提交 a30b0522 编写于 作者: A Andy Polyakov

x86 assembly pack: update performance results.

Reviewed-by: NRichard Levitte <levitte@openssl.org>
上级 f15eed3b
...@@ -62,7 +62,9 @@ ...@@ -62,7 +62,9 @@
# Westmere 3.77/1.37 1.37 1.52 1.27 # Westmere 3.77/1.37 1.37 1.52 1.27
# * Bridge 5.07/0.98 0.99 1.09 0.91 1.10 # * Bridge 5.07/0.98 0.99 1.09 0.91 1.10
# Haswell 4.44/0.80 0.97 1.03 0.72 0.76 # Haswell 4.44/0.80 0.97 1.03 0.72 0.76
# Skylake 2.68/0.65 0.65 0.66 0.64 0.66
# Silvermont 5.77/3.56 3.67 4.03 3.46 4.03 # Silvermont 5.77/3.56 3.67 4.03 3.46 4.03
# Goldmont 3.84/1.39 1.39 1.63 1.31 1.70
# Bulldozer 5.80/0.98 1.05 1.24 0.93 1.23 # Bulldozer 5.80/0.98 1.05 1.24 0.93 1.23
$PREFIX="aesni"; # if $PREFIX is set to "AES", the script $PREFIX="aesni"; # if $PREFIX is set to "AES", the script
......
...@@ -28,6 +28,7 @@ ...@@ -28,6 +28,7 @@
# Westmere 9.50/+45% 3.35 # Westmere 9.50/+45% 3.35
# Sandy Bridge 10.5/+47% 3.20 # Sandy Bridge 10.5/+47% 3.20
# Haswell 8.15/+50% 2.83 # Haswell 8.15/+50% 2.83
# Skylake 7.53/+22% 2.75
# Silvermont 17.4/+36% 8.35 # Silvermont 17.4/+36% 8.35
# Goldmont 13.4/+40% 4.36 # Goldmont 13.4/+40% 4.36
# Sledgehammer 10.2/+54% # Sledgehammer 10.2/+54%
......
...@@ -32,6 +32,7 @@ ...@@ -32,6 +32,7 @@
# Sandy Bridge 8.31/+42% 5.45/6.76 2.72 # Sandy Bridge 8.31/+42% 5.45/6.76 2.72
# Ivy Bridge 6.71/+46% 5.40/6.49 2.41 # Ivy Bridge 6.71/+46% 5.40/6.49 2.41
# Haswell 5.92/+43% 5.20/6.45 2.42 1.23 # Haswell 5.92/+43% 5.20/6.45 2.42 1.23
# Skylake 5.87/+39% 4.70/- 2.31 1.19
# Silvermont 12.0/+33% 7.75/7.40 7.03(iii) # Silvermont 12.0/+33% 7.75/7.40 7.03(iii)
# Goldmont 10.6/+17% 5.10/- 3.28 # Goldmont 10.6/+17% 5.10/- 3.28
# Sledgehammer 7.28/+52% -/14.2(ii) - # Sledgehammer 7.28/+52% -/14.2(ii) -
......
...@@ -29,6 +29,7 @@ ...@@ -29,6 +29,7 @@
# Westmere 4.58/+100% 1.43 # Westmere 4.58/+100% 1.43
# Sandy Bridge 3.90/+100% 1.36 # Sandy Bridge 3.90/+100% 1.36
# Haswell 3.88/+70% 1.18 0.72 # Haswell 3.88/+70% 1.18 0.72
# Skylake 3.10/+60% 1.14 0.62
# Silvermont 11.0/+40% 4.80 # Silvermont 11.0/+40% 4.80
# Goldmont 4.10/+200% 2.10 # Goldmont 4.10/+200% 2.10
# VIA Nano 6.71/+90% 2.47 # VIA Nano 6.71/+90% 2.47
......
