Pollard's kangaroo ECDLP solver

[Marynarz]

I am using Visual Studio 2019 for compilation, the problem also exists on the compiled version from Jean_Luc.

[1714805727]

1714805727

[1714805727]

1714805727

[1714805727]

1714805727

[PawGo]

I am using Visual Studio 2019 for compilation, the problem also exists on the compiled version from Jean_Luc.

Great, but what is your ccap parameter?

[MrFreeDragon]

Which ccap you use for compilation?

+1

Yes, right question: ccap depends on the card, it is compute capability of the card.
For example, for Titan Z, ccap=35 (that means version 3.5)

Source: https://en.wikipedia.org/wiki/CUDA

[Marynarz]

Code:

Kangaroo v2.2
GPU #0 GeForce GTX TITAN Z (15x192 cores) (Cap 3.5) (6144.0 MB) (Multiple host t
hreads)
GPU #1 GeForce GTX TITAN Z (15x192 cores) (Cap 3.5) (6144.0 MB) (Multiple host t
hreads)

[MrFreeDragon]

Code:

Kangaroo v2.2
GPU #0 GeForce GTX TITAN Z (15x192 cores) (Cap 3.5) (6144.0 MB) (Multiple host t
hreads)
GPU #1 GeForce GTX TITAN Z (15x192 cores) (Cap 3.5) (6144.0 MB) (Multiple host t
hreads)

Your ccap looks correct.
Try to play with grid size (change used values 30x384 while find the best ones for your card):

GPU: GPU #0 GeForce GTX TITAN Z (15x192 cores) Grid(30x384)

[PawGo]

The list is recommended values is here: https://developer.nvidia.com/cuda-gpus

[Marynarz]

I set 175,256,175,256 and my image in the monitor turned black and sometimes it flashed to a delicate white color, after restart the image works, does it mean that this card is damaged?  is it simply the fault of the drivers?


My friend says that the card has been left unused for 5 years, I will start with replacing the paste with a new one because it does not fulfill its role and it is possible that the card has just overheated

If that doesn't help, I'll just give him back 

[PawGo]

Compile with ccap=35 and then see the result. Then try to set grid 30,768.

https://en.m.wikipedia.org/wiki/Thread_block_(CUDA_programming)

[Marynarz]

Compile with ccap=35 and then see the result. Then try to set grid 30,768.

https://en.m.wikipedia.org/wiki/Thread_block_(CUDA_programming)

30,768,30,768

Code:

GPUEngine: Kernel: too many resources requested for launch
GPUEngine: Kernel: too many resources requested for launch
GPUEngine: Kernel: too many resources requested for launch
GPUEngine: Kernel: too many resources requested for launch
GPUEngine: Kernel: too many resources requested for launch
GPUEngine: Kernel: too many resources requested for launch
GPUEngine: Kernel: too many resources requested for launch
GPUEngine: Kernel: too many resources requested for launch
GPUEngine: Kernel: too many resources requested for launch
GPUEngine: Kernel: too many resources requested for launch
GPUEngine: Kernel: too many resources requested for launch
GPUEngine: Kernel: too many resources requested for launch
GPUEngine: Kernel: too many resources requested for launch
GPUEngine: Kernel: too many resources requested for launch
GPUEngine: Kernel: too many resources requested for launch
GPUEngine: Kernel: too many resources requested for launch
GPUEngine: Kernel: too many resources requested for launch
GPUEngine: Kernel: too many resources requested for launch
GPUEngine: Kernel: too many resources requested for launch
GPUEngine: Kernel: too many resources requested for launch
GPUEngine: Kernel: too many resources requested for launch
GPUEngine: Kernel: too many resources requested for launch
GPUEngine: Kernel: too many resources requested for launch
GPUEngine: Kernel: too many resources requested for launch

With other settings, after manually stopping the kangaroo, a black screen appears.

[PawGo]

And on released version from githuib?

[Marynarz]

Yes, suppose the card is damaged, is it worth fixing it?

[HardwareCollector]

Hi,


Does anyone know why my speed is so small, the card is GTX Titan Z.

I am using the compiled version of Jean_Luc.

Your card is the equivalent of two underclocked GTX 780 Tis and it is performing as expected. It is based on the Kepler architecture which is very old by today's standards. 

[Marynarz]

So I will give the card back, what do you think of this card?  Gigabyte Rtx 2060 6gb OC

[Jean_Luc]

Unfortunately I am not using Kangaroo in server mode.

The wsplit option works also for normal mode. It is a very useful option to avoid large hashtable in RAM which decrease performance.
If you plan to have a large master workfile (>100GB), you can use partitioned workfiles for speeding up the merge process.
https://github.com/JeanLucPons/Kangaroo#how-to-deal-with-work-files

[WanderingPhilospher]

Anyone besides me still chasing #120?

I do have a question for the community...I've been using Etar's dp reconstructor (already compiled, can't view source code) and it works great for taming wild kangaroos.

However, is it necessary to have to know the private key (that you were searching for) in the work file or do you only need to know the pub key?

