Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Add an experimental batch module #1134

Open
wants to merge 9 commits into
base: master
Choose a base branch
from
Open

Add an experimental batch module #1134

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
9 commits
Select commit Hold shift + click to select a range
db210f4
batch: Initialize an experimental batch module
siv2r Aug 19, 2022
e0dbd66
batch: Add `create` and `destroy` APIs
siv2r Aug 19, 2022
b7c7b58
batch, ecmult: Add `batch_verify` API and refactor `strauss_batch`
siv2r Aug 19, 2022
9c120ef
batch: Add `batch_add_*` APIs
siv2r Aug 19, 2022
b3d370f
batch: Add API usage example
siv2r Aug 19, 2022
d0446a8
batch,ecmult: Add tests for core batch APIs and `strauss_batch` refactor
siv2r Aug 19, 2022
8a77b2c
batch: Add tests for `batch_add_*` APIs
siv2r Aug 19, 2022
d3fbae5
batch, extrakeys: Add benchmark for batch verify and `tweak_add_check`
siv2r Aug 19, 2022
eb49e93
batch: Generate graphs for batch verification speed up
siv2r Aug 20, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
15 changes: 15 additions & 0 deletions doc/speedup-batch.md
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,15 @@
# Schnorrsig Batch Verification Speedup

![Speedup over single verification](speedup-batch/schnorrsig-speedup-batch.png)

# Tweak Pubkey Check Batch Verification Speedup

![Speedup over single verification](speedup-batch/tweakcheck-speedup-batch.png)

Build steps
-----------
To generate the above graphs on your local machine:

$ cd doc/speedup-batch
$ make
$ make speedup-batch.png
1 change: 1 addition & 0 deletions doc/speedup-batch/.gitignore
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1 @@
*.dat
23 changes: 23 additions & 0 deletions doc/speedup-batch/Makefile
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,23 @@
schnorrsig_data = schnorrsig_batch.dat schnorrsig_single.dat
tweak_data = tweak_batch.dat tweak_single.dat

bench_output.txt: bench.sh
SECP256K1_BENCH_ITERS=500000 ./bench.sh bench_output.txt

schnorrsig_batch.dat: bench_output.txt
cat bench_output.txt | grep -v "schnorrsig_batch_verify_1 " | awk '{ gsub(/ /,""); print }' | awk -F, 'match($$0, /schnorrsig_batch_verify_([0-9]+)/, arr) {print arr[1] " " $$3}' > schnorrsig_batch.dat

schnorrsig_single.dat: bench_output.txt
cat bench_output.txt | awk '{ gsub(/ /,""); print }' | awk -F, 'match($$0, /schnorrsig_verify/) {print $$3}' > schnorrsig_single.dat

tweak_batch.dat: bench_output.txt
cat bench_output.txt | grep -v "tweak_check_batch_verify_1 " | awk '{ gsub(/ /,""); print }' | awk -F, 'match($$0, /tweak_check_batch_verify_([0-9]+)/, arr) {print arr[1] " " $$3}' > tweak_batch.dat

tweak_single.dat: bench_output.txt
cat bench_output.txt | awk '{ gsub(/ /,""); print }' | awk -F, 'match($$0, /tweak_add_check/) {print $$3}' > tweak_single.dat

speedup-batch.png: $(schnorrsig_data) $(tweak_data) plot.gp
gnuplot plot.gp

clean:
rm *.log *.txt *.dat *.png
13 changes: 13 additions & 0 deletions doc/speedup-batch/bench.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,13 @@
#!/bin/bash

output_file=$1
cur_dir=$(pwd)

cd ../../
echo "HEAD: $(git rev-parse --short HEAD)" > "$cur_dir/$output_file.log"
make clean
./autogen.sh
./configure --enable-experimental --enable-module-batch --enable-module-schnorrsig >> "$cur_dir/$output_file.log"
make -j
./bench schnorrsig > "$cur_dir/$output_file"
./bench extrakeys >> "$cur_dir/$output_file"
137 changes: 137 additions & 0 deletions doc/speedup-batch/bench_output.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,137 @@
Benchmark , Min(us) , Avg(us) , Max(us)

schnorrsig_sign , 50.4 , 50.5 , 50.7
schnorrsig_verify , 89.1 , 89.2 , 89.3
schnorrsig_batch_verify_1 , 104.0 , 104.0 , 104.0
Copy link
Contributor

@sipa sipa Aug 23, 2022 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

batch_verify_1 shouldn't be slower than non-batch verify. Is it possible to revert to using non-batch validation logic for this case?

