Might have a better option for you (my EVGA SC 1070ti showed up a bit early and I just got done doing some testing on it). https://bitcointalk.org/index.php?topic=2324651.msg23996225#msg23996225 Push the shortie in slot 1 for clearance (I'd go with a Zotac mini like we previously discussed over the MSI, those MSI cards don't seem to cool well), push 7 x 1070ti at 104 watts each in the other 7 slots, for over 3500 sol at 850ish watts total draw. I think Zotac is supposedly working on a "mini" version of the 1070ti, might be able to go 8 of them in that case at the cost of a few more watts. I ordered 2 of these https://www.evga.com/products/product.aspx?pn=08G-P4-5678-KRthey may be the best card for the rig. Better cooling for tight quarters, makes sense to me. Still need something short for that first slot, unfortunately.
|
|
|
2k gridcoin is enough to stake at least monthly, if they fix the beacon bug they seem to have added in 3.6.3 whils supposedly FIXING a beacon bug.
|
|
|
Got them running around 666 sol/s or 10200 sol/s total now as I dropped the powerlimits to reduce heat and noise. I have mostly founder's editions as well as a few gigabyte gaming OCs, 2 gigabyte turbos and two pny 1080ti blowers. Overclocks don't really have so much affect on hashrate with equihash I believe, but everything is running at 150+ core and 450+ memory.
Equihash can be wierdly picky about overclocks, sometimes MORE overclock gives less hashrate. It doesn't seem to be super picky about memory clocks though. My GTX 1080 ti Aorus pulls 620 sol/s at 150 watts (60% tdp) for a bit over 4.0 efficiency - which seems to be it's most efficient point or very close to it's most efficient point. GTX 1070 ti (I got my EVGA SC today) at 60% can manage quite a bit better than 4 - more like 4.4 - at +200 core +700 memory in afterburner (but the card UNDERCLOCKS the memory by default a little for some reason). Up side of the 1080 ti though is rig density and fewer cards to deal with, even if it IS a little less efficient than some of the alternatives.
|
|
|
Could anyone tell me if the EWBF miner will be ok with the new Nvidia GTX 1070 Ti if ill have any problems with this miner and the card since its new not sure how that works, was going to try a few for a rig and i use this miner does anyone know if there will be any problems or consdering its like the same as the 1070 if it will be ok, would really apprecaite the help thank you.
There is no way to know for 100% sure before it's tried - but there is also no reason to believe that this miner won't work well on the 1070 ti, since there are no major changes in the basic architecture of the GPU that would cause issues. Ask again in about a week, when folks that pre-ordered 1070ti start getting their cards and have a chance to put them in machines. Thank you appreciate it. And just to be even more certain, since my EVGA SC 1070 ti showed up early (this morning, wasn't DUE 'till next Monday) - it works VERY WELL INDEED with EBWF. +200 core, +700 memory, 60% TDP in Afterburner (the card UNDERCLOCKS the memory by default a bit, 7600ish "effective - memory setting bumped it up to 8996). Bit over 100 watts, over 450 sol/s and appx. 4.4 efficiency on the current NVidia 388.13 drivers under Windows 7. I had a full review in one of the 1070ti threads, linked to it in 1 or 2 other 1070 ti related threads, and also linked to it in the current Philllma thread. I think there may still be some room to play with on this card, especially at the high end of it's range, never DID manage to get it past 66C at about 72% fan no matter how hard I pushed it (but I didn't bump TDP up past 100% in my testing).
|
|
|
There is no reason for it to NEED mining-specific cards to run 19 cards. Just tons of power connections, risers, etc.
720 is kinda low for a 1080 ti at 100% tdp, unless perhaps they are FE cards (which have relatively poor cooling solutions and don't boost as high as pretty much any aftermarket cards do). I get that much out of my Aorus cards at 200 watts (80% TDP) +100 core +100 memory and very close on my Gigabyte Windforce and EVGA SC cards at the same settings. Which 1080ti cards do you have? Have you done any setting testing on core clock and memory clock on them?
|
|
|
I got an early delivery today - just spent an hour and some testing my new EVGA SC 1070 ti. https://bitcointalk.org/index.php?topic=2324651.msg23996225#msg23996225 In my personal experience, this is a new "efficiency" king for Equihash - and does AMAZINGLY well at fairly low TDP settings, though I don't understand why they underclocked the RAM by default. Definitely a better option for Equihash (ZEC/ZEN/ZCL and such) than the 1070 across the board, and better than the 1080 except for "raw highest-end max push hashrate" once you optimise the settings. This is NOT a good choice for ETH/ETC and such mining, it's memory limited and NO faster than the 1070. It was a THEORY (that available facts support) that the 1070 ti GPU might be 1080 GPUs with one failed "compute unit" - but it's a more widely mentioned theory that the 1070 ti is a direct response to the RX Vega 56 being slightly better than the 1070 while being cheaper than the 1080. (The theories are not incompatable - it could be BOTH, at least for the initial production run). Which reminds me of one test I forgot to do on my 1070 ti - the Dnet/Moo! test - where it should beat the 1080 and STOMP the 1070 on keyrate per $. (later add). Yup, 1070ti is a legit 6 gigakey/sec card there while the 1070 is a little better than 4 and the 1080 is a little better than 6 - and the same price range Vega 56 is capable of well over 8 while also staying under 100% TDP. Sorry Nvidia, no price-range performance crown for you on THAT project.
|
|
|
So I decided to push the six 1070's
1- 90 msi itx
2-110 evga hybrid 3-110 evga hybrid 4-110 evga hybrid 5-110 msi seahawk 6-110 msi seahawk
7- 90 evga founders 8- 90 evga founders
this would be 820 for gpus and 30 for system total 850 watts at about 3200 sols
Might have a better option for you (my EVGA SC 1070ti showed up a bit early and I just got done doing some testing on it). https://bitcointalk.org/index.php?topic=2324651.msg23996225#msg23996225 Push the shortie in slot 1 for clearance (I'd go with a Zotac mini like we previously discussed over the MSI, those MSI cards don't seem to cool well), push 7 x 1070ti at 104 watts each in the other 7 slots, for over 3500 sol at 850ish watts total draw. I think Zotac is supposedly working on a "mini" version of the 1070ti, might be able to go 8 of them in that case at the cost of a few more watts.
|
|
|
we tried to place order for 50 pcs on the day it announced (hyperbitshop) for sale. Can't get it. plus the cost and delivery is per bitmain style. nothing innovative there.
