CPU support

[backplanellc]

Any plans for CPU support?

This works great thank you. I had to do some tweaks in Makefile for my gpu architecture but that was it.

I get roughly 50mkey/s on a 1070 gtx with 600mb address file.

Thanks again for your great work.

[bill32767]

You should get more that 180mkeys/s on your card. Increase the value of -b -t -p (for instance -b 64 -t 256 -p 1024). You can keep playing around with the values of -t -p until you reach a maximum.

…

On Mon, Nov 12, 2018, 4:26 PM backplanellc ***@***.*** wrote: Any plans for CPU support? This works great thank you. I had to do some tweaks in Makefile for my gpu architecture but that was it. I get roughly 50mkey/s on a 1070 gtx with 600mb address file. Thanks again for your great work. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#71>, or mute the thread <https://github.com/notifications/unsubscribe-auth/Ae0S2b0aesg8JkMR37xBEcgcFH2UuN2bks5uuZMsgaJpZM4YZ3xz> .

[brichard19]

See #24

I'd like to do a CPU version since AMD is now offering 32-core chips.

The priority right now is GPU. OpenCL does not perform very well so we are looking into ROCm to get the AMD GPU performance up to par with Nvidia.

And bill is right, you should probably be getting more than 50 million keys per second on a GTX 1070.

[bill32767]

1080 gtx card gets about 343mkeys/s for compressed address and a little lesser for uncomprosed address On Mon, Nov 12, 2018, 4:40 PM Bill Russel Wallace < [email protected] wrote:

…

You should get more that 180mkeys/s on your card. Increase the value of -b -t -p (for instance -b 64 -t 256 -p 1024). You can keep playing around with the values of -t -p until you reach a maximum. On Mon, Nov 12, 2018, 4:26 PM backplanellc ***@***.*** wrote: > Any plans for CPU support? > > This works great thank you. I had to do some tweaks in Makefile for my > gpu architecture but that was it. > > I get roughly 50mkey/s on a 1070 gtx with 600mb address file. > > Thanks again for your great work. > > — > You are receiving this because you are subscribed to this thread. > Reply to this email directly, view it on GitHub > <#71>, or mute the thread > <https://github.com/notifications/unsubscribe-auth/Ae0S2b0aesg8JkMR37xBEcgcFH2UuN2bks5uuZMsgaJpZM4YZ3xz> > . >

[backplanellc]

After adjusting those settings as advised, i am able to get about 300MKey/s!

I'm dying to use my 20 core/40 thread system with this so i anxiously await any CPU support. I'd even be willing to donate to the cause.

-Dennis

[xvgPedro]

You get 300Mkeys/s with how many addresses in the file ?

[Coinpiet]

I have 200k addresses plus the 160 puzzle addresses and I have 5 gtx 1070 cards. But I have to open a new terminal for every card and I only get 23 mkeys per card what can I do I use the ndv oclvanitygen

[ogronome]

oclvanitygen from ndv is twice slower then bitcrack. I funded NDV's work but not using it any more. I wish I've donated to brichard :(

on single gtx1070 I get 245MK with 4k list and only compressed addresses
-b 15 -t 512 -p 4096

there is a recommendation by bill
-b 64 -t 256 -p 1024
but it makes only 176MKps for me.

[bill32767]

Yes that was an example I gave for gtx 1080i. When you reduce -b and increase -t and -p the number of keys/sec increases. I think your gtx1070 can do more than 245MK, just slightly increase -t and -p (with no. 8 or 16 or 32 or 64) until it reaches its limit.

…

On Sun, Nov 18, 2018 at 6:34 PM ogronome ***@***.***> wrote: oclvanitygen from ndv is twice slower then bitcrack. I funded NDV's work but not using it any more. I wish I've donated to brichard :( on single gtx1070 I get 245MK with 4k list and only compressed addresses -b 15 -t 512 -p 4096 there is a recommendation by bill -b 64 -t 256 -p 1024 but it makes only 176MKps for me. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#71 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/Ae0S2RWgc7uQtUh58bVNZz5FlohHldqOks5uwhjCgaJpZM4YZ3xz> .

[Coinpiet]

wow ok so i do now with my 5 GTX1070 190 Mkeys with vanitygen
can Bitcrack do both like LBC uncompressed and compressed ?
can someone help me to start the Bitcrack ?

[Coinpiet]

and i have a 42k list to search and i let every card run in another space

[backplanellc]

Got your donation addresses. Look forward to giving back soon!

[reppolice]

@backplanellc what were your settings? I only managed 90M with variations of the suggested parameters above