Author

Topic: Mac Pro 5,1 with Titan Xp - trying to mine Ethereum (Read 1406 times)

newbie
Activity: 5
Merit: 0
Thanks for the love?  The Titan Xp is $1,200 from Nvidia direct.  Yes there are cheaper cards available that are nearly as fast but I wanted to reward Nvidia for their support of the Mac Community.
legendary
Activity: 4256
Merit: 8551
'The right to privacy matters'
Hi, all.

I have been digging through the interwebs for several days now without luck. I rolled the dice and bought an Nvidia Titan Xp for my Mac Pro 5,1 tower running Sierra 10.12.4 with the latest beta Nvidia Web Driver (presently 378.05.05f02) and CUDA (8.0.81) drivers.  Now I can at least load Unreal Engine whereas my stock ATI Radeon 5870 would not even allow that.  I've also compiled and run the C++ CUDA code from Nvidia and the card clearly excels at parallel processing tasks. https://devblogs.nvidia.com/parallelforall/even-easier-introduction-cuda/

I'm presently running the latest Howebrew build of cpp-ethereum but the --with-gpu-mining option is not available so the -G option is not available on the command line.  As such the cpp-ethereum is running on CPU only (as evidenced by 2,200%CPU usage) and I'm only getting 0.7-2.2 MH/s on Nanopool with stratum enabled.  I tried compiling Genoil's CUDA mod for cpp-ethereum under Parallels 12 (https://github.com/Genoil/cpp-ethereum).  Unfortunately neither Windows nor Ubuntu VMs access the Titan Xp directly with Nvidia drivers and instead used Parallel's GPU driver so I get 'no supported GPU' type errors.  I've also tried compiling Genoil's C++ code via CMake on MacOS and got all the way up the linking steps when I got a bunch of errors relating to processor architecture deficiencies.  Next I will try to compile Genoil's mod in XCode but I'm not holding my breath and expect the same linking or other errors.  I also read somewhere that CMake might not be using the clang++ compiler so that's probably worth trying to switch from clang to clang++.

I feel like the Xp would make quick work of Ethereum mining but the impediments are presently winning.  Have any other tenacious Mac users solved this dilemma?

Thanks in advance!

=G=

Wow so smart and so dumb at the same time.  This is with love not hate.

I had the same Mac you did.

As a long time Mac user I just built a Linux pc. Using smOS

How much did the Titan Xp set you back ?
newbie
Activity: 5
Merit: 0
Hi, all.

I have been digging through the interwebs for several days now without luck. I rolled the dice and bought an Nvidia Titan Xp for my Mac Pro 5,1 tower running Sierra 10.12.4 with the latest beta Nvidia Web Driver (presently 378.05.05f02) and CUDA (8.0.81) drivers.  Now I can at least load Unreal Engine whereas my stock ATI Radeon 5870 would not even allow that.  I've also compiled and run the C++ CUDA code from Nvidia and the card clearly excels at parallel processing tasks. https://devblogs.nvidia.com/parallelforall/even-easier-introduction-cuda/

I'm presently running the latest Howebrew build of cpp-ethereum but the --with-gpu-mining option is not available so the -G option is not available on the command line.  As such the cpp-ethereum is running on CPU only (as evidenced by 2,200%CPU usage) and I'm only getting 0.7-2.2 MH/s on Nanopool with stratum enabled.  I tried compiling Genoil's CUDA mod for cpp-ethereum under Parallels 12 (https://github.com/Genoil/cpp-ethereum).  Unfortunately neither Windows nor Ubuntu VMs access the Titan Xp directly with Nvidia drivers and instead used Parallel's GPU driver so I get 'no supported GPU' type errors.  I've also tried compiling Genoil's C++ code via CMake on MacOS and got all the way up the linking steps when I got a bunch of errors relating to processor architecture deficiencies.  Next I will try to compile Genoil's mod in XCode but I'm not holding my breath and expect the same linking or other errors.  I also read somewhere that CMake might not be using the clang++ compiler so that's probably worth trying to switch from clang to clang++.

I feel like the Xp would make quick work of Ethereum mining but the impediments are presently winning.  Have any other tenacious Mac users solved this dilemma?

Thanks in advance!

=G=
Jump to: