BitCrack - A tool for brute-forcing private keys

[NotATether]

Well boys, I think I may have just fixed RTX 30xx compatibility  



No debug switch used - speed is awful because I'm using tiny -t -p and -b settings. I didn't want to wait several minutes to see if it worked.

That being said, there's no optimization in this program either. I guess the optimization was too aggressive and it broke things. Maybe I will inch it up a few increments to see the highest that'll work without crashing.

Code and benchmarks tomorrow - I'm dead tired now.

[pacnload]

I have a question about the commands i found here many times: -b 64 -t 128 -p 256
i have an old GTX 660 OC II with 960 Shader-Units / 80 TMUs / 24ROPs. How must i config the commans -b -t and -p with my GPU?

With the cudaInfo it will give me this info:
ID:  0
Name: GeForce GTX 660
Capability: 3.0
MP: 5
Cores: 960 (192 per MP)
Memory: 2048MB

for now i have 17.79Mkeys/s i thinks i must be more than this for normal because it have also CUDA.
And another Question if i have some adresses i want to can why i must start with 0000000000000000000000000000000000000000000000000000000000000001 is that only a count number or have it something to do with the Private key?
I Only want to use it for education that i can see if there is really a way to hack my wallet and if bitcoin is secure.

thx

[WanderingPhilospher]

I have a question about the commands i found here many times: -b 64 -t 128 -p 256
i have an old GTX 660 OC II with 960 Shader-Units / 80 TMUs / 24ROPs. How must i config the commans -b -t and -p with my GPU?

With the cudaInfo it will give me this info:
ID:  0
Name: GeForce GTX 660
Capability: 3.0
MP: 5
Cores: 960 (192 per MP)
Memory: 2048MB

for now i have 17.79Mkeys/s i thinks i must be more than this for normal because it have also CUDA.
And another Question if i have some adresses i want to can why i must start with 0000000000000000000000000000000000000000000000000000000000000001 is that only a count number or have it something to do with the Private key?
I Only want to use it for education that i can see if there is really a way to hack my wallet and if bitcoin is secure.

thx

To be honest, you will have to play with those settings to see what gives your card, and hardware, the best results.

You do not have to start with 000....0001; use the --keyspace option with a start and end range.

--keyspace 400000000:FFFFFFFF; that tells the program to start at key 400000000 and end at FFFFFFFF

Also, there is a "counting by" option and maybe that is what you saw with the 000....0001. If that's the case, that is normal/default to count keys sequentially, by 1. 

[renedx]

Well boys, I think I may have just fixed RTX 30xx compatibility  



No debug switch used - speed is awful because I'm using tiny -t -p and -b settings. I didn't want to wait several minutes to see if it worked.

That being said, there's no optimization in this program either. I guess the optimization was too aggressive and it broke things. Maybe I will inch it up a few increments to see the highest that'll work without crashing.

Code and benchmarks tomorrow - I'm dead tired now.

Amazing job sir. A hero we all needed within this chaos.
Optimize can’t be set per file base? Split the code that can’t be optimized?

Can’t wait to see this. 

[WanderingPhilospher]

If you had a quantum computer, you could decipher anything and steal all the cryptocurrency...

If if was a fifth, the whole world would be drunk

[pacnload]

has it any effect if i use more than 1 target some people use 10 some 80.000 do it need longer with more targets? with 1 target it will count 20mio keys/s with 10 targets also 20mio keys/s or dont i see anything?

and can someone told me what exactly BitCracker does? many tools are using a password database for the passphrase other only generate 12 or 24 words.

and one more question in the beginning of btc how mutch words was used for a wallet?

[WanderingPhilospher]

has it any effect if i use more than 1 target some people use 10 some 80.000 do it need longer with more targets? with 1 target it will count 20mio keys/s with 10 targets also 20mio keys/s or dont i see anything?

and can someone told me what exactly BitCracker does? many tools are using a password database for the passphrase other only generate 12 or 24 words.

and one more question in the beginning of btc how mutch words was used for a wallet?