Copy link
Contributor Author

@siv2r siv2r Aug 23, 2022 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Not possible with the current design.

The non-batch validation (secp256k1_schnorrsig_verify) logic looks something like this:

  • calc rj using secp256k1_ecmult: Rj = sG - eP
  • convert rj (gej) to r (ge)
  • check if the r.x = sig[0:32] and r.y = even

one schnorrsig occupies two points in the batch, and one tweak check occupies one point in the batch. If a batch contains two points, there is no guarantee that they are from a schnorrsig (R, P). It could be from two tweak checks. So, we can't use the r.y = even check.

Hence, I tried implementing a slightly modified schnorrsig_verify logic (not implement in this PR):

  • calc neg_rj using secp256k1_ecmult: neg_Rj = -s*G + batch.scalars[1]*batch.points[1]
  • check if neg_rj + batch.points[0] == inf using _gej_add_var
    • batch.scalars[0] = 1 always. So, we don't need to use ecmult again

This gives somewhat better benchmarks than before:

Benchmark                          ,    Min(us)    ,    Avg(us)    ,    Max(us)    

schnorrsig_sign                    ,    49.1       ,    50.1       ,    53.4    
schnorrsig_verify                  ,    86.6       ,    87.2       ,    88.4    
schnorrsig_batch_verify_1          ,    94.7       ,    95.0       ,    95.2

But schnorrsig_batch_verify_1 is still slower than schnorrsig_verify.

schnorrsig_batch_verify_2 , 89.0 , 89.1 , 89.1
schnorrsig_batch_verify_3 , 84.1 , 84.1 , 84.1
schnorrsig_batch_verify_4 , 81.5 , 81.5 , 81.5
schnorrsig_batch_verify_5 , 79.9 , 79.9 , 79.9
schnorrsig_batch_verify_7 , 78.0 , 78.1 , 78.3
schnorrsig_batch_verify_9 , 77.0 , 77.0 , 77.1
schnorrsig_batch_verify_11 , 76.2 , 76.3 , 76.3
schnorrsig_batch_verify_14 , 75.6 , 75.6 , 75.6
schnorrsig_batch_verify_17 , 75.2 , 75.2 , 75.2
schnorrsig_batch_verify_21 , 74.8 , 74.8 , 74.8
schnorrsig_batch_verify_26 , 74.5 , 74.6 , 74.9
schnorrsig_batch_verify_32 , 74.3 , 74.5 , 74.7
schnorrsig_batch_verify_39 , 74.1 , 74.1 , 74.1
schnorrsig_batch_verify_47 , 73.9 , 73.9 , 73.9
schnorrsig_batch_verify_57 , 74.5 , 74.5 , 74.5
schnorrsig_batch_verify_69 , 74.3 , 74.3 , 74.5
schnorrsig_batch_verify_83 , 74.1 , 74.1 , 74.2
schnorrsig_batch_verify_100 , 73.9 , 74.0 , 74.1
schnorrsig_batch_verify_121 , 74.1 , 74.1 , 74.2
