Vanitygen: Vanity bitcoin address generator/miner [v0.22]

[aarons6]

Vanitygen hasnt been updated in a while by samr7 but you can check out my fork with this fix or wyseninjas fork

https://github.com/exploitagency/vanitygen-plus

If i forgot to add this pr just open an issue and i will add it to my repo and include it in next release.

yes but this is just vanitygen not vanityminer..

maybe you can fix that one too?

[adneena]

How much secure is the vanity address than a simple bitcoin address?  I am not familiar with vanitygen .

[elebit]

How much secure is the vanity address than a simple bitcoin address?

Exactly the same. If you leak the keys you lose. But they're not easier to crack than other keys.

[ExploitAgency]

Vanitygen hasnt been updated in a while by samr7 but you can check out my fork with this fix or wyseninjas fork

https://github.com/exploitagency/vanitygen-plus

If i forgot to add this pr just open an issue and i will add it to my repo and include it in next release.

yes but this is just vanitygen not vanityminer..

maybe you can fix that one too?

didnt catch your using windows, I'll compile a windows release of oclvanityminer later on in week if i get time, click on watch repo and you'll see when it comes out

[TiiStyle]

Hello ExploitAgency,

today I talk directly to you and your vanitygen plus.
I have to report that it completely doesn't work for me.

For the start compare the difficulty with the total tries. You'll see it's much more than 10x it. No percentage is shown anymore but it does not find anything for me. That's oclvanitygen.

Code:

\Desktop\vanitygen-plus-master\windows-binary>oclvanitygen.exe -v 1Test
Prefix difficulty:            264104224 1Test
Difficulty: 264104224
Device: GeForce GTX 1080
Vendor: NVIDIA Corporation (10de)
Driver: 378.66
Profile: FULL_PROFILE
Version: OpenCL 1.2 CUDA
Max compute units: 20
Max workgroup size: 1024
Global memory: 0
Max allocation: -2147483648
OpenCL compiler flags: -DPRAGMA_UNROLL -cl-nv-verbose
Loading kernel binary 95c9219a7c5dc06bf36550d94a316dc0.oclbin
Grid size: 2560x4096
Modular inverse: 20480 threads, 512 ops each
Using OpenCL prefix matcher
GPU idle: 10.39%
GPU idle: 4.20%
GPU idle: 3.95%
GPU idle: 3.65%
GPU idle: 3.79%
GPU idle: 3.70%
GPU idle: 3.95%
GPU idle: 3.97%
GPU idle: 4.25%
GPU idle: 3.69%
GPU idle: 3.74%
GPU idle: 4.12%
GPU idle: 3.54%
GPU idle: 4.01%
GPU idle: 3.36%
GPU idle: 3.68%
GPU idle: 3.81%
GPU idle: 3.70%
GPU idle: 3.15%
GPU idle: 3.99%
GPU idle: 3.71%
GPU idle: 3.88%
GPU idle: 3.54%
GPU idle: 4.17%
GPU idle: 3.53%
GPU idle: 3.76%
GPU idle: 3.65%
GPU idle: 3.67%
GPU idle: 3.69%
GPU idle: 3.99%
GPU idle: 3.48%
GPU idle: 3.23%
GPU idle: 3.97%
GPU idle: 3.41%
GPU idle: 4.08%
GPU idle: 4.02%
GPU idle: 3.54%
GPU idle: 3.69%
GPU idle: 3.58%
GPU idle: 3.73%
GPU idle: 3.95%
[71.93 Mkey/s][total 34225520640]                                              ^C
C:\Users\

In my opinion this should return the Address for bitcoin. But it just runs and runs and never get's any result. So let's try with specifying -C BTC:

Code:

\Desktop\vanitygen-plus-master\windows-binary>oclvanitygen.exe -v -C BTC 1Test
Generating BTC Address
Prefix difficulty:            264104224 1Test
Difficulty: 264104224
Device: GeForce GTX 1080
Vendor: NVIDIA Corporation (10de)
Driver: 378.66
Profile: FULL_PROFILE
Version: OpenCL 1.2 CUDA
Max compute units: 20
Max workgroup size: 1024
Global memory: 0
Max allocation: -2147483648
OpenCL compiler flags: -DPRAGMA_UNROLL -cl-nv-verbose
Loading kernel binary 95c9219a7c5dc06bf36550d94a316dc0.oclbin
Grid size: 2560x4096
Modular inverse: 20480 threads, 512 ops each
Using OpenCL prefix matcher
GPU idle: 9.03%
GPU idle: 3.48%
GPU idle: 3.76%
GPU idle: 3.70%
GPU idle: 3.80%
GPU idle: 3.44%
GPU idle: 3.29%
GPU idle: 3.73%
GPU idle: 3.69%
GPU idle: 3.78%
GPU idle: 4.01%
GPU idle: 3.99%
[71.56 Mkey/s][total 10055843840]                                              ^C
C:\Users\r

