delta/go-git.git/src/internal/bytealg, branch dev.cmdgo

all: replace runtime SSE2 detection with GO386 setting

2021-08-23T21:22:58+00:00

When GO386=sse2 we can assume sse2 to be present without
a runtime check. If GO386=softfloat is set we can avoid
the usage of SSE2 even if detected.

This might cause a memcpy, memclr and bytealg slowdown of Go
binaries compiled with softfloat on machines that support
SSE2. Such setups are rare and should use GO386=sse2 instead
if performance matters.

On targets that support SSE2 we avoid the runtime overhead of
dynamic cpu feature dispatch.

The removal of runtime sse2 checks also allows to simplify
internal/cpu further by removing handling of the required
feature option as a followup after this CL.

Change-Id: I90a853a8853a405cb665497c6d1a86556947ba17
Reviewed-on: https://go-review.googlesource.com/c/go/+/344350
Trust: Martin Möhrmann 
Run-TryBot: Martin Möhrmann 
TryBot-Result: Go Bot 
Reviewed-by: Keith Randall

[dev.typeparams] runtime, internal/bytealg: remove regabi fallback code on AMD64

2021-08-11T17:08:54+00:00

As we commit to always enabling register ABI on AMD64, remove the
fallback code.

Change-Id: I30556858ba4bac367495fa94f6a8682ecd771196
Reviewed-on: https://go-review.googlesource.com/c/go/+/341152
Trust: Cherry Mui 
Run-TryBot: Cherry Mui 
TryBot-Result: Go Bot 
Reviewed-by: Michael Knyszek 
Reviewed-by: Austin Clements

[dev.typeparams] all: merge master (785a8f6) into dev.typeparams

2021-06-16T07:59:13+00:00

- test/run.go

  CL 328050 added fixedbugs/issue46749.go to -G=3 excluded files list

Merge List:

+ 2021-06-16 785a8f677f cmd/compile: better error message for invalid untyped operation
+ 2021-06-16 a752bc0746 syscall: fix TestGroupCleanupUserNamespace test failure on Fedora
+ 2021-06-15 d77f4c0c5c net/http: improve some server docs
+ 2021-06-15 219fe9d547 cmd/go: ignore UTF8 BOM when reading source code
+ 2021-06-15 723f199edd cmd/link: set correct flags in .dynamic for PIE buildmode
+ 2021-06-15 4d2d89ff42 cmd/go, go/build: update docs to use //go:build syntax
+ 2021-06-15 033d885315 doc/go1.17: document go run pkg@version
+ 2021-06-15 ea8612ef42 syscall: disable c-shared test when no cgo, for windows/arm
+ 2021-06-15 abc56fd1a0 internal/bytealg: remove duplicate go:build line
+ 2021-06-15 4061d3463b syscall: rewrite handle inheritance test to use C rather than Powershell
+ 2021-06-15 cf4e3e3d3b reflect: explain why convertible or comparable types may still panic
+ 2021-06-14 7841cb14d9 doc/go1.17: assorted fixes
+ 2021-06-14 8a5a6f46dc debug/elf: don't apply DWARF relocations for ET_EXEC binaries
+ 2021-06-14 9d13f8d43e runtime: update the variable name in comment
+ 2021-06-14 0fd20ed5b6 reflect: use same conversion panic in reflect and runtime
+ 2021-06-14 6bbb0a9d4a cmd/internal/sys: mark windows/arm64 as c-shared-capable
+ 2021-06-14 d4f34f8c63 doc/go1.17: reword "results" in stack trace printing

Change-Id: I60d1f67c4d48cd4093c350fc89bd60c454d23944

internal/bytealg: remove duplicate go:build line

2021-06-15T10:13:08+00:00

Change-Id: I6b71bf468b9544820829f02e320673f5edd785fa
GitHub-Last-Rev: 8082ac5fba18e630dd2a21771837e6f0b1f9853f
GitHub-Pull-Request: golang/go#46683
Reviewed-on: https://go-review.googlesource.com/c/go/+/326730
Reviewed-by: Ian Lance Taylor 
Trust: Tobias Klauser

[dev.typeparams] runtime, internal/bytealg: port performance-critical functions to register ABI on ARM64

2021-06-03T16:28:19+00:00

This CL ports a few performance-critical assembly functions to use
register arguments directly. This is similar to CL 308931 and
CL 310184.

Change-Id: I6e30dfff17f76b8578ce8cfd51de21b66610fdb0
Reviewed-on: https://go-review.googlesource.com/c/go/+/324400
Trust: Cherry Mui 
Run-TryBot: Cherry Mui 
Reviewed-by: Than McIntosh 
Reviewed-by: Michael Knyszek 
TryBot-Result: Go Bot

[dev.typeparams] internal/bytealg: call memeqbody directly in memequal_varlen on ARM64

2021-06-03T14:29:14+00:00

Currently, memequal_varlen opens up a frame and call memequal,
which then tail-calls memeqbody. This CL changes memequal_varlen
tail-calls memeqbody directly.

This makes it simpler to switch to the register ABI in the next
CL.