schnorrsig_batch_verify_146 , 73.9 , 73.9 , 74.0
schnorrsig_batch_verify_176 , 74.0 , 74.2 , 74.5
schnorrsig_batch_verify_212 , 73.9 , 74.1 , 74.1
schnorrsig_batch_verify_255 , 74.0 , 74.0 , 74.1
schnorrsig_batch_verify_307 , 73.9 , 74.0 , 74.1
schnorrsig_batch_verify_369 , 73.9 , 73.9 , 73.9
schnorrsig_batch_verify_443 , 73.9 , 74.1 , 74.3
schnorrsig_batch_verify_532 , 74.0 , 74.0 , 74.1
schnorrsig_batch_verify_639 , 73.9 , 74.0 , 74.0
schnorrsig_batch_verify_767 , 73.9 , 73.9 , 73.9
schnorrsig_batch_verify_921 , 74.0 , 74.0 , 74.1
schnorrsig_batch_verify_1106 , 73.9 , 73.9 , 73.9
schnorrsig_batch_verify_1328 , 73.9 , 74.1 , 74.2
schnorrsig_batch_verify_1594 , 74.0 , 74.1 , 74.1
schnorrsig_batch_verify_1913 , 74.0 , 74.0 , 74.0
schnorrsig_batch_verify_2296 , 74.0 , 74.0 , 74.0
schnorrsig_batch_verify_2756 , 73.9 , 74.0 , 74.1
schnorrsig_batch_verify_3308 , 74.1 , 74.1 , 74.2
schnorrsig_batch_verify_3970 , 74.1 , 74.2 , 74.4
schnorrsig_batch_verify_4765 , 74.0 , 74.1 , 74.2
schnorrsig_batch_verify_5719 , 74.0 , 74.1 , 74.1
schnorrsig_batch_verify_6863 , 74.0 , 74.1 , 74.1
schnorrsig_batch_verify_8236 , 74.0 , 74.1 , 74.1
schnorrsig_batch_verify_9884 , 74.0 , 74.1 , 74.3
schnorrsig_batch_verify_11861 , 74.0 , 74.0 , 74.1
schnorrsig_batch_verify_14234 , 73.9 , 74.0 , 74.1
schnorrsig_batch_verify_17081 , 73.9 , 73.9 , 73.9
schnorrsig_batch_verify_20498 , 73.9 , 74.0 , 74.0
schnorrsig_batch_verify_24598 , 73.9 , 74.0 , 74.1
schnorrsig_batch_verify_29518 , 73.9 , 74.0 , 74.1
schnorrsig_batch_verify_35422 , 73.9 , 73.9 , 73.9
schnorrsig_batch_verify_42507 , 73.9 , 74.0 , 74.0
schnorrsig_batch_verify_51009 , 73.9 , 74.1 , 74.3
schnorrsig_batch_verify_61211 , 73.9 , 73.9 , 74.0
schnorrsig_batch_verify_73454 , 73.9 , 74.0 , 74.3
schnorrsig_batch_verify_88145 , 73.9 , 74.0 , 74.1
schnorrsig_batch_verify_105775 , 74.0 , 74.1 , 74.1
schnorrsig_batch_verify_126931 , 73.9 , 74.0 , 74.1
schnorrsig_batch_verify_152318 , 73.9 , 73.9 , 74.0
schnorrsig_batch_verify_182782 , 73.9 , 73.9 , 74.0
schnorrsig_batch_verify_219339 , 73.9 , 73.9 , 74.0
schnorrsig_batch_verify_263207 , 74.0 , 74.1 , 74.3
schnorrsig_batch_verify_315849 , 73.9 , 74.0 , 74.0
schnorrsig_batch_verify_379019 , 73.9 , 73.9 , 73.9
schnorrsig_batch_verify_454823 , 74.0 , 74.0 , 74.0
Benchmark , Min(us) , Avg(us) , Max(us)