Thing that gets me, why did they pick a "primary distributor" in AUSTRALIA of all places? Do they actually WANT to do any retail sales at all?
|
|
|
I think that the price is little high for the overall improvements of the card, i think we agree on that ? But, on other hand we will get cheaper 1070, yaay
Cheaper 1070s are happening because folks aren't buying them as "can't get RX series mining cards, they're the next best option" any more, not due to the 1070 ti introduction. If anything, the 1070ti seems to be driving 1080 prices up a hair. Mine isn't due 'till Monday (EVGA SC model) - will play around with it some then post data (like I did in a couple threads about the Vega 56 I just got today that is happily Mooing bigtime).
|
|
|
I want to get a few 1070s. Any advice on where to buy (newegg ? ) and which 1070 is the "best "?
Wind force or gtx 1070 Sc gaming ?
I'd go with the EVGA - I just had to fix fans on *2* of my Windforce 1070 cards that froze up this week (they ARE a bit over a year old IIRC), but my EVGA 1070 SC cards are older and having no issues to date. One of the last 2 cards I bought was an EVGA 1080 ti SC - just got it today and installed it in the B250 rig, along with a Gigabyte I was shifting from an old single-card rig (the other was a Sapphire Vega 56 I have set up to do lots of MOOing). The Gigabyte Extreme and Aorus cards appear to use ball bearing fans, but I'm avoiding the rest since I'm moving away from non-riser rigs.
|
|
|
For a 30 amp circuit, a L6-30 is the most common used outlet, but feeding a Power Distribution Unit to break it down to 2 or 3 smaller circuits.
You COULD have 4 or so NEMA 6-15 outlets put in on that circuit instead - just don't push 15 amps through ALL of them at one time.
|
|
|
Define "considerable".
To some folks, $1000 a month would be "considerable" - to others, it would be "pocket change".
|
|
|
Spondoolies wasn't competative with Bitmain because they build their miners to be configurable, NOT because their chips were too expensive or not comparable.
If they had been SMART, they would have settled for fewer sales and NOT tried to compete on price - the same route Caanan is taking with the Avalon - except that the SP20 when set for the same hashrate as the S5 was almost identical efficiency level and ran a hair cooler.
The SP20 compared to the S5 was a night and day difference in many respects - the SP20 could get down semi-close to the high end of S7 efficiency range (at a cost in hashrate), or it could be tuned to do 50% or so more hashrate than the S5 (at a cost in efficiency), and that tuning also allowed it to handle high temps a LOT better than the S5 could. You could also STACK SP20s, they were a lot more compact design in actual usage (longer, but lower in height and no cables/lack of a top made the EFFECTIVE height quite a bit lower, and almost identical width). The SP20 was also overall a much more PROFESSIONAL design - it's STILL a more professional design than the S7 or S9 or even the Avalon 741.
Configurability and professional design came at a cost though.
|
|
|
I was reading the specs of EWBF, the software used to mine Zcash and BTG on Nvidia GPUs as I wasn't familiar with it. I read that there is a 2% developer fee. Is it true? This plus the 1% pool fee makes a 3% dent on your gains.
It defaults to 2% for a dev fee but unlike Claymore you can set it lower (or higher) if you prefer to. I default to setting my miners up with 0.2% - which still nets the dev a few $$ a month from my fairly small farm.
|
|
|
ETH is not the end-all and be-all for mining. Being no better on ONE ALGORITHM does not make the card "bad to mine".
|
|
|
But then there is the important question - IS THE CARD WORKING?
Miss-ID isn't all that critical if the card is doing the job.
|
|
|
Yeah, I saw that about 20 minutes after my last post. I also note that hyperbitshop.io seems to be out of stock on everything right now....
|
|
|
That would be correct - single PCI-E, single Molex, are both OK to power a riser, 2 x SATA via an adapter should be OK.
Mine right now are all connected via PCI-E, but I'm starting to think about changing on future rigs to move to Molex for powering at least some of the risers.
|
|
|
Not a silly idea at all my man.
Bitmain Antminer L3+ and BW.com BW-L21 are 28nm node process ASIC chip miners.
Innosilicon A4+ LTC Master is a 14nm node process ASIC chip miner.
I sit corrected on the BW.com BW-L21, per their page on it. I begin to understand how the A4+ is managing to get the kind of hashrate and efficiency Innosilicon has been posting videos about recently. The question now becomes "can Bitmain make a RELIABLE miner on the 14/16nm node", and "will BW.com ever get end user sales going or set up to sell through a distributor to end users sometime"....
|
|
|
I meant no difference regarding Ethash. Honestly, didn't still have time to check Equihash results, but I guess it should be like Dovaogedys already pointed.
Given that AMD "good ETH mining cards" have dropped back under $300, why is anyone looking at a 1070 to mine ETH with anyway? 1070 ti turning out to be equal to the 1070 is not exactly a big deal - and it was already known that the 1070 was probably memory-limited on ETH anyway, so it's not a shock.
|
|
|
|