NVIDIA GTX 1080 ti GPUs


#551

Last GPU (4) problem


#552

Rafa… this isn’t really conclusive. For any given PEC readout, you’re going to see some variation in card solution rate. I personally see differentials of 30-35 sol/s between my top card and my lowest card for a given data set. This kind of variability (per your screenshot, ~3%) is to be expected, even among cards of the same make/model.

Have you seen more extreme differentials?


#553

Last time I turn off mining server and restart GPU4 mining only (580) sol I restart 3x time and work ok.


#554

These are Galax Hof watercooled cards


#555

Is this with a third-party water cooler? Thought they just came with 3 fans


#556

There’s a version that comes with the water block pre-installed. you then need to attach it to a custom loop


#557

Ok interesting thanks, I have the corsair hydro just for ease of setup but seems like the Galax version are a decent step above…assuming the total cost for the custom loop, etc might come in a bit more expensive though and in which case it would be interesting to see a ROI comparison.


#558

hi everyone,

have you spotted any difference in performance using different drivers versions?
same for win7/win10?


#559

With the current difficulty oscillating around 7million I don’t think you will get ROI any time soon.


#560

New miners entering now are going to have to rely almost exclusively on Zcash valuation for ROI. 7000 H/s barely gets you 2.5 ZEC/month at current average difficulty, and God knows it’s probably not going to get better for any appreciable amount of time.


#561

Its ok when zcash reach $1000 everything will be fine :slight_smile:


#562

If you expect zcash at $1000 - you better spend money on buying zcash instead of buying mining equipment. Buying zcash is more profitable than mining in any conditions. Actually the greater prices you would expect - the more profitable to buy, not to mine. Because the difficulty will increase instantly.


#563

That’s true with zec at this point in time. Luckily got in with my 1080ti a few months ago, if I really got stuck now I could sell the card and get the couple of hundred still owing on the card and would have a few dollars extra left over.


#564

Depends on where you sell it. If you sell it on eBay then expect at least 20% cut.


#565

Now is good


#566

True although at this stage I’d only need to get $240usd to break even. I’m not in a situation where I have to sell it but it’s a bit of peace of mind neither the less.


#567

People forget that mining in general has always been slim margins. Equihash and other memory hard POW have tried to stem the tide, but as we all saw - people still found a way to throw more hash power per dollar invested and per kWh. 13gpus per mobo, $4 pcie risers from china, cheap refurbished server PSUs, etc etc etc.

To me, the threat of the algorithm changing to render an ASIC useless is slightly hollow. If you developed one, youd probably keep it a secret and reap the benefits of efficiency. Or sell it to whale miners under NDA.

Monero’s cryptonight POW is a good example of something they think is difficult to tackle with an asic, but actually it’s not. 2MB of on chip fast memory isnt that big of a deal anymore. Nor is making it low latency. They also wouldnt dare increase the scratchpad size beyond 2MB lest they alienate their users/miners that CPU mine (cache sizes are not increases all that fast). That’s the while reason for picking that size - make the scratchpad about the same size as L3 cache. So yeah. ASICs will be used for cryptonight. I’d say in about 6 months max. But you’ll never see it, except in hashrates.


#568

There is already people making private GPUs 100x faster than 1080Ti on ZEC. Like you said, people will only notice on the increased hashrates. Many people joining now buying normal GPUs will get seriously hurt.


#569

Hey, how did you made your miner to show “gpu power usage” and “sol/w efficency” in same window?


#570

just add “–pec” inside of your .bat file.