...@@ -104,10 +104,12 @@ ...@@ -104,10 +104,12 @@
# Sandy Bridge 8.8 6.2/+40% 5.1(**)/+73% # Sandy Bridge 8.8 6.2/+40% 5.1(**)/+73%
# Ivy Bridge 7.2 4.8/+51% 4.7(**)/+53% # Ivy Bridge 7.2 4.8/+51% 4.7(**)/+53%
# Haswell 6.5 4.3/+51% 4.1(**)/+58% # Haswell 6.5 4.3/+51% 4.1(**)/+58%
# Skylake 6.4 4.1/+55% 4.1(**)/+55%
# Bulldozer 11.6 6.0/+92% # Bulldozer 11.6 6.0/+92%
# VIA Nano 10.6 7.5/+41% # VIA Nano 10.6 7.5/+41%
# Atom 12.5 9.3(*)/+35% # Atom 12.5 9.3(*)/+35%
# Silvermont 14.5 9.9(*)/+46% # Silvermont 14.5 9.9(*)/+46%
# Goldmont 8.8 6.7/+30% 1.7(***)/+415%
# #
# (*) Loop is 1056 instructions long and expected result is ~8.25. # (*) Loop is 1056 instructions long and expected result is ~8.25.
# The discrepancy is because of front-end limitations, so # The discrepancy is because of front-end limitations, so
...@@ -115,6 +117,8 @@ ...@@ -115,6 +117,8 @@
# limited parallelism. # limited parallelism.
# #
# (**) As per above comment, the result is for AVX *plus* sh[rl]d. # (**) As per above comment, the result is for AVX *plus* sh[rl]d.
#
# (***) SHAEXT result
$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1; $0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
push(@INC,"${dir}","${dir}../../perlasm"); push(@INC,"${dir}","${dir}../../perlasm");
......
...@@ -57,14 +57,17 @@ ...@@ -57,14 +57,17 @@
# Sandy Bridge 25 - 15.9 12.4 11.6 # Sandy Bridge 25 - 15.9 12.4 11.6
# Ivy Bridge 24 - 15.0 11.4 10.3 # Ivy Bridge 24 - 15.0 11.4 10.3
# Haswell 22 - 13.9 9.46 7.80 # Haswell 22 - 13.9 9.46 7.80
# Skylake 20 - 14.9 9.50 7.70
# Bulldozer 36 - 27/22 17.0 13.6 # Bulldozer 36 - 27/22 17.0 13.6
# VIA Nano 36 - 25/22 16.8 16.5 # VIA Nano 36 - 25/22 16.8 16.5
# Atom 50 - 30/25 21.9 18.9 # Atom 50 - 30/25 21.9 18.9
# Silvermont 40 - 34/31 22.9 20.6 # Silvermont 40 - 34/31 22.9 20.6
# Goldmont 29 - 20 16.3(***)
# #
# (*) numbers after slash are for unrolled loop, where applicable; # (*) numbers after slash are for unrolled loop, where applicable;
# (**) x86_64 assembly performance is presented for reference # (**) x86_64 assembly performance is presented for reference
# purposes, results are best-available; # purposes, results are best-available;
# (***) SHAEXT result is 4.1, strangely enough better than 64-bit one;
$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1; $0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
push(@INC,"${dir}","${dir}../../perlasm"); push(@INC,"${dir}","${dir}../../perlasm");
......
...@@ -32,6 +32,7 @@ ...@@ -32,6 +32,7 @@
# Sandy Bridge 58 - 35 11.9 11.2 # Sandy Bridge 58 - 35 11.9 11.2
# Ivy Bridge 50 - 33 11.5 8.17 # Ivy Bridge 50 - 33 11.5 8.17
# Haswell 46 - 29 11.3 7.66 # Haswell 46 - 29 11.3 7.66
# Skylake 40 - 26 13.3 7.25
# Bulldozer 121 - 50 14.0 13.5 # Bulldozer 121 - 50 14.0 13.5
# VIA Nano 91 - 52 33 14.7 # VIA Nano 91 - 52 33 14.7
# Atom 126 - 68 48(***) 14.7 # Atom 126 - 68 48(***) 14.7
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册