Yes, the amount of addresses you are searching for does have an effect on speed.  But to start seeing a decline in speed, you will have to have thousands of addresses loaded in input file.

Bitcrack searches private keys, sequentially, one by one, from your start range until your end range, whatever you input as start/end ranges - using the --keyspace option.

Bitcrack has nothing to do with databases or passprhases.

[NotATether]

Amazing job sir. A hero we all needed within this chaos.
Optimize can’t be set per file base? Split the code that can’t be optimized?

Can’t wait to see this. 

GCC's .cpp files are already compiled with -O2, and here are the NVCC flags I put in the Makefile:

NVCCFLAGS=-std=c++11 --ptxas-options="-v --opt-level 0" -Xcicc -O0 --compile --compiler-options -O2 -gencode=arch=compute_${COMPUTE_CAP},code=sm_${COMPUTE_CAP}

So the C++ files are compiled with -O2 but I completely disabled cicc and ptxas optimization. This resulted in several more CUDA functions being included in the Cubin (versus about 6 when the default optimization level 3 is used).

I want to see if t behaves properly under optimization levels 1 and 2 first. At 20 MKey/s it looks like there's still more work to do though I am aware that I can't push the optimization all the way up, so speed's never going to be as fast as it should be.

(Default block thread and points settings this time)

[zahid888]

Well boys, I think I may have just fixed RTX 30xx compatibility  



No debug switch used - speed is awful because I'm using tiny -t -p and -b settings. I didn't want to wait several minutes to see if it worked.

That being said, there's no optimization in this program either. I guess the optimization was too aggressive and it broke things. Maybe I will inch it up a few increments to see the highest that'll work without crashing.

Code and benchmarks tomorrow - I'm dead tired now.

Excellent job  please give us the compiled beta version for various benchmarks. it may helps you to get things done..

[NotATether]

Excellent job  please give us the compiled beta version for various benchmarks. it may helps you to get things done..

Compiled binary? Pffft. Then the next guy won't be able to speed up Bitcrack  I'll just put the source on GitHub once I'm find the optimal level of tuning.

[zahid888]

Excellent job  please give us the compiled beta version for various benchmarks. it may helps you to get things done..

Compiled binary? Pffft. Then the next guy won't be able to speed up Bitcrack  I'll just put the source on GitHub once I'm find the optimal level of tuning.

Source code is useless for me as i am a noob in those programming Except little bit in python and can't able to compile too   

[NotATether]

Excellent job   please give us the compiled beta version for various benchmarks. it may helps you to get things done..

Compiled binary? Pffft. Then the next guy won't be able to speed up Bitcrack   I'll just put the source on GitHub once I'm find the optimal level of tuning.

Source code is useless for me as i am a noob in those programming Except little bit in python and can't able to compile too    

Whether you're using Linux or Windows the compilation process is actually straightforward. On Linux you just open a terminal and run make BUILD_CUDA=1, and I have an autodetection script that will detect the compute cap of the GPUs.

On Windows you go to Visual Studio > Build Solution, and I believe there's an environment variable for the CUDA compute cap you can set from the Properties, however I haven't compiled Windows stuff in a while so I need to recheck where it is.

In all seriousness though, it is better not to make a compiled binary because assembly for all of the CUDA compute caps have to be included in the binary and that takes a long time to make, longer than just building against one compute cap.

(Though I think I see your point about Windows compilation, VS is a pain in the ass to download, so maybe there will be a compiled Windows version, but for Linux this is infeasible unless I find out how to make an AppImage).

[zahid888]

Excellent job   please give us the compiled beta version for various benchmarks. it may helps you to get things done..

Compiled binary? Pffft. Then the next guy won't be able to speed up Bitcrack   I'll just put the source on GitHub once I'm find the optimal level of tuning.

Source code is useless for me as i am a noob in those programming Except little bit in python and can't able to compile too    

Whether you're using Linux or Windows the compilation process is actually straightforward. On Linux you just open a terminal and run make BUILD_CUDA=1, and I have an autodetection script that will detect the compute cap of the GPUs.

