Topic: EWBF's CUDA Zcash miner - page 175. (Read 2164327 times)

I am hoping for control of individual GPU core/memory clocks and powerlimit to be built into the miner, so then I can use EWBF's miner in Linux more easily and get rid of Windows and MSI Afterburner.

currently there is status quo on ZEC miners market.
EWBF's for Pascal is a little faster then Nicehash's Excavator and lot more efficient then Claymore's on AMD.
There is no need for further speed rush.
If Nicehash or Claymore will come with new faster version then, sure, EWBF will take out the rabbit from the hat ))

You did a great job and I don't even mind your 2% dev fee.   I say this sincerely since I've written several miners and I have a good sense in what is involved.   For this one the hashing is simple but tracking, finding and handling the collisions quickly isn't so easy.  

ok thanks

Are you doing anything special to cool your card? I can't get mine under 81C with stock settings, and only 660 Sol/s with Linux.

This algorithm is complex and consists of many parts, which gives many opportunities for its improvement. i think that at this moment a couple percents of improvements are still possible. And i think that after that some opportunities is still be here, and if the  developers don't see them today it absolutely does not mean that tommorow will be the same.
Well i hope i gave you the answer 

EWBF, as the developer to the miner do you feel that you've reached the limit for NVidia miners or is there still optimization room?

Anyone with 780 ti ?

I get between 330 and 350 depending on overclock and settings.

Any one with 780ti , what is the hash rate for this card?
And what is the hashrate for 980 non ti?

@EWBF_
I noticed a new option in the Nvidia control panel when installing the 381.65 drivers (seems to be in there since 378.72 or something) for my GTX 980.
It is called: "Optimize for Compute Performance"
screenshot

At the moment when I enable that setting the miner (v3.3b) gives an error message: can not run workers

Maybe making use of those optimizations could improve efficiency of the miner as it does for some other applications ?

?? why deleted?
i was about to suggest adding || /bin/true , usually rc.local is picky because the -e

Anyway, thanks for sharing the solution.

@EWBF - firstly thank you very much for your fantastic miner.

I have a big request for your miner, which I believe will make it a killer for those of us who wish to use Nvidia based Linux rigs for ZEC or ZCL mining.

If you can somehow include the setting of individual core & memory clocks and powerlimit, in either percent like MSI afterburner or just in watts - for each card in a mining rig using your software - you could probably totally corner the market for people mining with Nvidia cards in equihash on Linux.

I currently use WIN8.1 and MSI afterburner on a rig with 2 x 1070s and 2 x 1060 6GBs. I would so much rather use Linux! I tried looking at PiMP's Nvidia version, which includes your miner but I don't think they have separate OC/powerlimit control. I also had a look at the link provided in a previous post https://forum.ethereum.org/discussion/7780/gtx1070-linux-installation-and-mining-clue-goodbye-amd-welcome-nvidia-for-miners/p1 but that stuff was way over my head.

Maybe I should just accept that I have a working Win8.1 rig doing 1400sol/s for only 500W but I just HATE using Windows if I can avoid it.

Many/most of the other major miner software variants out there seem to incorporate overclocking control. Claymore has it in his miners but he doesn't yet have a CUDA capable equihash miner. He has it in his ETH/ETC miner.

I have no idea how much work this will add to your effort, so I don't know whether this is a big deal for you or not, but I think it would be awesome to have overclocking in your miner - then it would have pretty much everything!

Just yesterday I added my 4th card (Gigabyte GTX1070 G1 Gaming) and at first I had a similar problem. The rig had already 1 of those 1070s and 2 Galax GTX1060 6GB. It is an Asrock H81 Pro BTC motherboard running Windows 8.1 (unfortunately - would rather run Linux), all GPUs on the powered risers that use USB3 cables.

When I plugged in the new 1070 one of the 1060s "disappeared". First I tried a different riser because my experience tells me that 70% of problems I've had in the past (mining ETH for about a year with AMD rigs) were riser problems - most risers you buy are just badly made, untested mass manufactured shit.

Eventually I "fixed" the problem by just plugging one of the 1060s into a different PCIe slot. The Asrock H81 Pro BTC has a x1 PCIE slot above the x16 slot, so I put the riser into that slot instead of the one it was in (2nd x1 slot below the x16 slot).

I have no idea if it was just because I moved the riser and seated it in the slot properly, or if I have a bad slot.

Anyway - the point is, my suggestion is to try different PCIe slots and ALWAYS have a spare riser or three around to try because they are mostly shit and they fail over time. Actually I have found that at least 1/5 of them don't even work when you first buy them. I always order twice as many as I need.

[ POST DELETED -- I figured out the problem. PEBKAC]

Okay, thank you very much. 

Any news for update and improvement on nvidia gtx 1070 or 1080ti .

Asked this couple of days ago, he said there is an incoming update

Hi was the same problem that you had on 4 graphics cards 1060 3GB found a solution that I had 3 were connected through a riser and a 4-mouth in the slot on the board was a snap from the slot on the board and inserted into the riser and it all worked well and without crashing.