bitcointalk AS A WHOLE on the other hand is significant - there are some large farm operators that hang out here, and probably some we don't know about that lurk rather than post.

 Bad word of mouth has killed or help killed more than a few ASIC makers (can YOU say KNC? Butterfly Labs? Just for 2 examples where BAD PUBLICITY ABOUT ISSUES BADLY HANDLED cost the company big time).

 I've never seen a GTX 1070 in the 310 range, especially not recently, from ANY source.
 AMD cards HAVE been dropping noticeably under 250 the past couple weeks, as low as $209 on Newegg on Black Friday/Cyber Monday weekend though they bounced back up a little since then (the LOWEST 1070 on Newegg is $399 for comparison).
 
 RX series cards DO need a BIOS modification to achieve top hashrates, but 22-24 with a STOCK card is common with just overclocking/undervolting.

Power supplies don't CARE what orientation you put them in - the only component in a PC that USED to care was hard drives, MANY years back, but they eventually went to "balanced actuator" designs for speed of access reasons that also made them pretty much immune to orientation issues.

 If you look at cases, "mounted with one side up" isn't real common but certainly isn't unknown in towers, is VERY common in desktop cases, and is THE NORM in 3u and 4u rack mount cases.

Based on the pics in the thread, it looked like a combination of hot dry weather, VERY poor wire management, and probably something got overloaded - coupled with NO fire prevention/retardation measures used.

Monero soaks enough resources out of a GPU that I doubt dual-mining is a practical option.

Only reason it works on ETH is that ETH is VERY strongly memory-limited while the cores tend to loaf....

I would drill holes in those brackets, put them on the END of the power supply, and screw the PS into the bracket for a SOLID "hold it in place" - optimally with a small air space between the PS and the wood, as some PS DO get pretty hot on that side.

 I NORMALLY mount my PS on their sides though, for space reasons.

Litecoin pricing has increased enough that the old Gridseed GC3355 based miners MIGHT be profitable again - if your cost of electric is SUPER low.
The Moonlander and Moonlander 2 ARE profitable, though the original needs semi-cheap electric to be making money.

 ROI is a different issue - but if you already HAVE the miners and they are profitable, might as well use them.

FE is basically a Vega RX 64 with a different cooling solution and slightly lower default clocks.

Mine Monero with it and treat it as a Vega 64 *SHOULD* work.

For ETH, the AMD RX 470/570/480/580 are the kings as far as current new cards go on hash/$ and they are very very close on efficiency to anything else on hash/watt.

The only reason folks started building NVidia-based ETH mining rigs was that those RX series cards hit crazy pricing for a while due to demand GREATLY exceeding supply to the point they cost MORE than NVidia 1070 cards - and started driving NVidia 1060 and 1070 card pricing up as well.

 At current "getting back towards MSRP" pricing on the AMD cards, no reason to waste time looking at anything else (possible exception for the GTX 1060).

 In used cards, if you can get R9 290/290x/390/390x cards for enough less than those RX cards AND if you have low cost power, they MIGHT be a better choice - they'll pull a hair higher hashrate but they eat a TON more power to do so.

 1070 ti pulling 480 sol/s isn't pushing hard at all.

 70C though is high on temp for 115 watts, especially with 80% fan - SOMETHING is blocking airflow to the cards or the room is VERY VERY hot.
 Those Aero blower cards shouldn't have THAT bad of cooling, though they would tend to run a LITTLE warmer than good fan-cooled cards unless they're in a case.

 My EVGA and Zotac 1070 ti cards in my "new shelf/rack" first rig only run 62-63C at 50% fan in a room that is 86 F measured right above the "room fan" that blows air into that rig.


 It's POSSIBLE that you got a batch of flaky cards - but if they're working for video playing it sounds more like you're pushing them too hard, probably on the memory clock.

The EVGA 2-connector cables don't have enough reach to do that, but from a power draw standpoint it would be fine.
You could use an "extender" cable to make it work - or just plan on powering the risers from MOLEX instead.

Most of my mining machines are on XUbuntu LINUX - and as I make progress on my "rationalization" project they ALL should end up there (possible exception for a while for the Vega 56 system since it REQUIRES bleeding Win10 to be able to run the miners at full hashrate and LINUX drivers for Vega are STILL borked).

 You will need C14 to C13 power cords for your supplies.
 You shouldn't need fancy high-power cords - at 220 your EVGA power supplies (even the 1600) would draw less than 9 amps so 16AWG is plenty. 18AWG would be fine for anything under a 1000 watt unit.
 They should all run on 220 without issue - I've not seen ANY gold-rated power supply that was incapable of running on 220, but look at the "input" info sticker that should be right next to the AC power socket on the power supply for details to be sure.
 Some older supplies needed you to switch between 110 and 220 manually, but I've not seen one like that in years.

 The outlet you show is correct for the plug the PDU is described as using.

One of the large Chinese bitcoin farms had a fire that burnt out most of their facility.
I suspect most fires involving miners have been small though - probably only burnt part of the board on one miner due to power overdraw at the connectors.

 $480 is pretty high for a GTX 1070 ti - quite a few of them at the $469 price point or under on Newegg most of the time (5 right now) a couple of which are at $449 (MSI blower model and Zotac mini).
 $500 at current pricing is LOW for a 1080 there - exactly ONE listed at $499 (ASUS blower model) one at $509 one at $519 one at $529 and most at $539 or higher.
 $20 is NOT a normal difference between comparable models of 1070 ti and 1080 cards.

 Your "80% power limit" on the 1080 is for a MSI GAMING model that has a 240 "default" TDP - which is 60 watts HIGHER than almost any other 1080 ti default TDP - and the shown 180 watt draw in your picture is actually 75% TDP for THAT card, while being 100% TDP for most 1080 models.
 80% TDP on most 1080 and 1070 ti cards is 144 watts NOT 180 watts.

 As it turns out, in my testing on 2 1070 ti models to date (EVGA SC and Zotac Mini) vs 4 different 1080 models (Gigabyte 2-fan and 3-fan Windforce models, EVGA SC, and Zotac Mini), the most efficient point for ALL of the cards is right about 105 watts and they all hash at right about 460 sol/s at that point - for basically the SAME hashrate and efficiency on both sides of the question.
 The 1080 picks up more hashrate as you bump the power draw up, but not a ton faster, so it loses less efficiency as you get closer to the rated 180 watt TDP on MOST models of both cards - but even at 180 watts it's not 10% faster than a 1070 ti while SAME MODEL versions of each card tends to be 10% or more difference in price, making the ROI longer on the 1080 vs the 1070 ti *IF YOU RUN THEM AT THE SAME POWER LEVEL OR EFFICIENCY LEVEL*.

 As I already said though, if you can get a "same model" 1080 on sale for close enough to the price of the 1070 ti version, it's worthwhile - but they do NOT normally sell for a $20 difference in price for SAME MODEL cards, more commonly $50 or more price difference.

 And yes, if your power cost is low enough best efficiency is NOT best profitablity - unless you're pushing the limits on available power.


 4.7 is NOT average efficiency for a 1070 ti - that was a peak figure out of his video and as I recall was at the 60% TDP setting (106 watts) not 110 watts.
 Average is closer to 4.4-4.5 range at the BEST efficiency point, though it's not much different at the 110 watt level (still 4.3-4.4 ballpark on average in my testing).

 Trying to figure an "average" out of DSTM though is a lot tougher than EWBF, as DSTM reported hashrates bounce around a lot more from what I've seen in my testing.






 

My "less than 300 watts" IS THE ENTIRE SYSTEM as I have already stated that's "AT THE WALL" on a system that has NOT been optimised for power consumption.
That figure is NOT "only considering the GPU load".
It's also a very NON-OPTIMAL setup in general - ONE non-optimized GPU in an old high-power-draw SINGLE CARD system.
It would be nice if you would pay attention and quit trying to prop up your strawman "THE GPU eats 250+ watts to mine with" argument with lies while ignoring presented FACTS.

The FACT is that 6 KW will power a lot closer to 30 Vega GPUs including total rig consumption - and possibly MORE than that - than your TOTALLY BS 17 claim.
If that "130 watt" TDP figure I've seen claimed on one "how to build it" site is correct, that would put a 4 card Vega rig right close to 600 watts - for *40* Vega GPUs in 6 KW including total system draw - while still respecting the widely reported issue of trying to get more than 4 Vegas mining Monero in a single system at high hashrates.

Does the 835 have a GPU that's useable for anything other than driving a tiny display?
Does it support OpenCL or CUDA?
If not, it IS useless for mining.

Comparing floating point capabilities is MEANINGLESS for cryptocoin work, as floating point does NOTHING to help cryptocoin work in any way shape or form.
There is a reason any RX 470 blows away the R9 280x on most cryptocoin work, even though the R9 280x does more than DOUBLE the FP64 work of any RX 470 - Floating Point is MEANINGLESS.
Again you insist on ignoring FACTS to try to prop up a strawman argument.


 Nvidia cards in general are more efficient - but there are exceptions, they're a tossup at best on ETH mining, they just aren't competative at all in Monero mining where Vega blows away the hashrate of ANY Nvidia card by a wide margin and has better hash/watt when both sides are well optimized, and outside of cryptocoin MINING as such the Vega 56 can very close to MATCH the keyrate of the GTX 1080 ti on Distributed.net RC5-72 work while soaking a little less power (much better on keyrate/$ when you can find the Vega 56 at close to MSRP, tossup on keyrate/watt).

 In many cases though, even where Nvidia cards are more efficient, they are NOT "much more efficient".

 Generally yes, as the PCI-E spec rates an 8-pin power connector for 150 watts and the wiring for such a connector is normally capable of handling a little more than that safely - *IF* the 1070 in question is something like the EVGA SC 1070 or the Gigabyte ITX that is only rated for 150 watts.
 If you are running the cards at a "lower than spec" TDP where the power you are SETTING the card to is 150 watts or less, you should also be OK - though sometimes when Windows crashes it will cause Afterburner to lose it's settings on one or more card and allows the card to reset itself to FULL TDP.
 180 watt cards should be OK - most PCI-E connectors have wiring that will handle that safely even on "single connector" cables, and any DOUBLE connector cable (the kind that have both an 8 pin (or 6+2) AND a 6-pin on the same cable) should be safe to at least 225 watts.

 That MSI Gaming monster card with the dual connector - might be a bit marginal but SHOULD be ok, it might be rated for 240 watts but pushing a 1070 hard enough to actually DRAW that much power for more than a short peak timeframe takes a ton of work.

 Technically, the actual CONNECTOR is rated for 288 watts, but it's rare that the wiring FEEDING the connector is rated for quite that much.

 In most cases, you can run 2 risers from a single Molex power string - but check it after the rig has been on for 10 minutes or so and make sure the wires aren't running hot. Not all power supply makers use the same gauge wiring for their "peripheral" connectors, and while the CONNNECTORS are rated for a bit over 150 watts it's rare for any device connected to them to draw more than 50 or so (and most hard drives draw less than 1 AMP or under 12 watts from the 12VDC line, though some high-end FANS can draw close to 4 amps at 12VDC).

At current pricing (for the USA anyway), the 1080 is inferior hands down to the 1070 ti - 10% higher cost for the SAME MODEL card, but it doesn't hash 10% better and when both cards are run at their most efficient point they give almost IDENTICAL hashrate (though also almost identical efficiency).

If you can get a 1080 for $20 or less more than a "same model" 1070 ti, it's definitely worth THAT much more - but not at 10% more which is the NORM - and on rare cases of a 1080 on sale for the SAME price as a 1070 ti same-model, the 1080 wins hands-down due to higher resale value and ability to be pushed to higher hashrates if you don't need maximum efficiency.


 1080 ti - is a debate.
 It does NOT match hash/$ or hash/watt vs the 1080 or 1070 ti, be it run for efficiency OR for high hashrate - but it's close, and if you factor in total cost of a SYSTEM for the same hashrate it gets a lot closer on hash/$
 The "higher density" that allows for the use of fewer machines is a benefit that to some outweighs the small loss of efficiency.


 Most algorithms don't gain any significant benefit from a wider bus, or else the Fury and Vega would be hands-down across the board DOMINANT players in mining on everything.
 Keep in mind that the GTX 1060 has an even narrower bus, yet is quite efficient on most algorithms.


 When I talk about a "same model" card, I'm talking comparing (as an example) a EVGA SC 1070 to an EVGA SC 1070 ti to a EVGA SC 1080.
 Any other comparison is NOT a fair one, though sometimes there is no "same model" to compare TO between the various GPUs.

It's plenty - lots of miners that run their farms using USB "pen drives" to boot/run from.
 Other than ETH and clones using older software that built the DAG file on the CPU/HD, and the "HD mining" coins like BURST/Storj/SIA, nothing in mining uses the HD very much if at all once it has started actually mining or needs a lot of bandwidth.

 Some of my early mining rigs ran on EIDE interface hard drives - which were a LOT slower than anything SATA.

The Seasonic X-series specifically designs their GPU and CPU cables to be interchangeable ON THE POWER SUPPLY END.

 They're NOT interchangeable on the other end.

 This also applies to the Snow Silent (platinum) series - and the cables from those are interchangeable with X-series cables as well.