• fanzha02's avatar
    cmd/internal/obj/arm64: encode float constants into FMOVS/FMOVD instructions · c430adf1
    fanzha02 authored
    Current assembler rewrites float constants to values stored in memory
    except 0.0, which is not performant. This patch uses the FMOVS/FMOVD
    instructions to move some available floating-point immediate constants
    into SIMD&FP destination registers. These available constants can be
    encoded into FMOVS/FMOVD instructions, checked by the chipfloat7() function.
    
    go1 benchmark results.
    name                     old time/op    new time/op    delta
    BinaryTree17-8              6.27s ± 1%     6.27s ± 1%    ~     (p=0.762 n=10+8)
    Fannkuch11-8                5.42s ± 1%     5.38s ± 0%  -0.63%  (p=0.000 n=10+10)
    FmtFprintfEmpty-8          92.9ns ± 1%    93.4ns ± 0%  +0.47%  (p=0.004 n=9+8)
    FmtFprintfString-8          169ns ± 2%     170ns ± 4%    ~     (p=0.378 n=10+10)
    FmtFprintfInt-8             197ns ± 1%     196ns ± 1%  -0.77%  (p=0.009 n=10+9)
    FmtFprintfIntInt-8          284ns ± 1%     286ns ± 1%    ~     (p=0.051 n=10+10)
    FmtFprintfPrefixedInt-8     419ns ± 0%     422ns ± 1%  +0.69%  (p=0.038 n=6+10)
    FmtFprintfFloat-8           458ns ± 0%     463ns ± 1%  +1.14%  (p=0.000 n=10+10)
    FmtManyArgs-8              1.35µs ± 2%    1.36µs ± 1%  +0.91%  (p=0.043 n=10+10)
    GobDecode-8                16.0ms ± 2%    15.5ms ± 1%   -3.39%  (p=0.000 n=10+10)
    GobEncode-8                11.9ms ± 3%    11.4ms ± 1%   -3.98%  (p=0.000 n=10+9)
    Gzip-8                      621ms ± 0%     625ms ± 0%   +0.59%  (p=0.000 n=9+10)
    Gunzip-8                   74.0ms ± 1%    74.3ms ± 0%     ~     (p=0.059 n=9+8)
    HTTPClientServer-8          116µs ± 1%     116µs ± 1%     ~     (p=0.165 n=10+10)
    JSONEncode-8               29.3ms ± 1%    29.5ms ± 0%   +0.72%  (p=0.001 n=10+10)
    JSONDecode-8                145ms ± 1%     148ms ± 2%   +2.06%  (p=0.000 n=10+10)
    Mandelbrot200-8            9.67ms ± 0%    9.48ms ± 1%   -1.92%  (p=0.000 n=8+10)
    GoParse-8                  7.55ms ± 0%    7.60ms ± 0%   +0.57%  (p=0.000 n=9+10)
    RegexpMatchEasy0_32-8       234ns ± 0%     210ns ± 0%  -10.13%  (p=0.000 n=8+10)
    RegexpMatchEasy0_1K-8       753ns ± 1%     729ns ± 0%   -3.17%  (p=0.000 n=10+8)
    RegexpMatchEasy1_32-8       225ns ± 0%     224ns ± 0%   -0.44%  (p=0.000 n=9+9)
    RegexpMatchEasy1_1K-8      1.03µs ± 0%    1.04µs ± 1%   +1.29%  (p=0.000 n=10+10)
    RegexpMatchMedium_32-8      320ns ± 3%     296ns ± 6%   -7.50%  (p=0.000 n=10+10)
    RegexpMatchMedium_1K-8     77.0µs ± 5%    73.6µs ± 1%     ~     (p=0.393 n=10+10)
    RegexpMatchHard_32-8       3.93µs ± 0%    3.89µs ± 1%   -0.95%  (p=0.000 n=10+9)
    RegexpMatchHard_1K-8        120µs ± 5%     115µs ± 1%     ~     (p=0.739 n=10+10)
    Revcomp-8                   1.07s ± 0%     1.08s ± 1%   +0.63%  (p=0.000 n=10+9)
    Template-8                  165ms ± 1%     163ms ± 1%   -1.05%  (p=0.001 n=8+10)
    TimeParse-8                 751ns ± 1%     749ns ± 1%     ~     (p=0.209 n=10+10)
    TimeFormat-8                759ns ± 1%     751ns ± 1%   -0.96%  (p=0.001 n=10+10)
    
