Commits · ec8caf696ae44b9ea4624025a613dad4301398c4 · Kirill Smelkov / go

21 Mar, 2013 10 commits

runtime: mark strings without going through an intermediate buffer · ec8caf69
Jan Ziak authored Mar 21, 2013
```
R=rsc
CC=golang-dev
https://golang.org/cl/7949043
```
ec8caf69
go/format: fix documentation · e95c41f3
Robert Griesemer authored Mar 21, 2013
```
R=r
CC=golang-dev
https://golang.org/cl/7920048
```
e95c41f3

crypto/sha1: faster amd64, 386 implementations · 2f32138a

Russ Cox authored Mar 21, 2013

-- amd64 --

On a MacBookPro10,2 (Core i5):

benchmark              old ns/op    new ns/op    delta
BenchmarkHash8Bytes          785          592  -24.59%
BenchmarkHash1K             8727         3014  -65.46%
BenchmarkHash8K            64926        20723  -68.08%

benchmark               old MB/s     new MB/s  speedup
BenchmarkHash8Bytes        10.19        13.50    1.32x
BenchmarkHash1K           117.34       339.71    2.90x
BenchmarkHash8K           126.17       395.31    3.13x

For comparison, on the same machine, openssl 0.9.8r reports
its sha1 speed as 341 MB/s for 1K and 404 MB/s for 8K.

On an Intel Xeon E5520:

benchmark              old ns/op    new ns/op    delta
BenchmarkHash8Bytes          984          707  -28.15%
BenchmarkHash1K            11141         3466  -68.89%
BenchmarkHash8K            82435        23411  -71.60%

benchmark               old MB/s     new MB/s  speedup
BenchmarkHash8Bytes         8.13        11.31    1.39x
BenchmarkHash1K            91.91       295.36    3.21x
BenchmarkHash8K            99.37       349.91    3.52x

For comparison, on the same machine, openssl 1.0.1 reports
its sha1 speed as 286 MB/s for 1K and 394 MB/s for 8K.

-- 386 --

On a MacBookPro10,2 (Core i5):

benchmark              old ns/op    new ns/op    delta
BenchmarkHash8Bytes         1041          713  -31.51%
BenchmarkHash1K            15612         3382  -78.34%
BenchmarkHash8K           110152        22733  -79.36%

benchmark               old MB/s     new MB/s  speedup
BenchmarkHash8Bytes         7.68        11.21    1.46x
BenchmarkHash1K            65.59       302.76    4.62x
BenchmarkHash8K            74.37       360.36    4.85x

On an Intel Xeon E5520:

benchmark              old ns/op    new ns/op    delta
BenchmarkHash8Bytes         1221          842  -31.04%
BenchmarkHash1K            14643         4137  -71.75%
BenchmarkHash8K           108722        27394  -74.80%

benchmark               old MB/s     new MB/s  speedup
BenchmarkHash8Bytes         6.55         9.49    1.45x
BenchmarkHash1K            69.93       247.51    3.54x
BenchmarkHash8K            75.35       299.04    3.97x

R=agl, dave
CC=golang-dev
https://golang.org/cl/7763049

2f32138a

crypto/md5: faster amd64, 386 implementations · 25cbd534

Russ Cox authored Mar 21, 2013

-- amd64 --

On a MacBookPro10,2 (Core i5):

benchmark                       old ns/op    new ns/op    delta
BenchmarkHash8Bytes                   471          524  +11.25%
BenchmarkHash1K                      3018         2220  -26.44%
BenchmarkHash8K                     20634        14604  -29.22%
BenchmarkHash8BytesUnaligned          468          523  +11.75%
BenchmarkHash1KUnaligned             3006         2212  -26.41%
BenchmarkHash8KUnaligned            20820        14652  -29.63%

benchmark                        old MB/s     new MB/s  speedup
BenchmarkHash8Bytes                 16.98        15.26    0.90x
BenchmarkHash1K                    339.26       461.19    1.36x
BenchmarkHash8K                    397.00       560.92    1.41x
BenchmarkHash8BytesUnaligned        17.08        15.27    0.89x
BenchmarkHash1KUnaligned           340.65       462.75    1.36x
BenchmarkHash8KUnaligned           393.45       559.08    1.42x

For comparison, on the same machine, openssl 0.9.8r reports
its md5 speed as 350 MB/s for 1K and 410 MB/s for 8K.