On Windows you go to Visual Studio > Build Solution, and I believe there's an environment variable for the CUDA compute cap you can set from the Properties, however I haven't compiled Windows stuff in a while so I need to recheck where it is.

In all seriousness though, it is better not to make a compiled binary because assembly for all of the CUDA compute caps have to be included in the binary and that takes a long time to make, longer than just building against one compute cap.

(Though I think I see your point about Windows compilation, VS is a pain in the ass to download, so maybe there will be a compiled Windows version, but for Linux this is infeasible unless I find out how to make an AppImage).

I don't care if I use it later or earlier. There are many people in our community who can compile it in the Windows version. I will use it after that, I am very happy to see that People like you share their works and help each other. This is definitely the identity of a good community. I really appreciate all of you Those who carry out such a difficult tasks. hats off 

[zahid888]

why this will happen when i run 2 windows in same time then close the other one then speed will boost for few seconds  




https://imgur.com/coMnsp7  https://imgur.com/tVHZL84

[pacnload]

has it any effect if i use more than 1 target some people use 10 some 80.000 do it need longer with more targets? with 1 target it will count 20mio keys/s with 10 targets also 20mio keys/s or dont i see anything?

and can someone told me what exactly BitCracker does? many tools are using a password database for the passphrase other only generate 12 or 24 words.

and one more question in the beginning of btc how mutch words was used for a wallet?

Yes, the amount of addresses you are searching for does have an effect on speed.  But to start seeing a decline in speed, you will have to have thousands of addresses loaded in input file.

Bitcrack searches private keys, sequentially, one by one, from your start range until your end range, whatever you input as start/end ranges - using the --keyspace option.

Bitcrack has nothing to do with databases or passprhases.

is it better to use more than 1 target? for example if i use one target and it needs 1 year to crack and then i let it work on the 2nd one and it also needs 1 year will it found 2 targets together in maybe 2 years or faster?
then it cracks from 20 targets one of them in 20 years? (only example)

and it is possible to run the script offline?

[fxsniper]

and it is possible to run the script offline?

What you mean offline?

BitCrack is work offline already no need internet connection
try unplug LAN cable or off WIFI it still work

[NotATether]

why this will happen when i run 2 windows in same time then close the other one then speed will boost for few seconds  

Well one of the two CUDA kernels that were running on the GPU have exited so naturally the other CUDA kernel should get all the clock cycles that were previously reserved for the other kernel.

is it better to use more than 1 target? for example if i use one target and it needs 1 year to crack and then i let it work on the 2nd one and it also needs 1 year will it found 2 targets together in maybe 2 years or faster?
then it cracks from 20 targets one of them in 20 years? (only example)

I'm sure there's an optimal number of targets to search for simultaneously, but Bitcrack can't calculate this number yet. Too low and you're not making full use of the GPU, while too high and you're going to make each iteration slower.

and it is possible to run the script offline?

Bitcrack doesn't have a networking feature so yes.


PS. Bear with me everyone I'm trying to push the mods on Github

[zahid888]

why this will happen when i run 2 windows in same time then close the other one then speed will boost for few seconds  

Well one of the two CUDA kernels that were running on the GPU have exited so naturally the other CUDA kernel should get all the clock cycles that were previously reserved for the other kernel.

Very satisfying and technical answer as i expect..


PS. Bear with me everyone I'm trying to push the mods on Github

[NotATether]

So it looks like that BitCrack hates any kind of cicc optimization and will immediately throw a misaligned address even on optimization level 1.

Ptxas optimization still works though, and setting it to the default level 3 triples the speed on my T4. Now before anyone gets excited about this, the speed still sucks compared to older cards (only 60MKeys/s now).  

Meanwhile I have pushed the updates to https://github.com/ZenulAbidin/BitCrack-3000

Go get 'em boys.

[coolindark]

If you had a quantum computer, you could decipher anything and steal all the cryptocurrency...

lol