Thanks man I'll give that a go, for the cuda code, I have cuda 10.1 and it's built for 9.2 and I can't be arsed compiling on windows as it's a pain lol
Edit: Actually just built it there and was much quicker and easier than I thought, just have to change all references to 9.2 over to 10.1. The problem now is with cuBitCrack I'm only getting 7Mkey/s..
I suggest you try again on the compiled version, which is ultimately on 9.2. You don't need to install another version of CUDA ... just put in the same folder as cuBitCrack.exe two .dll files from CUDA 9.2. I have prepared them for you:
==>
http://bitchain.pl/100btc/cuda92.zip <==
Just unpack them and you will run out! To two .dll files I also added a BitCrack .exe which I personally use with these files. I don't remember if it is a pure version of "pikachu" or I compiled it for SM70, anyway for the lazy it may be useful to someone.
I had my time 1080Ti (AORUS XTREME EDITION) and I am firmly convinced that cuBitCrack is better. Check the results on the original version 9.2 and let me know ...
I wish I could remember these settings. However, I have 100% confidence in the properties for each of the setting values in terms of the BEST setting ... and here information for EVERYONE concerned:
{-b} - is nothing but the CU value of the "Compute Units" card. To be honest, I have not found another source that gives this value (including websites that create benchmarks and other things with cards) than one application, which I had to use once, noticing this fact :-) it is Bitcoin Password Recovery which you can download from here: https : //www.thegrideon.com/bitcoin-password-recovery.html
The values appear immediately after starting in the start log. The program requires no installation.
In this example screen the detected card is GT630 and the value CU = 2 which means that -b best for this card would be equal to -b 2.
Someone recently wrote according to my assessment of the issue -b with the difference that the setting should be twice the number of CU. In the middle, compliance with the truth aroused my interest and I checked it ... I was very surprised to see how high the results are at different settings -p ... nevertheless - finally none of the levels I set to doubled - b did not give me the result 1590Mkeys [and out of curiosity - I devoted 10 hours to it today trying various options], so finally I think that no multiplication of values will be better here.
{-p} - you need to set the maximum value that does not end this way:
[2019-08-24.22: 09: 17] [Info] 10.0%
[2019-08-24.22: 09: 18] [Info] 20.0%
[2019-08-24.22: 09: 19] [Info] 30.0%
[2019-08-24.22: 09: 20] [Info] 40.0%
[2019-08-24.22: 09: 20] [Info] 50.0%
[2019-08-24.22: 09: 20] [Info] 60.0%
[2019-08-24.22: 09: 20] [Info] 70.0%
[2019-08-24.22: 09: 21] [Info] 80.0%
[2019-08-24.22: 09: 21] [Info] 90.0%
[2019-08-24.22: 09: 21] [Info] 100.0%
[2019-08-24.22: 09: 21] [Info] Done
[2019-08-24.22: 09: 22] [Info] Error: too many resources requested for launch
For most cards it will be "256", although in the case of the 2080Ti certainly "512" as in my case. More than 512 I have not met so that any card can let go. On 1080Ti you have to check because I don't remember. No in-between values will be better. If you set 512 and the loading process does not finish as in the log above - you can be sure that this is the best value.
{-p} - and here you have a patience test, because this value requires your own invention and luck. Someone wrote earlier that he doesn't see the goal of setting more than 1024 ... I see ... My setting exceeds the number "2800" and I'm still not sure if this is the best value for me, because with the 1600 configuration I only have 100Mkey / s lower than above 2800 [where the very middle between these values gives much less]. Here you can only set the max which the card will support and try to follow the chosen pattern. Interestingly, changing this value by 1-2 units can bring a lot .... but I do not intend to test over two thousand combinations, because I waste my time :-) remember - patience will lead you to the peak value. I pulled the V100 to 1590Mkeys / s - where someone once said they would pull 1500. What is important for you and anyone who will try to find the best setting - during the tests you need to set the card to factory clocki ... No overclocking. Clean drivers and you're flying with it!
*********************************
try here
--keyspace 13C96A3742F64906:7CCE5EFDACCF6808
but above people write their results, limit yourself
Haha. This is very interesting advice for #62. Good luck for you