    name                     old speed      new speed      delta
    GobDecode-8              48.0MB/s ± 2%  49.6MB/s ± 1%   +3.50%  (p=0.000 n=10+10)
    GobEncode-8              64.5MB/s ± 3%  67.1MB/s ± 1%   +4.08%  (p=0.000 n=10+9)
    Gzip-8                   31.2MB/s ± 0%  31.1MB/s ± 0%   -0.55%  (p=0.000 n=9+8)
    Gunzip-8                  262MB/s ± 1%   261MB/s ± 0%     ~     (p=0.059 n=9+8)
    JSONEncode-8             66.3MB/s ± 1%  65.8MB/s ± 0%   -0.72%  (p=0.001 n=10+10)
    JSONDecode-8             13.4MB/s ± 1%  13.2MB/s ± 1%   -2.02%  (p=0.000 n=10+10)
    GoParse-8                7.67MB/s ± 0%  7.63MB/s ± 0%   -0.57%  (p=0.000 n=9+10)
    RegexpMatchEasy0_32-8     136MB/s ± 0%   152MB/s ± 0%  +11.45%  (p=0.000 n=10+10)
    RegexpMatchEasy0_1K-8    1.36GB/s ± 1%  1.40GB/s ± 0%   +3.25%  (p=0.000 n=10+8)
    RegexpMatchEasy1_32-8     142MB/s ± 0%   143MB/s ± 0%   +0.35%  (p=0.000 n=10+9)
    RegexpMatchEasy1_1K-8     992MB/s ± 0%   980MB/s ± 1%   -1.27%  (p=0.000 n=10+10)
    RegexpMatchMedium_32-8   3.12MB/s ± 3%  3.38MB/s ± 6%   +8.17%  (p=0.000 n=10+10)
    RegexpMatchMedium_1K-8   13.3MB/s ± 5%  13.9MB/s ± 1%     ~     (p=0.362 n=10+10)
    RegexpMatchHard_32-8     8.14MB/s ± 0%  8.21MB/s ± 1%   +0.95%  (p=0.000 n=10+9)
    RegexpMatchHard_1K-8     8.54MB/s ± 5%  8.90MB/s ± 1%     ~     (p=0.636 n=10+10)
    Revcomp-8                 238MB/s ± 0%   236MB/s ± 1%   -0.63%  (p=0.000 n=10+9)
    Template-8               11.8MB/s ± 1%  11.9MB/s ± 1%   +1.07%  (p=0.001 n=8+10)
    
    Change-Id: I57b372d8dcd47e6aec39893843b20385d5d9c37e
    Reviewed-on: https://go-review.googlesource.com/129555
    Run-TryBot: Cherry Zhang <cherryyz@google.com>
    TryBot-Result: Gobot Gobot <gobot@golang.org>
    Reviewed-by: 's avatarCherry Zhang <cherryyz@google.com>
    c430adf1
Name
Last commit
Last update
..
archive Loading commit data...
bufio Loading commit data...
builtin Loading commit data...
bytes Loading commit data...
cmd Loading commit data...
compress Loading commit data...
container Loading commit data...
context Loading commit data...
crypto Loading commit data...
database/sql Loading commit data...
debug Loading commit data...
encoding Loading commit data...
errors Loading commit data...
expvar Loading commit data...
flag Loading commit data...
fmt Loading commit data...
go Loading commit data...
hash Loading commit data...
html Loading commit data...
image Loading commit data...
index/suffixarray Loading commit data...
internal Loading commit data...
io Loading commit data...
log Loading commit data...
math Loading commit data...
mime Loading commit data...
net Loading commit data...
os Loading commit data...
path Loading commit data...
plugin Loading commit data...
reflect Loading commit data...
regexp Loading commit data...
runtime Loading commit data...
sort Loading commit data...
strconv Loading commit data...
strings Loading commit data...
sync Loading commit data...
syscall Loading commit data...
testing Loading commit data...
text Loading commit data...
time Loading commit data...
unicode Loading commit data...
unsafe Loading commit data...
vendor/golang_org/x Loading commit data...
Make.dist Loading commit data...
all.bash Loading commit data...
all.bat Loading commit data...
all.rc Loading commit data...
androidtest.bash Loading commit data...
bootstrap.bash Loading commit data...
buildall.bash Loading commit data...
clean.bash Loading commit data...
clean.bat Loading commit data...
clean.rc Loading commit data...
cmp.bash Loading commit data...
iostest.bash Loading commit data...
make.bash Loading commit data...
make.bat Loading commit data...
make.rc Loading commit data...
naclmake.bash Loading commit data...
nacltest.bash Loading commit data...
race.bash Loading commit data...
race.bat Loading commit data...
run.bash Loading commit data...
run.bat Loading commit data...
run.rc Loading commit data...