ee8972cd12
Rewrite addVW to use a fast path and remove the original vector and non vector implementation of addVW in assembly. This CL uses a similar idea as CL 164968, where we copy the rest of words when we know carry bit is zero. In addition, since we are copying vector of words, a faster implementation of copy is written in this CL to copy a word or multiple words at a time. Benchmarks: name old time/op new time/op delta AddVW/1-18 4.56ns ± 0% 4.01ns ± 6% -12.14% (p=0.000 n=18+20) AddVW/2-18 5.54ns ± 0% 4.42ns ± 5% -20.20% (p=0.000 n=18+20) AddVW/3-18 6.55ns ± 0% 4.61ns ± 0% -29.62% (p=0.000 n=16+18) AddVW/4-18 6.11ns ± 2% 5.12ns ± 6% -16.19% (p=0.000 n=20+20) AddVW/5-18 7.32ns ± 4% 5.14ns ± 0% -29.77% (p=0.000 n=20+19) AddVW/10-18 10.6ns ± 2% 7.2ns ± 1% -31.47% (p=0.000 n=20+20) AddVW/100-18 49.6ns ± 2% 18.0ns ± 0% -63.63% (p=0.000 n=20+20) AddVW/1000-18 465ns ± 3% 244ns ± 0% -47.54% (p=0.000 n=20+20) AddVW/10000-18 4.99µs ± 4% 2.97µs ± 0% -40.54% (p=0.000 n=20+20) AddVW/100000-18 48.3µs ± 3% 30.8µs ± 1% -36.29% (p=0.000 n=20+20) [Geo mean] 58.1ns 38.0ns -34.57% name old speed new speed delta AddVW/1-18 1.76GB/s ± 0% 2.00GB/s ± 6% +14.04% (p=0.000 n=20+20) AddVW/2-18 2.89GB/s ± 0% 3.63GB/s ± 5% +25.55% (p=0.000 n=18+20) AddVW/3-18 3.66GB/s ± 0% 5.21GB/s ± 0% +42.25% (p=0.000 n=18+19) AddVW/4-18 5.24GB/s ± 2% 6.27GB/s ± 6% +19.61% (p=0.000 n=20+20) AddVW/5-18 5.47GB/s ± 4% 7.78GB/s ± 0% +42.28% (p=0.000 n=20+18) AddVW/10-18 7.55GB/s ± 2% 11.04GB/s ± 1% +46.09% (p=0.000 n=20+20) AddVW/100-18 16.1GB/s ± 2% 44.3GB/s ± 0% +174.77% (p=0.000 n=20+20) AddVW/1000-18 17.2GB/s ± 3% 32.8GB/s ± 1% +90.58% (p=0.000 n=20+20) AddVW/10000-18 16.0GB/s ± 4% 26.9GB/s ± 0% +68.11% (p=0.000 n=20+20) AddVW/100000-18 16.6GB/s ± 3% 26.0GB/s ± 1% +56.94% (p=0.000 n=20+20) [Geo mean] 7.03GB/s 10.75GB/s +52.93% Change-Id: Idbb73f3178311bd2b18a93bdc1e48f26869d2f6a Reviewed-on: https://go-review.googlesource.com/c/go/+/209679 Reviewed-by: Michael Munday <mike.munday@ibm.com> Run-TryBot: Michael Munday <mike.munday@ibm.com> TryBot-Result: Gobot Gobot <gobot@golang.org> |
||
|---|---|---|
| .github | ||
| api | ||
| doc | ||
| lib/time | ||
| misc | ||
| src | ||
| test | ||
| .gitattributes | ||
| .gitignore | ||
| AUTHORS | ||
| CONTRIBUTING.md | ||
| CONTRIBUTORS | ||
| LICENSE | ||
| PATENTS | ||
| README.md | ||
| SECURITY.md | ||
| favicon.ico | ||
| robots.txt | ||
README.md
The Go Programming Language
Go is an open source programming language that makes it easy to build simple, reliable, and efficient software.
Gopher image by Renee French, licensed under Creative Commons 3.0 Attributions license.
Our canonical Git repository is located at https://go.googlesource.com/go. There is a mirror of the repository at https://github.com/golang/go.
Unless otherwise noted, the Go source files are distributed under the BSD-style license found in the LICENSE file.
Download and Install
Binary Distributions
Official binary distributions are available at https://golang.org/dl/.
After downloading a binary release, visit https://golang.org/doc/install or load doc/install.html in your web browser for installation instructions.
Install From Source
If a binary distribution is not available for your combination of operating system and architecture, visit https://golang.org/doc/install/source or load doc/install-source.html in your web browser for source installation instructions.
Contributing
Go is the work of thousands of contributors. We appreciate your help!
To contribute, please read the contribution guidelines: https://golang.org/doc/contribute.html
Note that the Go project uses the issue tracker for bug reports and proposals only. See https://golang.org/wiki/Questions for a list of places to ask questions about the Go language.