The same. it just runs and runs and never gives me a result. I just didnt let it run SUCH far.
There is no change in the behaviour when I try to use it in -S -mode as you can see here:

Code:

\Desktop\vanitygen-plus-master\windows-binary>oclvanitygen.exe -v -C BTC -S 1Test
Generating BTC Address
Prefix difficulty:            264104224 1Test
Difficulty: 264104224
Device: GeForce GTX 1080
Vendor: NVIDIA Corporation (10de)
Driver: 378.66
Profile: FULL_PROFILE
Version: OpenCL 1.2 CUDA
Max compute units: 20
Max workgroup size: 1024
Global memory: 0
Max allocation: -2147483648
OpenCL compiler flags: -cl-nv-verbose
Compiling kernel, can take minutes...done!
Build log:

Grid size: 2560x4096
Modular inverse: 20480 threads, 512 ops each
Using OpenCL prefix matcher
GPU idle: 8.10%
GPU idle: 2.39%
GPU idle: 1.97%
GPU idle: 2.25%
GPU idle: 2.40%
GPU idle: 2.15%
GPU idle: 2.34%
GPU idle: 2.06%
GPU idle: 2.25%
GPU idle: 2.38%
GPU idle: 2.35%
GPU idle: 2.26%
GPU idle: 2.18%
GPU idle: 2.20%
GPU idle: 2.11%
GPU idle: 2.34%
GPU idle: 2.14%
[49.16 Mkey/s][total 10013900800]                                              ^C
C:\Users\

When I run this in hardwareverificationmode I get a very different result from what I see in normal oclvanitygen.
It permanently shows the kernel mismatches like this:

Code:

Mismatch for kernel 0, offset 40296 (125,296)
Column X: 31CEB67EFF2D377A49B6B09D5DBD4CA92F832BEC8D1691F4853CFA7B380DC63C
Column Y: 1E3FFDE8A016587B6FCD022076070259C7BFD4CBC49FAC33C238588C0522D00E
Expect X: F125E2550C9060404342F3860ED028E2C617515B254B255FF0C7DA2AB08D0766
Device X: 1350A2EF2062962C937C8E02FCE850DDB87478BA80AB7C1ED321E569C41FD0CD
Expect Y: 96C115DEE545855896AC7576B2715611F5633C1BEB424F1C9A10C6E7315BD3DA
Device Y: 9DDBB285A7BCF93E75A397FCF62E67017AB506ED281A2853E515BF39120CD664
Mismatch for kernel 0, offset 40297 (125,297)
Column X: CF91EBCA82C23FC07BFD031FD96604ECC4034C3126C44E1BDCB93C130C9D98E6
Column Y: C748D92EBFB12E6F0445CC18553BE1D131BEDA0254141D2057EB5828A1F2CE2A
Expect X: D3B822CAE1A78FD16D8B422A500E971E6578883DF64BE677B42F51007628AB7B
Device X: 7C79E0EC4420BF184FDA6D99FFDA23E353A96A4C14E51D641E06350FA7A06D4F
Expect Y: DAA8482C288E600FEDEA67FB9F9D3417F3E1A2F1A0AE60E1715C6467BCDF22EC
Device Y: 9E70812DE454DEA31A9C596310565BF543C3BE79DCC11227FF7901ECBF85BB27
Mismatch for kernel 0, offset 40298 (125,298)
Column X: 5610FC1D5283768F3CB3DF522BD7E71ADD8106BE6B03B4FE104642F7B8287894
Column Y: 4A215ADE32E58853CB0A2B19ED0A4E1E497D90A8A3CDC0F29502F1475E5B4E13
Expect X: B30D8F578470BCFE2E0CD738B8091EE8399B67417AF316FDF1BD2DA5DA3C10F5
Device X: 2FFB371784F515B115CE32C8309D043452AF51B0E230FA368610FEFA1B5E49C6
Expect Y: A1F3E01E91B5B3763060ED072A7FCD4E4527128734D0EDD009E8BA1FE6C8A70B
Device Y: 244DF1508FED252F53B9967F791025B7223656BEEE239E17805392D525666916

If I shall let this run longer or give you more information about hardwareverificationmode just post or pm me.

With all this I can conclusively say: oclvanitygen.exe DOES NOT WORK FOR ME AT ALL.

And like in my last post you clearly can identify the same overflow in max. allocation like with the normal oclvanitygen.exe. It's a overflow of integer32.