tweak_add_check , 64.7 , 64.7 , 65.0
tweak_check_batch_verify_1 , 69.7 , 69.8 , 69.8
tweak_check_batch_verify_2 , 57.2 , 57.2 , 57.3
tweak_check_batch_verify_3 , 52.0 , 52.1 , 52.2
tweak_check_batch_verify_4 , 49.4 , 49.5 , 49.5
tweak_check_batch_verify_5 , 47.9 , 47.9 , 47.9
tweak_check_batch_verify_7 , 46.1 , 46.1 , 46.2
tweak_check_batch_verify_9 , 45.2 , 45.2 , 45.4
tweak_check_batch_verify_11 , 44.5 , 44.6 , 44.6
tweak_check_batch_verify_14 , 43.9 , 43.9 , 43.9
tweak_check_batch_verify_17 , 43.5 , 43.5 , 43.5
tweak_check_batch_verify_21 , 43.1 , 43.1 , 43.1
tweak_check_batch_verify_26 , 42.8 , 42.8 , 42.8
tweak_check_batch_verify_32 , 42.5 , 42.6 , 42.6
tweak_check_batch_verify_39 , 42.3 , 42.4 , 42.4
tweak_check_batch_verify_47 , 42.2 , 42.2 , 42.2
tweak_check_batch_verify_57 , 42.1 , 42.2 , 42.3
tweak_check_batch_verify_69 , 42.0 , 42.1 , 42.1
tweak_check_batch_verify_83 , 41.9 , 41.9 , 41.9
tweak_check_batch_verify_100 , 41.8 , 41.9 , 41.9
tweak_check_batch_verify_121 , 42.1 , 42.1 , 42.1
tweak_check_batch_verify_146 , 42.0 , 42.0 , 42.0
tweak_check_batch_verify_176 , 41.9 , 41.9 , 42.0
tweak_check_batch_verify_212 , 41.8 , 41.9 , 41.9
tweak_check_batch_verify_255 , 41.9 , 41.9 , 41.9
tweak_check_batch_verify_307 , 41.8 , 41.9 , 41.9
tweak_check_batch_verify_369 , 41.9 , 42.0 , 42.1
tweak_check_batch_verify_443 , 41.9 , 41.9 , 41.9
tweak_check_batch_verify_532 , 41.9 , 41.9 , 41.9
tweak_check_batch_verify_639 , 41.9 , 41.9 , 42.0
tweak_check_batch_verify_767 , 41.9 , 41.9 , 41.9
tweak_check_batch_verify_921 , 41.9 , 41.9 , 41.9
tweak_check_batch_verify_1106 , 41.9 , 41.9 , 41.9
tweak_check_batch_verify_1328 , 41.9 , 41.9 , 42.0
tweak_check_batch_verify_1594 , 41.9 , 41.9 , 42.0
tweak_check_batch_verify_1913 , 41.9 , 41.9 , 41.9
tweak_check_batch_verify_2296 , 41.9 , 41.9 , 41.9
tweak_check_batch_verify_2756 , 41.8 , 41.9 , 41.9
tweak_check_batch_verify_3308 , 41.9 , 41.9 , 42.0
tweak_check_batch_verify_3970 , 41.9 , 41.9 , 41.9
tweak_check_batch_verify_4765 , 41.8 , 41.9 , 41.9
tweak_check_batch_verify_5719 , 41.9 , 42.0 , 42.1
tweak_check_batch_verify_6863 , 42.0 , 42.0 , 42.0
tweak_check_batch_verify_8236 , 42.0 , 42.0 , 42.0
tweak_check_batch_verify_9884 , 41.9 , 41.9 , 42.0
tweak_check_batch_verify_11861 , 41.9 , 42.0 , 42.1
tweak_check_batch_verify_14234 , 41.9 , 42.0 , 42.0
tweak_check_batch_verify_17081 , 41.8 , 41.9 , 41.9
tweak_check_batch_verify_20498 , 41.8 , 41.9 , 41.9
tweak_check_batch_verify_24598 , 41.8 , 41.9 , 41.9
tweak_check_batch_verify_29518 , 41.9 , 41.9 , 41.9
tweak_check_batch_verify_35422 , 41.9 , 41.9 , 41.9
tweak_check_batch_verify_42507 , 41.8 , 41.8 , 41.9
tweak_check_batch_verify_51009 , 41.9 , 41.9 , 41.9
tweak_check_batch_verify_61211 , 41.8 , 41.8 , 41.8
tweak_check_batch_verify_73454 , 41.8 , 42.0 , 42.2
tweak_check_batch_verify_88145 , 41.9 , 41.9 , 41.9
tweak_check_batch_verify_105775 , 41.8 , 41.8 , 41.8
tweak_check_batch_verify_126931 , 41.8 , 41.9 , 41.9
tweak_check_batch_verify_152318 , 41.8 , 41.9 , 42.0
tweak_check_batch_verify_182782 , 41.9 , 41.9 , 41.9
tweak_check_batch_verify_219339 , 41.9 , 42.0 , 42.0
tweak_check_batch_verify_263207 , 41.9 , 42.0 , 42.1
tweak_check_batch_verify_315849 , 41.9 , 41.9 , 41.9
tweak_check_batch_verify_379019 , 41.9 , 41.9 , 42.0
tweak_check_batch_verify_454823 , 41.9 , 41.9 , 41.9
127 changes: 127 additions & 0 deletions doc/speedup-batch/bench_output.txt.log
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,127 @@
HEAD: 6ddb0d0c
checking build system type... x86_64-pc-linux-gnu
checking host system type... x86_64-pc-linux-gnu
checking for a BSD-compatible install... /usr/bin/install -c
checking whether build environment is sane... yes
checking for a thread-safe mkdir -p... /usr/bin/mkdir -p
checking for gawk... gawk
checking whether make sets $(MAKE)... yes
checking whether make supports nested variables... yes
checking whether make supports nested variables... (cached) yes
checking for gcc... gcc
checking whether the C compiler works... yes
checking for C compiler default output file name... a.out
checking for suffix of executables...
checking whether we are cross compiling... no
checking for suffix of object files... o
checking whether we are using the GNU C compiler... yes
checking whether gcc accepts -g... yes
checking for gcc option to accept ISO C89... none needed
checking whether gcc understands -c and -o together... yes
checking whether make supports the include directive... yes (GNU style)
checking dependency style of gcc... gcc3
checking dependency style of gcc... gcc3
