19, dvyukov wrote: > > What if you set max threads to 10000? ...
10 years, 9 months ago (2013-06-27 11:38:36 UTC) #61
dvyukov
On Thu, Jun 27, 2013 at 3:36 PM, <alex.brainman@gmail.com> wrote: > On 2013/06/27 11:25:45, dvyukov ...
10 years, 9 months ago (2013-06-27 11:39:26 UTC) #62
brainman
On 2013/06/27 11:37:15, dvyukov wrote: > > Does it happen deterministically or not? ... As ...
10 years, 9 months ago (2013-06-27 11:44:14 UTC) #63
brainman
On 2013/06/27 11:39:26, dvyukov wrote: > > I would not be surprised if kernel uses ...
10 years, 9 months ago (2013-06-27 11:45:44 UTC) #64
dvyukov
On 2013/06/27 11:45:44, brainman wrote: > On 2013/06/27 11:39:26, dvyukov wrote: > > > > ...
10 years, 9 months ago (2013-06-27 12:12:23 UTC) #65
brainman
On 2013/06/27 12:12:23, dvyukov wrote: > > Yeah, there is no good to add them, ...
10 years, 9 months ago (2013-06-27 12:14:48 UTC) #66
brainman
> ... TestVariousDeadlines4Proc timeout too? Yes. > ... set max threads to 10000, see if ...
10 years, 9 months ago (2013-06-28 02:43:12 UTC) #67
brainman
On 2013/06/28 02:43:12, brainman wrote: > > I have collected some benchmarks > https://code.google.com/p/go/issues/detail?id=5284#c4 (there ...
10 years, 9 months ago (2013-06-28 06:31:48 UTC) #68
dvyukov
On 2013/06/28 02:43:12, brainman wrote: > > ... TestVariousDeadlines4Proc timeout too? > > Yes. OK, ...
10 years, 9 months ago (2013-06-28 11:15:14 UTC) #69
dvyukov
Regarding the test failure, I don't have any good ideas now... You've said that if ...
10 years, 9 months ago (2013-06-28 11:22:28 UTC) #70
brainman
On 2013/06/28 11:15:14, dvyukov wrote: > I would include the following benchmarks: > BenchmarkTCP4OneShot 500 ...
10 years, 9 months ago (2013-06-28 12:00:55 UTC) #71
dvyukov
On Fri, Jun 28, 2013 at 4:00 PM, <alex.brainman@gmail.com> wrote: > On 2013/06/28 11:15:14, dvyukov ...
10 years, 9 months ago (2013-06-28 12:10:15 UTC) #72
brainman
On 2013/06/28 11:22:28, dvyukov wrote: > Regarding the test failure, I don't have any good ...
10 years, 9 months ago (2013-06-28 12:14:30 UTC) #73
dvyukov
On Fri, Jun 28, 2013 at 4:14 PM, <alex.brainman@gmail.com> wrote: > On 2013/06/28 11:22:28, dvyukov ...
10 years, 9 months ago (2013-06-28 12:28:27 UTC) #74
brainman
On 2013/06/28 12:28:27, dvyukov wrote: > What other things can it do? There must be ...
10 years, 9 months ago (2013-06-28 12:32:28 UTC) #75
dvyukov
Maybe it correlates with some cold error paths, e.g. WSARecv returns nil instead of ERROR_IO_PENDING ...
10 years, 9 months ago (2013-06-28 12:51:15 UTC) #76
brainman
message: On 2013/06/28 12:28:27, dvyukov wrote: > ... > What other things can it do? ...
10 years, 8 months ago (2013-07-05 02:58:58 UTC) #77
brainman
Hello dvyukov@google.com (cc: golang-dev@googlegroups.com), Please take another look.
10 years, 8 months ago (2013-07-05 02:59:33 UTC) #78
dvyukov
On 2013/07/05 02:58:58, brainman wrote: > message: On 2013/06/28 12:28:27, dvyukov wrote: > > ...
10 years, 8 months ago (2013-07-17 10:42:15 UTC) #79
brainman
On 2013/07/17 10:42:15, dvyukov wrote: > ... Different runnable thread still should be > alternated ...
10 years, 8 months ago (2013-07-19 04:50:16 UTC) #80
dvyukov
On 2013/07/19 04:50:16, brainman wrote: > On 2013/07/17 10:42:15, dvyukov wrote: > > ... Different ...
10 years, 8 months ago (2013-07-19 12:00:02 UTC) #81
brainman
10 years, 8 months ago (2013-07-22 02:50:04 UTC) #82
*** Submitted as https://code.google.com/p/go/source/detail?r=e6a6dc0d9c22 ***

net: implement netpoll for windows

Moves the network poller from net package into runtime.

benchmark                           old ns/op    new ns/op    delta
BenchmarkTCP4OneShot                   316386       287061   -9.27%
BenchmarkTCP4OneShot-2                 339822       313424   -7.77%
BenchmarkTCP4OneShot-3                 330057       306589   -7.11%
BenchmarkTCP4OneShotTimeout            341775       287061  -16.01%
BenchmarkTCP4OneShotTimeout-2          380835       295849  -22.32%
BenchmarkTCP4OneShotTimeout-3          398412       328070  -17.66%
BenchmarkTCP4Persistent                 40622        33392  -17.80%
BenchmarkTCP4Persistent-2               44528        35736  -19.74%
BenchmarkTCP4Persistent-3               44919        36907  -17.84%
BenchmarkTCP4PersistentTimeout          45309        33588  -25.87%
BenchmarkTCP4PersistentTimeout-2        50289        38079  -24.28%
BenchmarkTCP4PersistentTimeout-3        51559        37103  -28.04%
BenchmarkTCP6OneShot                   361305       345645   -4.33%
BenchmarkTCP6OneShot-2                 361305       331976   -8.12%
BenchmarkTCP6OneShot-3                 376929       347598   -7.78%
BenchmarkTCP6OneShotTimeout            361305       322212  -10.82%
BenchmarkTCP6OneShotTimeout-2          378882       333928  -11.86%
BenchmarkTCP6OneShotTimeout-3          388647       335881  -13.58%
BenchmarkTCP6Persistent                 47653        35345  -25.83%
BenchmarkTCP6Persistent-2               49215        35736  -27.39%
BenchmarkTCP6Persistent-3               38474        37493   -2.55%
BenchmarkTCP6PersistentTimeout          56637        34369  -39.32%
BenchmarkTCP6PersistentTimeout-2        42575        38079  -10.56%
BenchmarkTCP6PersistentTimeout-3        44137        37689  -14.61%

R=dvyukov
CC=golang-dev
https://codereview.appspot.com/8670044
Sign in to reply to this message.

Powered by Google App Engine
RSS Feeds Recent Issues | This issue
This is Rietveld f62528b