VanitySearch (Yet another address prefix finder)

[eozdem]

Hi,

First of all I need to say I really amazed on calculation speed of your program. I am trying to use it with Tesla V100 device but I am getting the following error;

Cuda version is 10.2 , Ubuntu server 18.04.4 LTS

command is :
./VanitySearch -gpu 1sdfa

result is:

VanitySearch v1.17
Difficulty: 264104224
Search: 1sdfa [Compressed]
Start Mon Mar 30 10:19:36 2020
Base Key: C3A84BA970A2B558DDC6C25670DB4257E70E544BEECEBE72B3A7EB02B4EC22C9
Number of CPU thread: 1
GPU: GPU #0 Tesla V100-SXM3-32GB (80x64 cores) Grid(640x128)
GPUEngine: Kernel: no kernel image is available for execution on the device

how can I solve it?

Thanks

[Jean_Luc]

Hi,
Yes there is an issue with Cuda 10.2 (both on Linux and Windows), it works fine with 10.0.
I didn't try it by myself but some users already reported this issue.

Edit: May be it is a different problem there, when you compiled what ccap did you use ?
Could you try to execute VanitySearch -l ?

[eozdem]

Hi,
Yes there is an issue with Cuda 10.2 (both on Linux and Windows), it works fine with 10.0.
I didn't try it by myself but some users already reported this issue.

Edit: May be it is a different problem there, when you compiled what ccap did you use ?
Could you try to execute VanitySearch -l ?

I compiled with ccap 75

-l results is below:

[eozdem]

Hi,
Yes there is an issue with Cuda 10.2 (both on Linux and Windows), it works fine with 10.0.
I didn't try it by myself but some users already reported this issue.

Edit: May be it is a different problem there, when you compiled what ccap did you use ?
Could you try to execute VanitySearch -l ?

I compiled with ccap 75

-l results is below:

I've tried with ccap 70 also ...

[Jean_Luc]

So good ccap is 70 for your hardware.
Did you get the same error message with ccap 70 ?
Could you try with cuda 10.0 ?

[eozdem]

So good ccap is 70 for your hardware.
Did you get the same error message with ccap 70 ?
Could you try with cuda 10.0 ?

I made clean OS installation (Ubuntu 18.04 LTS) and now it works with cuda 10.2 drivers without any problem. Do you suggest better command for my system? I feel speed should be more ... ?

[Jean_Luc]

Try to play with gridsize, ex: -g 640,256,640,256,....
On the benchmark by DaveF, for a Tesla V100-SXM2-16GB we have 1.815 GK/s per board.
You have 40GK/s , so 2.5 GK/s per board, not so bad...

[eozdem]

Try to play with gridsize, ex: -g 640,256,640,256,....
On the benchmark by DaveF, for a Tesla V100-SXM2-16GB we have 1.815 GK/s per board.
You have 40GK/s , so 2.5 GK/s per board, not so bad...

OMG, I changed the grid size as you suggested now I am getting 51 GK/s, amazing ...

Is there any chance to use regex with gpu?

[Jean_Luc]

OMG, I changed the grid size as you suggested now I am getting 51 GK/s, amazing ...

Great:)

Is there any chance to use regex with gpu?

You can only use wildcard characters ? and *:
C:\C++\VanitySearch\x64\Release>VanitySearchCUDA8.exe -stop -gpu 1AB*C?D
VanitySearch v1.15
Search: 1AB*C?D [Compressed]
Start Fri Apr  3 08:57:58 2020
Base Key: 181D206E917CAA73BBD771F81508B51E6EAD5B63A4ACFE7C0EB334436372AEE9
Number of CPU thread: 3
GPU: GPU #0 GeForce GTX 645 (3x192 cores) Grid(24x128)

PubAddress: 1ABsAWqsewBsiTAFzUuABocgEAP8JP9C7D
Priv (WIF): p2pkh:L3uif98QLuVLNb71SgTrhKt61SjKXdJk1czxNb4j8ekRhpHYcmMj
Priv (HEX): 0xC79B2930ABC2F63E601320D0412AC8BAF4E39933AF67FBC445AEEC73DC981FFA

[shlomogold]

Haven't been using the program for a couple of weeks, today gave it another try.
What I noticed:
Right now it is 20 Mkey/s GPU 20 Mkey/s and it periodically goes down to 10 / 10
two weeks ago it was 100 / 100 Mkeys. Why such a downgrade, what does it depend on?