Also, does anyone know how to go about taming the wilds via a python script?

Example:

here is a line from one of my wild workfile:

Code:

26820E1D54B30FE9F13F7B99BE9CD19F4E47AAD61FD6D1F3C59874C500000000 0000000000000000000000000000000000E7DCAB65369C5F7CC15D08EEE48C67

so let's say it was for some pubkey 03......at the 120 bit range. How would I tame the wilds to say some pubkey 02.... in the same 120 bit range?

All new pubkeys that I search for will be in the same range, so just need to reconstruct the wild kangaroos to the new pubkey. But would be ok if code can reconstruct in different range too.

[dextronomous]

jean luc any new compiled version ready with the added
perf improvements still 2.2, my vs is broken a while.
can't compile now.
got a question after looong search.
while searching

Unexpected wrong collision, reset kangaroo !
Found: Td *
Found: Wd *

what these numbers behind the tame and wild. any use there for them
or suddenly 1000's dead kang's, just continue this search?

[WanderingPhilospher]

Anyone besides me still chasing #120?

I do have a question for the community...I've been using Etar's dp reconstructor (already compiled, can't view source code) and it works great for taming wild kangaroos.

However, is it necessary to have to know the private key (that you were searching for) in the work file or do you only need to know the pub key?

Also, does anyone know how to go about taming the wilds via a python script?

Example:

here is a line from one of my wild workfile:

Code:

26820E1D54B30FE9F13F7B99BE9CD19F4E47AAD61FD6D1F3C59874C500000000 0000000000000000000000000000000000E7DCAB65369C5F7CC15D08EEE48C67

so let's say it was for some pubkey 03......at the 120 bit range. How would I tame the wilds to say some pubkey 02.... in the same 120 bit range?

All new pubkeys that I search for will be in the same range, so just need to reconstruct the wild kangaroos to the new pubkey. But would be ok if code can reconstruct in different range too.

So from further reading, it doesn't look like I need to change the distances, as they will be in the same range with same jumps. I would just need to redo the private keys (right side) to equal new pub being searched for. How to do?

[bigvito19]

Anyone besides me still chasing #120?

I do have a question for the community...I've been using Etar's dp reconstructor (already compiled, can't view source code) and it works great for taming wild kangaroos.

However, is it necessary to have to know the private key (that you were searching for) in the work file or do you only need to know the pub key?

Also, does anyone know how to go about taming the wilds via a python script?

Example:

here is a line from one of my wild workfile:

Code:

26820E1D54B30FE9F13F7B99BE9CD19F4E47AAD61FD6D1F3C59874C500000000 0000000000000000000000000000000000E7DCAB65369C5F7CC15D08EEE48C67

so let's say it was for some pubkey 03......at the 120 bit range. How would I tame the wilds to say some pubkey 02.... in the same 120 bit range?

All new pubkeys that I search for will be in the same range, so just need to reconstruct the wild kangaroos to the new pubkey. But would be ok if code can reconstruct in different range too.

So from further reading, it doesn't look like I need to change the distances, as they will be in the same range with same jumps. I would just need to redo the private keys (right side) to equal new pub being searched for. How to do?

What's the link to that script you're using?

[WanderingPhilospher]

Anyone besides me still chasing #120?

I do have a question for the community...I've been using Etar's dp reconstructor (already compiled, can't view source code) and it works great for taming wild kangaroos.

However, is it necessary to have to know the private key (that you were searching for) in the work file or do you only need to know the pub key?

Also, does anyone know how to go about taming the wilds via a python script?

Example:

here is a line from one of my wild workfile:

Code:

26820E1D54B30FE9F13F7B99BE9CD19F4E47AAD61FD6D1F3C59874C500000000 0000000000000000000000000000000000E7DCAB65369C5F7CC15D08EEE48C67

so let's say it was for some pubkey 03......at the 120 bit range. How would I tame the wilds to say some pubkey 02.... in the same 120 bit range?

All new pubkeys that I search for will be in the same range, so just need to reconstruct the wild kangaroos to the new pubkey. But would be ok if code can reconstruct in different range too.

So from further reading, it doesn't look like I need to change the distances, as they will be in the same range with same jumps. I would just need to redo the private keys (right side) to equal new pub being searched for. How to do?

What's the link to that script you're using?

https://bitcointalk.org/index.php?topic=5244940.msg54666524#msg54666524

[WanderingPhilospher]

So from further reading, it doesn't look like I need to change the distances, as they will be in the same range with same jumps. I would just need to redo the private keys (right side) to equal new pub being searched for. How to do?

After more reading and researching and taking samples from Etar's reconstructor, apparently all I have to do, to tame wilds, is to subtract range from wild's priv/xpub coord.  Leave distances alone, subtract range (which is odd because it will all be same range) and bam, new tamed wilds.
Edit: actually offset previous private key - start range.
Example: if previous search was for 8F400000000 and you want to tame the wilds for that search and search for a new pubkey in same range; take 8F400000000 - start range of 80000000000 = F400000000 ; now add that to previous wilds priv/pub xcoord