Topic: CCminer(SP-MOD) Modded NVIDIA Maxwell / Pascal kernels. - page 669. (Read 2347641 times)

windows 10 is okish too for those card with the right drivers?

The TDP is 150watt and is the default in the bios for these cards. (gainward reference cooler cards)
For the G1 gaming the tdp is 240Watt so a 1000W will probobobly be too little.
I mine Etherum. + decred (with -i 26 in my private dual miner (300-350MHASH decred for free)

windows 8.1

i think sp has some rig with 980ti, maybe he pulled it off with six in one rig

but are you using windows 7 or 10 and maybe you can try to run then with one psu only if you have the chance, try every thing

the only way to fix this kind of problem is to proceed with exclusion

with the 980ti - only 5 can run ... even though it IS possible with 6 - i could not get it to run ... all with powered high quality risers - all powered on the motherboard as well ... so there is sufficient power all round ... 2 x corsair ax1200i psu ... asrock h81btcpro ... 16gb hi performance ram ... ( cant go any higher with this motherboard due to bios restrictions apparently ) ...

example of the processing using spmod80 with no oc on 5 cards - qaurk approx 142mh - x11 approx 61mh ...

a sixth card would be very beneficial - but just refuses to work ...

all the lower cards ( gigabyte 750ti oc lp ) run on similar systems - all on corsair rm1000 psu - all 6 cards at a time ...

go figure ...

#crysx

perfect, tnx for the info it's the same setup that i have(not the gpu planning to reach six that why i asked), but what system are your unning, windows 7?

btw 1kw for six gpu isn't a bit far-stretched are those card heavily tdp-limited? like 50-55%?

I am running 6 gtx 970's stable on a 1000W gold rated psu (EVGA). The 970 cards have 2x 6 pin power connectors. H81BTC board with no powered risers, but with 2 extra 4pin molex connectors attached to the board.

You can try to enable x1 speed of the pciE slots in the Bios. Then it will unable to pull the full power through x16 slot.

among the rig you have with nvidia, you can confirm that you are running 6 970 or 80 and not only small nvidia like 750?

Simple, basically you need to have at least as much RAM plus pagefile combined as the total memory that'll be used for mining across all the cards.

So if you mine something which barely uses any memory it doesn't matter but if you have cards with more memory and you mine memory heavy algos like neoscrypt and/or with high intensity (which increases GPU memory usage) then you need more RAM+pagefile just for a second for the miner to start. It should have been long fixed in ccminer by not initializing all the cards at the same time but one after the other which should also solve the dualcore CPU issue.


I think that's two bad advices in one sentence but to each of his own I guess. (Though the GV-N75TOC-2GI cards are faulty by design and will die in a few months).

I have also been running a rig with 6 gpu's for about a year and no issues.

On Windows 7 64 bit using a 700 watt power supply and not a top of the line. All cards are on powered usb risers.

One thing to keep in mind that might help. Do not put more than 2 cards on each daisy chain. Those things are not designed to handle a lot of wattage through them. There have been many reports in the past of them melting at the molex connector and even catching fire.

Windows works and the drivers work so it's time to move on.

Start from the begining and try to get one card stable when mining. If you can't get any card working
with different combinations of slots and risers you have a bad mobo or PSU.

If you can get one card working, try a second card and repeat the process with two cards. Make note of which
combinations of slots, GPUs, and risers are known to work.

If you run in to probems along the way swap the suspect card and riser with a known good combo from another
slot. Then observe if the problem changes in any way Is the known good card/riser eorking in its new slot?
Is the suspect card/riser still not working in the slot you know is good?

Keep doing that until you've isolated the suspect card/riser combo, a suspect mobo slot or just the number of cards.
If it's the number of cards it could be the mobo or PSU. Even if it's rated high enough it may not be working properly.

If you've isolated to a PCIe slot, it's definitely the mobo. It's up to you what you want to do about it.

If you've isolated it to a GPU/riser combo them swap the riser with another card to isolate the problem.

And you're done.

Hi Longsnowsm, my advice:

Return those Gigabyte cards with 6pin pciex gimmick power connector and buy 750ti cards without additional power connector. You can also use unpowered risers if you choose to switch to H81pro BTC.