On an Intel Xeon E5520:

benchmark                       old ns/op    new ns/op    delta
BenchmarkHash8Bytes                   565          607   +7.43%
BenchmarkHash1K                      3753         2475  -34.05%
BenchmarkHash8K                     25945        16250  -37.37%
BenchmarkHash8BytesUnaligned          559          594   +6.26%
BenchmarkHash1KUnaligned             3754         2474  -34.10%
BenchmarkHash8KUnaligned            26011        16359  -37.11%

benchmark                        old MB/s     new MB/s  speedup
BenchmarkHash8Bytes                 14.15        13.17    0.93x
BenchmarkHash1K                    272.83       413.58    1.52x
BenchmarkHash8K                    315.74       504.11    1.60x
BenchmarkHash8BytesUnaligned        14.31        13.46    0.94x
BenchmarkHash1KUnaligned           272.73       413.78    1.52x
BenchmarkHash8KUnaligned           314.93       500.73    1.59x

For comparison, on the same machine, openssl 1.0.1 reports
its md5 speed as 443 MB/s for 1K and 513 MB/s for 8K.

-- 386 --

On a MacBookPro10,2 (Core i5):

benchmark                       old ns/op    new ns/op    delta
BenchmarkHash8Bytes                   602          670  +11.30%
BenchmarkHash1K                      4038         2549  -36.87%
BenchmarkHash8K                     27879        16690  -40.13%
BenchmarkHash8BytesUnaligned          602          670  +11.30%
BenchmarkHash1KUnaligned             4025         2546  -36.75%
BenchmarkHash8KUnaligned            27844        16692  -40.05%

benchmark                        old MB/s     new MB/s  speedup
BenchmarkHash8Bytes                 13.28        11.93    0.90x
BenchmarkHash1K                    253.58       401.69    1.58x
BenchmarkHash8K                    293.83       490.81    1.67x
BenchmarkHash8BytesUnaligned        13.27        11.94    0.90x
BenchmarkHash1KUnaligned           254.40       402.05    1.58x
BenchmarkHash8KUnaligned           294.21       490.77    1.67x

On an Intel Xeon E5520:

benchmark                       old ns/op    new ns/op    delta
BenchmarkHash8Bytes                   752          716   -4.79%
BenchmarkHash1K                      5307         2799  -47.26%
BenchmarkHash8K                     36993        18042  -51.23%
BenchmarkHash8BytesUnaligned          748          730   -2.41%
BenchmarkHash1KUnaligned             5301         2795  -47.27%
BenchmarkHash8KUnaligned            36983        18085  -51.10%

benchmark                        old MB/s     new MB/s  speedup
BenchmarkHash8Bytes                 10.64        11.16    1.05x
BenchmarkHash1K                    192.93       365.80    1.90x
BenchmarkHash8K                    221.44       454.03    2.05x
BenchmarkHash8BytesUnaligned        10.69        10.95    1.02x
BenchmarkHash1KUnaligned           193.15       366.36    1.90x
BenchmarkHash8KUnaligned           221.51       452.96    2.04x

R=agl
CC=golang-dev
https://golang.org/cl/7621049

25cbd534

crypto/rc4: faster amd64, 386 implementations · 1af96080

Russ Cox authored Mar 21, 2013

-- amd64 --

On a MacBookPro10,2 (Core i5):

benchmark           old ns/op    new ns/op    delta
BenchmarkRC4_128          470          421  -10.43%
BenchmarkRC4_1K          3123         3275   +4.87%
BenchmarkRC4_8K         26351        25866   -1.84%

benchmark            old MB/s     new MB/s  speedup
BenchmarkRC4_128       272.22       303.40    1.11x
BenchmarkRC4_1K        327.80       312.58    0.95x
BenchmarkRC4_8K        307.24       313.00    1.02x

For comparison, on the same machine, openssl 0.9.8r reports
its rc4 speed as somewhat under 350 MB/s for both 1K and 8K.
The Core i5 performance can be boosted another 20%, but only
by making the Xeon performance significantly slower.

