NVIDIA Kepler (K20) from 134MHash/s to 330MHash/s with CUDA

[psychocoder]

Bas news!!!

With K20 we can only get theoreticly 372 MHash/s.

We can run 120 bitshifts and 160 xor,and,add,rotate per clockcycle and multiprocessor. The asm code inside of the loop has 3725 and,or,.. and 162 bitshifts. K20 has 705 MHz and 13 multiprocessors. 705*13/(3725/160+162/120)=372

With GTX Titen we can get 475 MHash/s (theoreticly) because clock rate is 837 MHz and we have 14 multiprocessors.

All this calculations are without the caculation overhead before the loop and I am not really shure if 160 or 120 rotates can caculated per clock. At the moment I only get out real 330 MHash/s :-(
Keep in mind that this optimization not get a factor of two with old gpus, only with sm_35 (new Kepler GPUs)

[Edvin512]

Damn exciting, i have a 460 that needs a workout!

[cbuchner1]

Finally CUDA is getting a bit more attention from developers.

I am currently doing similar optimization work for the scrypt hashing used in Litecoin. About doubled the performance I am getting from most of my cards compared to OpenCL miners. This still sucks big time when compared to ATI cards, but it sucks a bit less than before.

With the scrypt hashing it appears much more difficult to lower the kernel's register count, as the required Salsa20/8 rounds are fairly complex beasts, also the memory-hard part of the algorithm really bangs on the memory controller.

Watch out for potential Windows binary releases in the next days. I will post into the alt cryptocurrency forum.

[Evan]

Finally CUDA is getting a bit more attention from developers.

I am currently doing similar optimization work for the scrypt hashing used in Litecoin. About doubled the performance I am getting from most of my cards compared to OpenCL miners. This still sucks big time when compared to ATI cards, but it sucks a bit less than before.

With the scrypt hashing it appears much more difficult to lower the kernel's register count, as the required Salsa20/8 rounds are fairly complex beasts, also the memory-hard part of the algorithm really bangs on the memory controller.

Watch out for potential Windows binary releases in the next days. I will post into the alt cryptocurrency forum.

for ltc or btc?

[camaro69327]

Finally CUDA is getting a bit more attention from developers.

I am currently doing similar optimization work for the scrypt hashing used in Litecoin. About doubled the performance I am getting from most of my cards compared to OpenCL miners. This still sucks big time when compared to ATI cards, but it sucks a bit less than before.

With the scrypt hashing it appears much more difficult to lower the kernel's register count, as the required Salsa20/8 rounds are fairly complex beasts, also the memory-hard part of the algorithm really bangs on the memory controller.

Watch out for potential Windows binary releases in the next days. I will post into the alt cryptocurrency forum.

for ltc or btc?

Looking forward to this.....

[Evan]

 If someone needs a testing I currently have:

GTX690
GTX680 x2
Quadro K2000 x32
Quadro K2000M x2

 (I think all of these are Keplers)

[Skami]

I will give  .25BTC for a Windows version to be created and maintained.... I have 37 nvidia cards that are BEGGING TO BE WORKED

I'll add another 0.1 BTC

[Evan]

I will give  .25BTC for a Windows version to be created and maintained.... I have 37 nvidia cards that are BEGGING TO BE WORKED

I'll add another 0.1 BTC

Added to list 

[RainmanCC]

i have compiled a windows version of the miner everyone who wants it send 0.5 btc to my adress and pm me



1ChFy7a3Lsyt3JkgG2NVWXBP4Yebs1YcVV

[Evan]

i have compiled a windows version of the miner everyone who wants it send 0.5 btc to my adress and pm me



1ChFy7a3Lsyt3JkgG2NVWXBP4Yebs1YcVV

Willing to Escrow? Screen shots? Support for future updates?

[RainmanCC]

i have compiled a windows version of the miner everyone who wants it send 0.5 btc to my adress and pm me



1ChFy7a3Lsyt3JkgG2NVWXBP4Yebs1YcVV

Willing to Escrow? Screen shots? Support for future updates?

i just compiled it its not difficult to compile it. sinsce i dont have any nvidia card i coult not test hashing.
just offering the binaries.

[Skami]

i have compiled a windows version of the miner everyone who wants it send 0.5 btc to my adress and pm me



1ChFy7a3Lsyt3JkgG2NVWXBP4Yebs1YcVV

Willing to Escrow? Screen shots? Support for future updates?

i just compiled it its not difficult to compile it. sinsce i dont have any nvidia card i coult not test hashing.
just offering the binaries.

0.5 BTC for untested bins, uhu...

[psychocoder]

Please wait and save your BTC, if I have time I add this code in a windows miner which supports CUDA and is still in maintenance. Keep in mind that the source code is open source and you must give out the source for free.
Knows everyone a Windows CUDA Miner? I am out of bitcoin mining since 2 years.

@Evan GTX690 has only sm_30 and not support rotate function :-(

[jeezy]

Knows everyone a Windows CUDA Miner?

This one? https://bitcointalk.org/index.php?topic=3486.0

[wndrbr3d]

CUDA Litecoin Miner + Source released here: https://bitcointalk.org/index.php?topic=167229.0

[GimpyPrime]

Any status updates on a compiled version for Windows?

[charliemaggot]

I created a build solution for VS2010. Pre-built binary in bin.

https://github.com/cdmackie/rpcminer-mod

Included both OpenCL and CUDA, but I can't check CUDA.

Appreciate if someone could test it out and see if it works the same.

[Evan]

Any status updates on a compiled version for Windows?

MSVCP100d.dll is missing from your computer please reinstall

[DastanX]

Any status updates on a compiled version for Windows?

MSVCP100d.dll is missing from your computer please reinstall

http://www.microsoft.com/de-de/download/details.aspx?id=14632

[GimpyPrime]

Any status updates on a compiled version for Windows?

MSVCP100d.dll is missing from your computer please reinstall

http://www.microsoft.com/de-de/download/details.aspx?id=14632

Won't let me install it since I have a newer version. I am running Windows 8 Pro / Visual Studio 2012 Ultimate.

So that probably isn't the issue.