Your best bet would be Gainward GTX750TI golden sample or Palit dual (around 245khs neoscrypt with opensource djm miner max stable OC).

Running my two 750GTX rigs since april 2014 without any issues.
My conf:
Corsair CS750M
Asrock H81pro BTC
CPU Intel Pentium G3220 Haswell+stock cooler
1x4GB DDR3 1333Mhz ram(win7/64 basic)
sata 2.5" 5400rpm HDD with bad sectors  
6x GTX 750TI 2Gb Palit dual/Gainward GS on "old" 352 drivers
6x unpowered cable risers  

Bottom line is: or mobo do not recognize card or card does not get enough juice to start regularly. Due to: riser, mobo, driver, etc.

Out of desperation I did "upgrade" to Win 10 to ensure that Win 7 wasn't the issue with the 6 gpu's and Win 10 did not fix the issue I was seeing with the Nvidia cards.  I can say that I have this exact same mobo, cpu, memory, ssd, psu combo working fine with 6 AMD gpu's for a different rig.  So I know this combo is capable of running 6 gpu's.  So I am not sure what the real issue is... I thought it might be drivers issues.  The other rig that I spoke of is running Linux, so there is a different OS in the mix. 

When I discovered one of the GPU's wasn't being seen by the Nvidia software I then pulled it out of the mix, and I could finally get the system to boot reliably and behave normally, but still when running ccminer it would fail and crash saying cannot validate to CPU. 

So if your hoping that Win 10 is the magic bullet like I was I wouldn't hold my breath.  The only reason I went this route with windows on this rig was the intention to buy SP_ private release.  Otherwise I would have stuck with Linux.  So now if I run into any Windows related issues with the AMD cards I will just reconfig the rig with Linux and call it a day. 

I hope someone is able to figure out what the issue is.  Nvidia cards are more efficient than the AMD cards and I would love to have a rig that takes me in a more efficient direction for the long haul.

AMD has to workaround for more than 4 cards on w 7, but W 10 works with 6 cards ( I'll try 7 soon ). Nvidia works 6 cards on w 7 but some strange things from time to time happen.  I have several rigs (with 6 nvidia ) and some need 8+GB RAM some not, and last fun is AMD mobo: 6 cards work only one way, if I change position(slots) of 2 cards- only 4 start ( AMD cards, 380 ). To me looks like mobo switches off PCIe in case of smallest irregularity. Usually that is fixed with trying risers (I have them all types) and to me was good idea to short pins that signals "it is GPU" to mobo  ( here:   http://www.gobitgo.com/articles/1001/How-To-Correctly-Use-and-Install-PCI-E-Riser-Cables/  )

does amd have the same issue? and did you tried another os, maybe windows 7 is the problem there

i would probably be in the same boat soon so i want to know

I don't think you understand what someones life is like. Everyone has different skillsets, lets look for a moment at you or me completely changing. Spending a couple years learning a different set of skills is not all that different that starting a giant company. I could more easily start a new company then learn to program to the same level of quality as current releases are at, let alone at yours. For me that's a complete waste of time because my expertise are in other areas. I pay people instead to do that work because that's not where my expertise is. I can't pay for what people don't sell or don't even advertise openly that that's what they sell.

Most people don't know the market exists because you don't present your wares and there is no market to actually buy from due to back room trading.

Not just that, but what market there is isn't a free market. You don't advertise prices and even then you sell them to whomever you want. You've even mentioned you only sell to a handful of people (which are big players in the crypto game, which further pushes down all the other miners) keeping it completely exclusive. This pushes other miners out of the market due to the nature of mining.

There is no other game in town for AMD. The current alternative is SP and friends, which is why I switched to Nvidia. Ethereum completely fucked up your stranglehold on AMD mining, but no one could see that coming.

Thanks!  .

Try release 68.

ccminer --algo=nist5 -o stratum+tcp://pwr.suprnova.cc:4242  -u user.worker -p pass

I want mining powercoin with my Nvidia GTX 970 .

what do I download and how do I configure the bat ?

i use https://pwr.suprnova.cc .

Thanks for any help .