For now I think this is all what makes oclvanitygen.exe unusable for me.
I don't have ANY problems when I use the normal vanitygen.exe. It works perfect for me. It is JUST oclvanitygen.exe.
So I don't can talk of a fixed version of oclvanitygen.exe here. :/
But I hope I can help you to fix this issue so that it really works fine for all coins for all users on all hardware.

Just let me know what I can do to help.

~Style

[ExploitAgency]

Do oclvanitygen.exe --help

If it isnt v1.41 then download and try the latest release
https://github.com/exploitagency/vanitygen-plus/releases

So the original oclvanitygen does not work for you either right?

Thats funny I have a GTX 1070 and my fork works great for me.  Albeit testing on Linux.  The Windows version has only been tested with Intel CPU OpenCL by me in a virtual machine.  Others seemed to have success as well.

Thanks for the feedback hopefully I can do some more testing and figure out the issue.

[TiiStyle]

Ok that's very strange. I simply pulled the directory to my computer and that does not work at all o.O
The windows binaries seem to be broken.
After separately downloading the executables separately from the newest release worked great.
But I still have the allocation problem.
Why can't you fix that? Just switch the variable-type to long instead of int?? Or isn't it that simple? (As told, I never learned C/C++ I can't read the code myself.)

I uploaded the tool I mentioned in my first post to github including executable and full source code for Microsoft Visual Studio 2015 written in C# as a WPF-program.
The program is VERY small and VERY simple and is just a "wrapper" around the keyconv file providing a simplified GUI showing only the most necessary Options for generating and combining keypairs.
The keyconv file needs to be placed in the same directory unless you did not change the directory yourself in the code.

Maybe you want to include that in your next release or maybe not, but I just wanted to give you the posibility.
You may test this all out in one of your vm's.

Would be nice if you could give some short feedback if you like this or not.

Here the Link:
https://github.com/TiiStyle/VanAddrKeyPairUtility


Have a nice day!

~Style

[TiiStyle]

I forgot to say:
No, the original oclvanitygen works for me.

[ExploitAgency]

So my fork of oclvanitygen works for you now?

It is just oclvanityminer thats broke for you?(which I don't include in my Windows release)

[TiiStyle]

yes, now your oclvanitygen works aswell, it seems if I simply clone the repository I get the wrong windows executables. I had to manually get the executables from the latest release as you said. Pulling the repository did not work.
And yes it was only the oclvanitygen that was broken, the normal vanitygen worked even from the pulled/cloned repository instantly and the keyconv worked also as far I can say, which is not far because I don't use altcoins atm.^^

[ExploitAgency]

Ok thats strange

I'll look into it.  Thanks for the feedback.

Can you try to delete and reclone the repo and see if it works this time?

I couldn't replicate this in my Windows VM albeit Intel CPU OpenCL... wondering if it was just some sort of file corruption.

[TiiStyle]

I did that but no it does not work.
I have the same bugs I posted before with a new downloaded one.

[ExploitAgency]

Ok so release zip file works, repo clone doesnt?

I'll look into the miner.  Was in a rush to get to work what int were you saying to convert to a long.  Didnt have time to read logs.

[TiiStyle]

Erm no.
I don't know if cloning works, there's just a combined button on the website and it's always a zip download. I'm not using a git utility so I did not "clone" the repository even if I used that term. Next time I'm more specific.

And the int to long that I mean is the maximum memory allocation.

[ExploitAgency]

I have a releases page i linked.  What method worked?  Downloading from the releases page which doesnt download any source just the binaries or what?

[TiiStyle]

Downloading from the releases page which doesnt download any source just the binaries worked.

[ExploitAgency]

Excellent thanks.  Will try to get out a copy of oclvanityminer soon for you to test.  Pm me your email address if you dont mind or i can contact you here.

[TiiStyle]

It was aarons6 who asked for a copy of vanityminer but I can test that for you for sure aswell, no problem you're welcome.
I pm you.

[aarons6]

It was Aarsons6 who asked for a copy of vanityminer but I can test that for you for sure aswell, no problem you're welcome.
I pm you.

yeah but you were saying it doesnt work, and i tested it and it does indeed not work.. so maybe he can fix it.

[TiiStyle]

ExploitAgency,

you have to make another improvement.
You should replace the parameter "-1" with a parameter "-X" or so (choose an unused letter). That I can specify something like 5 hits and then stop.
Currently I only have the possibility of choosing 1 or permanent run. That's a bit...  dumb..
So why not simply make it right. no parameter = 1, with parameter = the choosen amount, with parameter 0 = permanent run. With this construction you also could get rid of the -k parameter by fusing it into this option and provide a maximum of possible runs using an unsigned long or int.

I think that's a very practical but also important improvement for the user control of the program.

Think about it.


~ Style