On an Intel Xeon E5520:

benchmark           old ns/op    new ns/op    delta
BenchmarkRC4_128          774          417  -46.12%
BenchmarkRC4_1K          6121         3200  -47.72%
BenchmarkRC4_8K         48394        25151  -48.03%

benchmark            old MB/s     new MB/s  speedup
BenchmarkRC4_128       165.18       306.84    1.86x
BenchmarkRC4_1K        167.28       319.92    1.91x
BenchmarkRC4_8K        167.29       321.89    1.92x

For comparison, on the same machine, openssl 1.0.1
(which uses a different implementation than 0.9.8r)
reports its rc4 speed as 587 MB/s for 1K and 601 MB/s for 8K.
It is using SIMD instructions to do more in parallel.

So there's still some improvement to be had, but even so,
this is almost 2x faster than what it replaced.

-- 386 --

On a MacBookPro10,2 (Core i5):

benchmark           old ns/op    new ns/op    delta
BenchmarkRC4_128         3491          421  -87.94%
BenchmarkRC4_1K         28063         3205  -88.58%
BenchmarkRC4_8K        220392        25228  -88.55%

benchmark            old MB/s     new MB/s  speedup
BenchmarkRC4_128        36.66       303.81    8.29x
BenchmarkRC4_1K         36.49       319.42    8.75x
BenchmarkRC4_8K         36.73       320.90    8.74x

On an Intel Xeon E5520:

benchmark           old ns/op    new ns/op    delta
BenchmarkRC4_128         2268          524  -76.90%
BenchmarkRC4_1K         18161         4137  -77.22%
BenchmarkRC4_8K        142396        32350  -77.28%

benchmark            old MB/s     new MB/s  speedup
BenchmarkRC4_128        56.42       244.13    4.33x
BenchmarkRC4_1K         56.38       247.46    4.39x
BenchmarkRC4_8K         56.86       250.26    4.40x

R=agl
CC=golang-dev
https://golang.org/cl/7547050

1af96080

runtime: explicitly remove fd's from epoll waitset before close() · 44840786

Dmitriy Vyukov authored Mar 21, 2013

Fixes #5061.

Current code relies on the fact that fd's are automatically removed from epoll set when closed. However, it is not true. Underlying file description is removed from epoll set only when *all* fd's referring to it are closed.

There are 2 bad consequences:
1. Kernel delivers notifications on already closed fd's.
2. The following sequence of events leads to error:
   - add fd1 to epoll
   - dup fd1 = fd2
   - close fd1 (not removed from epoll since we've dup'ed the fd)
   - dup fd2 = fd1 (get the same fd as fd1)
   - add fd1 to epoll = EEXIST

So, if fd can be potentially dup'ed of fork'ed, it's necessary to explicitly remove the fd from epoll set.

R=golang-dev, bradfitz, dave
CC=golang-dev
https://golang.org/cl/7870043

44840786

runtime: faster parallel GC · d4c80d19

Dmitriy Vyukov authored Mar 21, 2013

Use per-thread work buffers instead of global mutex-protected pool. This eliminates contention from parallel scan phase.

benchmark old ns/op new ns/op delta
garbage.BenchmarkTree2-8 97100768 71417553 -26.45%
garbage.BenchmarkTree2LastPause-8 970931485 714103692 -26.45%
garbage.BenchmarkTree2Pause-8 469127802 345029253 -26.45%
garbage.BenchmarkParser-8 2880950854 2715456901 -5.74%
garbage.BenchmarkParserLastPause-8 137047399 103336476 -24.60%
garbage.BenchmarkParserPause-8 80686028 58922680 -26.97%

R=golang-dev, 0xe2.0x9a.0x9b, dave, adg, rsc, iant
CC=golang-dev
https://golang.org/cl/7816044

d4c80d19

cmd/gc: implement more cases in racewalk. · 656bc3eb

Rémy Oudompheng authored Mar 21, 2013

Add missing CLOSUREVAR in switch.
Mark MAKE, string conversion nodes as impossible.
Control statements do not need instrumentation.
Instrument COM and LROT nodes.
Instrument map length.

Update #4228

R=dvyukov, golang-dev
CC=golang-dev
https://golang.org/cl/7504047

656bc3eb

crypto/tls: use method values · 76d5e2ce

Brad Fitzpatrick authored Mar 20, 2013

Currently fails with a compiler error, though.