Also, my laptop now makes a quiet sound while scanning, something like tac-tac-tac-tac-tac-tac every 5 seconds

[nc50lc]

Also, my laptop now makes a quiet sound while scanning, something like tac-tac-tac-tac-tac-tac every 5 seconds

That sounds like your HDD's "actuator arm" (slider head) tapping the disk (platters).
If it's an SSD, then it must be a barred fan that can still slip through or something because there are no other mechanical parts that can produce a "tac-tac-tac" sound.

The latter could slow-down your GPU because of cooling failure.

[shlomogold]

Also, my laptop now makes a quiet sound while scanning, something like tac-tac-tac-tac-tac-tac every 5 seconds

That sounds like your HDD's "actuator arm" (slider head) tapping the disk (platters).
If it's an SSD, then it must be a barred fan that can still slip through or something because there are no other mechanical parts that can produce a "tac-tac-tac" sound.

The latter could slow-down your GPU because of cooling failure.

it does sound like a HDD working, exactly like that tapping sound, but my laptop actually has SSD.
also it gets extremely hot underneath after just 30 minutes so I guess it is a barred fan then

[bangbumbang]

So good ccap is 70 for your hardware.
Did you get the same error message with ccap 70 ?
Could you try with cuda 10.0 ?

I made clean OS installation (Ubuntu 18.04 LTS) and now it works with cuda 10.2 drivers without any problem. Do you suggest better command for my system? I feel speed should be more ... ?

10.2 really? can u tell me the exact driver version? maybe nvidia patched something like a bug or so.. i would give it another try then..

did you also run -check and it was without errors ?

did you edit cuda config or used the one from source?

[inlovewiththedj]

Okay, as a workaround to the workaround  
I've constructed this batch file that should do the trick:

Copy the code below to a text file then rename it into "filename.bat", change the extension from 'txt' into 'bat'.
Create an input file with the same name as the 'input=' in the batch file containing a single line (not 100) of your desired prefix.
Remove "-gpu" if you don't have a supported Video Card.

Code:

@set input=1test.txt
@set output=result.txt
@set command=vanitysearch -stop -gpu

set loop=0
:start
%command% -i %input% -o 1%output%
set /a loop=%loop%+1 
if "%loop%"=="100" goto next
goto start

:next
%command% -i %input% -o 2%output%
set /a loop=%loop%+1 
if "%loop%"=="200" goto next2
goto next

:next2
%command% -i %input% -o 3%output%
set /a loop=%loop%+1 
if "%loop%"=="300" goto next3
goto next2

:next3
%command% -i %input% -o 4%output%
set /a loop=%loop%+1 
if "%loop%"=="400" goto next4
goto next3

:next4
%command% -i %input% -o 5%output%
set /a loop=%loop%+1 
if "%loop%"=="500" goto next5
goto next4

:next5
%command% -i %input% -o 6%output%
set /a loop=%loop%+1 
if "%loop%"=="600" goto next6
goto next5

:next6
%command% -i %input% -o 7%output%
set /a loop=%loop%+1 
if "%loop%"=="700" goto next7
goto next6

:next7
%command% -i %input% -o 8%output%
set /a loop=%loop%+1 
if "%loop%"=="800" goto next8
goto next7

:next8
%command% -i %input% -o 9%output%
set /a loop=%loop%+1 
if "%loop%"=="900" goto next9
goto next8

:next9
%command% -i %input% -o 10%output%
set /a loop=%loop%+1 
if "%loop%"=="1000" goto next10
goto next9

:next10
echo end

Output files will be '1result.txt', '2result.txt', '3result.txt'..... '10result.txt'; with 100 keys each.
Just edit the values of the first three lines (after the '=' sign) depending in your preferences (it's self-explanatory).

Nice work! Thank You
But i have one more question.
How to move the newly created file to a new location after every loop?

I tried this code:

Code:

:next2
%command% -i %input% -o 3%output%
set /a loop=%loop%+1 
if "%loop%"=="300" goto next3
goto next2
move *.txt c:\test\

but it did't work

[nc50lc]

Nice work! Thank You
But i have one more question.
How to move the newly created file to a new location after every loop?

Really? Is it that hard to manually move the text files instead of automating it?  
Ok, just add this line to the batch file's first three lines and %directory% after each %output%, it should look like this:

Code:

@set directory=C:\test\

Code:

.
.
.
.
set loop=0
:start
%command% -i %input% -o %directory%1%output%
set /a loop=%loop%+1 
if "%loop%"=="100" goto next
goto start

:next
%command% -i %input% -o %directory%2%output%
set /a loop=%loop%+1 
if "%loop%"=="200" goto next2
goto next
.
.
.

Make sure that 'test' folder under C:\ is already existing or it will fail to write.

[inlovewiththedj]

Hi

sorry for simple question (for some people) but can You write I mean step by step
how to install VanitySearch on Linux ie. Mint, Ubuntu etc.

First of all.
1. I downloaded archive from github https://github.com/JeanLucPons/VanitySearch/archive/1.17.tar.gz
2. I extracted this archive to my desktop
3. Then In my Linux Mint I opened terminal and I navigated to this folder ie. VanitySearch-1.17
4. Then I did put and run the command:

Code:

make

5. Don't know what next to open program and start generating new addresses

Thanks!

[MrFreeDragon]

Hi

sorry for simple question (for some people) but can You write I mean step by step
how to install VanitySearch on Linux ie. Mint, Ubuntu etc.

First of all.
1. I downloaded archive from github https://github.com/JeanLucPons/VanitySearch/archive/1.17.tar.gz
2. I extracted this archive to my desktop
3. Then In my Linux Mint I opened terminal and I navigated to this folder ie. VanitySearch-1.17
4. Then I did put and run the command:

Code:

make

5. Don't know what next to open program and start generating new addresses

Thanks!

If you do not use CUDA, just run tthis comand:
$ make all

If you use CUDA (GPU device), you should compile in another way:
$ make gpu=1 ccap=20 all

where: gpu=1 is the number of your GPU device (gpu=0 is also possible for one device)
           ccap=20 is the compute capability depending on your GPU device

Find your device here https://en.wikipedia.org/wiki/CUDA and input the version in accordance with your device (for example 20 for version 2.0; 61 for version 6.1; etc)

As soon as you compile the program you can run it in the following way:

$ ./VanitySearch -t 0 -gpu 1TryMe

where -gpu means you use GPU, -t 0 means that you so not use CPU (0 threads)
"./" - important part to run the program in Linux 

[inlovewiththedj]

Hi

sorry for simple question (for some people) but can You write I mean step by step
how to install VanitySearch on Linux ie. Mint, Ubuntu etc.

First of all.
1. I downloaded archive from github https://github.com/JeanLucPons/VanitySearch/archive/1.17.tar.gz
2. I extracted this archive to my desktop
3. Then In my Linux Mint I opened terminal and I navigated to this folder ie. VanitySearch-1.17
4. Then I did put and run the command:

Code:

make

5. Don't know what next to open program and start generating new addresses

Thanks!

If you do not use CUDA, just run tthis comand:
$ make all

If you use CUDA (GPU device), you should compile in another way:
$ make gpu=1 ccap=20 all

where: gpu=1 is the number of your GPU device (gpu=0 is also possible for one device)
           ccap=20 is the compute capability depending on your GPU device

Find your device here https://en.wikipedia.org/wiki/CUDA and input the version in accordance with your device (for example 20 for version 2.0; 61 for version 6.1; etc)

As soon as you compile the program you can run it in the following way:

$ ./VanitySearch -t 0 -gpu 1TryMe

where -gpu means you use GPU, -t 0 means that you so not use CPU (0 threads)
"./" - important part to run the program in Linux 

I installed CUDA SDK and Toolkit but
I have an error. PLEASE HELP

Code:

make gpu=1 ccap=61 all
mkdir -p obj
cd obj &&	mkdir -p GPU
cd obj &&	mkdir -p hash
g++ -DWITHGPU -m64 -mssse3 -Wno-write-strings -O2 -I. -I/usr/local/cuda-8.0/include -o obj/Base58.o -c Base58.cpp
g++ -DWITHGPU -m64 -mssse3 -Wno-write-strings -O2 -I. -I/usr/local/cuda-8.0/include -o obj/IntGroup.o -c IntGroup.cpp
g++ -DWITHGPU -m64 -mssse3 -Wno-write-strings -O2 -I. -I/usr/local/cuda-8.0/include -o obj/main.o -c main.cpp
g++ -DWITHGPU -m64 -mssse3 -Wno-write-strings -O2 -I. -I/usr/local/cuda-8.0/include -o obj/Random.o -c Random.cpp
g++ -DWITHGPU -m64 -mssse3 -Wno-write-strings -O2 -I. -I/usr/local/cuda-8.0/include -o obj/Timer.o -c Timer.cpp
g++ -DWITHGPU -m64 -mssse3 -Wno-write-strings -O2 -I. -I/usr/local/cuda-8.0/include -o obj/Int.o -c Int.cpp
g++ -DWITHGPU -m64 -mssse3 -Wno-write-strings -O2 -I. -I/usr/local/cuda-8.0/include -o obj/IntMod.o -c IntMod.cpp
g++ -DWITHGPU -m64 -mssse3 -Wno-write-strings -O2 -I. -I/usr/local/cuda-8.0/include -o obj/Point.o -c Point.cpp
g++ -DWITHGPU -m64 -mssse3 -Wno-write-strings -O2 -I. -I/usr/local/cuda-8.0/include -o obj/SECP256K1.o -c SECP256K1.cpp
g++ -DWITHGPU -m64 -mssse3 -Wno-write-strings -O2 -I. -I/usr/local/cuda-8.0/include -o obj/Vanity.o -c Vanity.cpp
g++ -DWITHGPU -m64 -mssse3 -Wno-write-strings -O2 -I. -I/usr/local/cuda-8.0/include -o obj/GPU/GPUGenerate.o -c GPU/GPUGenerate.cpp
g++ -DWITHGPU -m64 -mssse3 -Wno-write-strings -O2 -I. -I/usr/local/cuda-8.0/include -o obj/hash/ripemd160.o -c hash/ripemd160.cpp
g++ -DWITHGPU -m64 -mssse3 -Wno-write-strings -O2 -I. -I/usr/local/cuda-8.0/include -o obj/hash/sha256.o -c hash/sha256.cpp
g++ -DWITHGPU -m64 -mssse3 -Wno-write-strings -O2 -I. -I/usr/local/cuda-8.0/include -o obj/hash/sha512.o -c hash/sha512.cpp
g++ -DWITHGPU -m64 -mssse3 -Wno-write-strings -O2 -I. -I/usr/local/cuda-8.0/include -o obj/hash/ripemd160_sse.o -c hash/ripemd160_sse.cpp
g++ -DWITHGPU -m64 -mssse3 -Wno-write-strings -O2 -I. -I/usr/local/cuda-8.0/include -o obj/hash/sha256_sse.o -c hash/sha256_sse.cpp
/usr/local/cuda-8.0/bin/nvcc -maxrregcount=0 --ptxas-options=-v --compile --compiler-options -fPIC -ccbin /usr/bin/g++-4.8 -m64 -O2 -I/usr/local/cuda-8.0/include -gencode=arch=compute_61,code=sm_61 -o obj/GPU/GPUEngine.o -c GPU/GPUEngine.cu
make: /usr/local/cuda-8.0/bin/nvcc: Command not found
Makefile:63: recipe for target 'obj/GPU/GPUEngine.o' failed
make: *** [obj/GPU/GPUEngine.o] Error 127

[MrFreeDragon]

Before recompilation use the clean option first:
$ make clean

and then:
$ make gpu=0 ccap=61 all

Is your device gpu=1? If only one device, it could be gpu=0

[MrFreeDragon]

-snip-
I installed CUDA SDK and Toolkit but
I have an error. PLEASE HELP
-snip
make: /usr/local/cuda-8.0/bin/nvcc: Command not found
Makefile:63: recipe for target 'obj/GPU/GPUEngine.o' failed
make: *** [obj/GPU/GPUEngine.o] Error 127

Probably you have the different vesrion of CUDA, not 8.0

Test the version of your cuda through the command:
$ nvcc --version

It will show the version of your CUDA compiler driver.
Also you can go to /usr/local and find the folder with CUDA (cuda-8.0 or cuda-10.2)

Then open the Makefile (in Kangaroo project) and edit the folder for CUDA
You will find a line:
CUDA    = /usr/local/cuda-8.0

Change it to your cuda driver folder like this:
CUDA    = /usr/local/cuda-10.2

Save the edited Makefile and run the make command again (do not forget to run make clean as I wrote before)

It should work now.