checking for ar... ar
checking the archiver (ar) interface... ar
checking how to print strings... printf
checking for a sed that does not truncate output... /usr/bin/sed
checking for grep that handles long lines and -e... /usr/bin/grep
checking for egrep... /usr/bin/grep -E
checking for fgrep... /usr/bin/grep -F
checking for ld used by gcc... /usr/bin/ld
checking if the linker (/usr/bin/ld) is GNU ld... yes
checking for BSD- or MS-compatible name lister (nm)... /usr/bin/nm -B
checking the name lister (/usr/bin/nm -B) interface... BSD nm
checking whether ln -s works... yes
checking the maximum length of command line arguments... 1572864
checking how to convert x86_64-pc-linux-gnu file names to x86_64-pc-linux-gnu format... func_convert_file_noop
checking how to convert x86_64-pc-linux-gnu file names to toolchain format... func_convert_file_noop
checking for /usr/bin/ld option to reload object files... -r
checking for objdump... objdump
checking how to recognize dependent libraries... pass_all
checking for dlltool... no
checking how to associate runtime and link libraries... printf %s\n
checking for archiver @FILE support... @
checking for strip... strip
checking for ranlib... ranlib
checking command to parse /usr/bin/nm -B output from gcc object... ok
checking for sysroot... no
checking for a working dd... /usr/bin/dd
checking how to truncate binary pipes... /usr/bin/dd bs=4096 count=1
checking for mt... mt
checking if mt is a manifest tool... no
checking how to run the C preprocessor... gcc -E
checking for ANSI C header files... yes
checking for sys/types.h... yes
checking for sys/stat.h... yes
checking for stdlib.h... yes
checking for string.h... yes
checking for memory.h... yes
checking for strings.h... yes
checking for inttypes.h... yes
checking for stdint.h... yes
checking for unistd.h... yes
checking for dlfcn.h... yes
checking for objdir... .libs
checking if gcc supports -fno-rtti -fno-exceptions... no
checking for gcc option to produce PIC... -fPIC -DPIC
checking if gcc PIC flag -fPIC -DPIC works... yes
checking if gcc static flag -static works... yes
checking if gcc supports -c -o file.o... yes
checking if gcc supports -c -o file.o... (cached) yes
checking whether the gcc linker (/usr/bin/ld -m elf_x86_64) supports shared libraries... yes
checking whether -lc should be explicitly linked in... no
checking dynamic linker characteristics... GNU/Linux ld.so
checking how to hardcode library paths into programs... immediate
checking whether stripping libraries is possible... yes
checking if libtool supports shared libraries... yes
checking whether to build shared libraries... yes
checking whether to build static libraries... yes
checking if gcc supports -Werror=unknown-warning-option... no
checking if gcc supports -std=c89 -pedantic -Wno-long-long -Wnested-externs -Wshadow -Wstrict-prototypes -Wundef... yes
checking if gcc supports -Wno-overlength-strings... yes
checking if gcc supports -Wall... yes
checking if gcc supports -Wno-unused-function... yes
checking if gcc supports -Wextra... yes
checking if gcc supports -Wcast-align... yes
checking if gcc supports -Wcast-align=strict... yes
checking if gcc supports -Wconditional-uninitialized... no
checking if gcc supports -fvisibility=hidden... yes
checking for valgrind support... yes
checking for x86_64 assembly availability... yes
configure: ******
configure: WARNING: experimental build
configure: Experimental features do not have stable APIs or properties, and may not be safe for production use.
configure: Building batch verification module: yes
configure: ******
checking that generated files are newer than configure... done
configure: creating ./config.status
config.status: creating Makefile
config.status: creating libsecp256k1.pc
config.status: creating src/libsecp256k1-config.h
config.status: src/libsecp256k1-config.h is unchanged
config.status: executing depfiles commands
config.status: executing libtool commands

Build Options:
with external callbacks = no
with benchmarks = yes
with tests = yes
with coverage = no
with examples = no
module ecdh = no
module recovery = no
module extrakeys = yes
module schnorrsig = yes
module batch = yes

asm = x86_64
ecmult window size = 15
ecmult gen prec. bits = 4

valgrind = yes
CC = gcc
CPPFLAGS =
SECP_CFLAGS = -O2 -std=c89 -pedantic -Wno-long-long -Wnested-externs -Wshadow -Wstrict-prototypes -Wundef -Wno-overlength-strings -Wall -Wno-unused-function -Wextra -Wcast-align -Wcast-align=strict -fvisibility=hidden
CFLAGS = -g -O2
LDFLAGS =
Loading