Change-Id: Ia1367c0abb7f4755fe736c404411793fb9e5c04f
Reviewed-on: https://go-review.googlesource.com/c/go/+/324399
Trust: Cherry Mui 
Run-TryBot: Cherry Mui 
TryBot-Result: Go Bot 
Reviewed-by: Michael Knyszek

all: add //go:build lines to assembly files

2021-05-13T09:12:17+00:00

Don't add them to files in vendor and cmd/vendor though. These will be
pulled in by updating the respective dependencies.

For #41184

Change-Id: Icc57458c9b3033c347124323f33084c85b224c70
Reviewed-on: https://go-review.googlesource.com/c/go/+/319389
Trust: Tobias Klauser 
Run-TryBot: Tobias Klauser 
TryBot-Result: Go Bot 
Reviewed-by: Russ Cox

internal/bytealg: add power9 version of bytes index

2021-04-21T21:14:04+00:00

This adds a power9 version of the bytes.Index function
for little endian.

Here is the improvement on power9 for some of the Index
benchmarks:

Index/10           -0.14%
Index/32           -3.19%
Index/4K          -12.66%
Index/4M          -13.34%
Index/64M         -13.17%
Count/10           -0.59%
Count/32           -2.88%
Count/4K          -12.63%
Count/4M          -13.35%
Count/64M         -13.17%
IndexHard1        -23.03%
IndexHard2        -13.01%
IndexHard3        -22.12%
IndexHard4         +0.16%
CountHard1        -23.02%
CountHard2        -13.01%
CountHard3        -22.12%
IndexPeriodic/IndexPeriodic2  -22.85%
IndexPeriodic/IndexPeriodic4  -23.15%

Change-Id: Id72353e2771eba2efbb1544d5f0be65f8a9f0433
Reviewed-on: https://go-review.googlesource.com/c/go/+/311380
Run-TryBot: Carlos Eduardo Seo 
TryBot-Result: Go Bot 
Reviewed-by: Carlos Eduardo Seo 
Trust: Lynn Boger

bytes: add asm implementation for index on ppc64x

2021-04-15T17:32:04+00:00

This adds an asm implementation of index on ppc64le and
ppc64. It results in a significant improvement in some
of the benchmarks that use bytes.Index.

The implementation is based on a port of the s390x asm
implementation. Comments on the design are found
with the code.

The following improvements occurred on power8:

Index/10       70.7ns ± 0%    18.8ns ± 0%   -73.4
Index/32        165ns ± 0%      95ns ± 0%   -42.6
Index/4K       9.23µs ± 0%    4.91µs ± 0%   -46
Index/4M       9.52ms ± 0%    5.10ms ± 0%   -46.4
Index/64M       155ms ± 0%      85ms ± 0%   -45.1

Count/10       83.0ns ± 0%    32.1ns ± 0%   -61.3
Count/32        178ns ± 0%     109ns ± 0%   -38.8
Count/4K       9.24µs ± 0%    4.93µs ± 0%   -46
Count/4M       9.52ms ± 0%    5.10ms ± 0%   -46.4
Count/64M       155ms ± 0%      85ms ± 0%   -45.1

IndexHard1     2.36ms ± 0%    0.13ms ± 0%   -94.4
IndexHard2     2.36ms ± 0%    1.28ms ± 0%   -45.8
IndexHard3     2.36ms ± 0%    1.19ms ± 0%   -49.4
IndexHard4     2.36ms ± 0%    2.35ms ± 0%    -0.1

CountHard1     2.36ms ± 0%    0.13ms ± 0%   -94.4
CountHard2     2.36ms ± 0%    1.28ms ± 0%   -45.8
CountHard3     2.36ms ± 0%    1.19ms ± 0%   -49.4

IndexPeriodic/IndexPeriodic2  146µs ± 0%       8µs ± 0%   -94
IndexPeriodic/IndexPeriodic4  146µs ± 0%       8µs ± 0%   -94

Change-Id: I7dd2bb7e278726e27f51825ca8b2f8317d460e60
Reviewed-on: https://go-review.googlesource.com/c/go/+/309730
Run-TryBot: Lynn Boger 
Reviewed-by: Paul Murphy 
Reviewed-by: Carlos Eduardo Seo 
Trust: Carlos Eduardo Seo 
Trust: Lynn Boger

internal/bytealg: port more performance-critical functions to ABIInternal

2021-04-15T04:10:33+00:00

CL 308931 ported several runtime assembly functions to ABIInternal so
that compiler-generated ABIInternal calls don't go through ABI
wrappers, but it missed the runtime assembly functions that are
actually defined in internal/bytealg.

This eliminates the cost of wrappers for the BleveQuery and
GopherLuaKNucleotide benchmarks, but there's still more to do for
Tile38.

                                      0-base                1-wrappers
                                     sec/op        sec/op            vs base
BleveQuery                          6.507 ± 0%    6.477 ± 0%  -0.46% (p=0.004 n=20)
GopherLuaKNucleotide                30.39 ± 1%    30.34 ± 0%       ~ (p=0.301 n=20)
Tile38IntersectsCircle100kmRequest 1.038m ± 1%   1.080m ± 2%  +4.03% (p=0.000 n=20)

For #40724.

Change-Id: I0b722443f684fcb997b1d70802c5ed4b8d8f9829
Reviewed-on: https://go-review.googlesource.com/c/go/+/310184
Trust: Austin Clements 
Reviewed-by: Michael Knyszek 
Reviewed-by: Cherry Zhang