R=golang-dev, agl, rsc
CC=golang-dev
https://golang.org/cl/7933043

76d5e2ce

cmd/gc: fix escape analysis of method values · 38e9b077
Russ Cox authored Mar 20, 2013
```
R=ken2
CC=golang-dev
https://golang.org/cl/7518050
```
38e9b077

20 Mar, 2013 17 commits

src/pkg: fixed various typos · 0ad265d4
Robert Griesemer authored Mar 20, 2013
```
R=golang-dev, bradfitz
CC=golang-dev
https://golang.org/cl/7936043
```
0ad265d4
runtime: free map structures more aggressively · bcb39a77
Keith Randall authored Mar 20, 2013
```
R=rsc, bradfitz, khr
CC=golang-dev
https://golang.org/cl/7849047
```
bcb39a77

spec: fix description of initialization · b636f192

Rob Pike authored Mar 20, 2013

The analysis does not depend on the values of the items.
Fixes #4648.

R=golang-dev, gri, rsc
CC=golang-dev
https://golang.org/cl/7593050

b636f192

go/parser: use method values · 5ae4012b
Robert Griesemer authored Mar 20, 2013
```
R=golang-dev, r
CC=golang-dev
https://golang.org/cl/7858045
```
5ae4012b
runtime: Use aligned loads for AES key schedule. · db53d97a
Keith Randall authored Mar 20, 2013
```
R=rsc, minux.ma, khr
CC=golang-dev
https://golang.org/cl/7763050
```
db53d97a
cmd/gc: add OCHECKNOTNIL to race walker (fix build) · a63d5dd1
Russ Cox authored Mar 20, 2013
```
TBR=dvyukov
CC=golang-dev
https://golang.org/cl/7635046
```
a63d5dd1
cmd/gc: reject methods on non-locals, even if symbol exists · 7c3694c4
Daniel Morsing authored Mar 20, 2013
```
Fixes #5089.

R=golang-dev, rsc
CC=golang-dev
https://golang.org/cl/7767044
```
7c3694c4
cmd/gc: implement method values · d3c758d7
Russ Cox authored Mar 20, 2013
```
R=ken2, ken
CC=golang-dev
https://golang.org/cl/7546052
```
d3c758d7

spec: define method values · 6e15683c

Russ Cox authored Mar 20, 2013

Fixes #2280.

R=golang-dev, r, bradfitz, iant, andybalholm, gri
CC=golang-dev
https://golang.org/cl/7816045

6e15683c

runtime: faster hashmap implementation. · 00224a35

Keith Randall authored Mar 20, 2013

Hashtable is arranged as an array of
8-entry buckets with chained overflow.
Each bucket has 8 extra hash bits
per key to provide quick lookup within
a bucket.  Table is grown incrementally.

Update #3885
Go time drops from 0.51s to 0.34s.

R=r, rsc, m3b, dave, bradfitz, khr, ugorji, remyoudompheng
CC=golang-dev
https://golang.org/cl/7504044

00224a35

runtime: prevent garbage collection during hashmap insertion (fix 2) · 2001f0c2

Jan Ziak authored Mar 20, 2013

Fixes #5074 in multi-threaded scenarios.

R=golang-dev, daniel.morsing, dave, dvyukov, bradfitz, rsc
CC=golang-dev, remyoudompheng
https://golang.org/cl/7916043

2001f0c2

net/http/fcgi: Request.Body should always be non-nil · bd21f7f1

Brad Fitzpatrick authored Mar 20, 2013

Found this inconsistency from net/http's Server while
debugging Issue 4183

Unfortunately this package lacks testing around this,
or most of child.go. :/

R=golang-dev, adg
CC=golang-dev
https://golang.org/cl/7735046

bd21f7f1

misc/swig: remove old broken Makefile · eeb98224
Ian Lance Taylor authored Mar 20, 2013
```
R=golang-dev, bradfitz
CC=golang-dev
https://golang.org/cl/7786046
```
eeb98224

cmd/6c, cmd/8c: fix stack allocated Biobuf leaking at exit · f701e1c3

Dave Cheney authored Mar 20, 2013

Fixes #5085.

{6,8}c/swt.c allocates a third Biobuf in automatic memory which is not terminated at the end of the function. This causes the buffer to be 'in use' when the batexit handler fires, confusing valgrind.

Huge thanks to DMorsing for the diagnosis.

R=golang-dev, daniel.morsing, rsc
CC=golang-dev
https://golang.org/cl/7844044

f701e1c3

cmd/ld: remove remaining occurrence of isobj. · b35019fe
Rémy Oudompheng authored Mar 20, 2013
```
Fixes cgo builds.

R=golang-dev, minux.ma
CC=golang-dev
https://golang.org/cl/7784044
```
b35019fe

net/http: new benchmark to measure server without network · 1ce0d72f

Brad Fitzpatrick authored Mar 19, 2013

No net package involved. And with ReportAllocs we can see
how much garbage is created per request.

