Topic: Donations to open source a GPU Protoshares miner (PTS) (Read 2427 times)

Yup - see the cudaminer thread in this forum.  Or just download a recent CudaMiner: The changes are all in there as of 12-18 (and Christian made some further improvements after).  For a technical summary of the changes, see:  http://da-data.blogspot.com/2013/12/inside-better-cuda-based-scrypt-miner.html

i havent heard about the kepler tuning for kepler-based core, is there any link?

thanks.

(btw, by "previous release", I meant the Kepler-tuned code for Scrypt mining... this PTS thing is new.  Though, as I said in another thread earlier, I'm pretty convinced someone is running a GPU in the wild already.)

Yeah.  I don't do any overlapping of kernel execution and CPU, or use multiple kernels.  During the execution of one of the phases the memory is locked up 100% and the compute cores are bored, and in another phase, the cores are locked up and the memory is bored, etc.

But, as yvg1900 convinced me, it's more healthy to let these things evolve a little bit gradually anyway.

really?? 

I know I posted earlier, but I seem to be unable to let this GPU on protoshares thing go.  Any "memory-hard, GPU resistant" problem is too good to give up.  So - and this explains why I'm up so late:

I have a working prototype, and I'm convinced that it's not going to utterly disrupt the ecosystem - the GPU is better than CPU, but not so much so that a good CPU won't compete.  At least, until other people improve upon my code.

I don't want to go the route of long-term maintaining this or releasing closed-source builds that siphon off a portion of the mining income, though I like that model a lot.  Instead, I'd like to ask for tips _now_ in advance of releasing it open source.  I contacted the Protoshares company about their bounty, but heard nothing.

Is there a kickstarter-equivalent in bitcoin, or a recommended best way to do this?

Here are the numbers I can put together.  Note that there's still work to be done to make this thing friendly.  Right now it's about in the same state as the Nvidia miner code I released for scrypt/LTC.  Gunky to compile, not platform friendly, etc.

(EDIT:  I typed c/s when I meant c/m.  Argh, fixed.  I'm too used to scrypt coins in kh/s.)

Dual GTX 690s:  2400 c/m at 115W at 10% CPU.
650Ti:  330-340 c/m at 1% CPU.
Macbook Pro with GT 650M:  200 c/m.
Tesla K20c:  ~800c/m.

As with my previous release, this may be Kepler-only, but I think the algorithm will work on Fermi.  Part of why I want to open source it is to let the community build upon it.  I know that there's more performance that could be tuned into it - probably on the order of 30% but maybe 200%.

  -Dave