R=adg, dave
CC=golang-dev
https://golang.org/cl/7913044

1ce0d72f

cmd/5l: fix arm build · 85773e6b

Dave Cheney authored Mar 20, 2013

Fixes build. One part of ../misc/cgo/test appears to be unhappy but this does not cause the build to fail.

R=rsc, r
CC=golang-dev
https://golang.org/cl/7769047

85773e6b

19 Mar, 2013 13 commits

doc/go1.1.html: add links to the text, write some new sections · ee5a794f
Rob Pike authored Mar 19, 2013
```
R=golang-dev, rsc, dave
CC=golang-dev
https://golang.org/cl/7914043
```
ee5a794f
cmd/gc: implement revised rule for shifts in equality. · 465ff4cf
Rémy Oudompheng authored Mar 20, 2013
```
R=rsc, daniel.morsing
CC=golang-dev
https://golang.org/cl/7613046
```
465ff4cf
bytes,strings: remove user name from BUG in comment · 1b6b1bc6
Rob Pike authored Mar 19, 2013
```
R=golang-dev, dave
CC=golang-dev
https://golang.org/cl/7856048
```
1b6b1bc6

runtime: prevent garbage collection during hashmap insertion · 54dffda2

Jan Ziak authored Mar 19, 2013

Inserting a key-value pair into a hashmap storing keys or values
indirectly can cause the garbage collector to find the hashmap in
an inconsistent	state.

Fixes #5074.

R=golang-dev, minux.ma, rsc
CC=golang-dev
https://golang.org/cl/7913043

54dffda2

debug/dwarf: support for DWARF 3 · b79afe1b
Ian Lance Taylor authored Mar 19, 2013
```
R=rsc
CC=golang-dev
https://golang.org/cl/7662045
```
b79afe1b

cmd/ld: fix typo · 47ce4bd3

Russ Cox authored Mar 19, 2013

R=golang-dev, minux.ma
CC=golang-dev
https://golang.org/cl/7552045

47ce4bd3

cmd/ld: generate relocated DWARF in hostobj mode · 40d356e9

Russ Cox authored Mar 19, 2013

While we're here, downgrade DWARF to version 2.
We're not using any version 3 features, and OS X gdb
only supports version 2.

Fixes #3436.

R=golang-dev, minux.ma
CC=golang-dev
https://golang.org/cl/7891044

40d356e9

misc/dashboard: show correct dir name in README · ddddd70c
Robert Griesemer authored Mar 19, 2013
```
R=golang-dev, bradfitz
CC=golang-dev
https://golang.org/cl/7909043
```
ddddd70c

cmd/ld: replace -hostobj with -linkmode · b4f3533c

Russ Cox authored Mar 19, 2013

Still disabled. Need to fix TLS.

R=golang-dev, minux.ma, bradfitz
CC=golang-dev
https://golang.org/cl/7783044

b4f3533c

runtime: optimize calls to addroot() · ab08eac5
Jan Ziak authored Mar 19, 2013
```
R=golang-dev, rsc
CC=dvyukov, golang-dev
https://golang.org/cl/7879043
```
ab08eac5
cmd/gc: support channel types in the garbage collector · 6e69df61
Jan Ziak authored Mar 19, 2013
```
R=golang-dev, dvyukov, rsc
CC=golang-dev
https://golang.org/cl/7473044
```
6e69df61
unicode: modify a comment to the convention format. · d8714ca4
Oling Cat authored Mar 19, 2013
```
R=golang-dev, r, rsc
CC=golang-dev
https://golang.org/cl/7869043
```
d8714ca4

go/doc, godoc: improved note reading · 5268119f

Robert Griesemer authored Mar 19, 2013

- A note doesn't have to be in the first
comment of a comment group anymore, and
several notes may appear in the same comment
group (e.g., it is fairly common to have a
TODO(uid) note immediately following another
comment).

- Define a doc.Note type which also contains
note uid and position info.

- Better formatting in godoc output. The position
information is not yet used, but could be used to
locate the note in the source text if desired.

Fixes #4843.

R=r, cnicolaou
CC=gobot, golang-dev
https://golang.org/cl/7496048

5268119f