Hmm, I lose - I NEVER expected the S9 to ship this early.

 On the otehr hand, that price is WAY high.
 Looks like I'm going with other options for a while, waiting for the inevetible price drop once competition shows up.


 The REAL loser might be Bitfury - they SHOULD have been first out of the gate on this generation, but now looks like they're going to be second at best....

 Depends - for Ethereum, the GTX 950 is almost identical efficiency on a per-card basis which gives it a HUGE edge on a per-system basis, 3x the hashrate PER CARD at about 3.5-4x the power (system with 1xGTX 950 and 2xGTX 960 hashing 31+ MH (had almost identical hashrate when I had all 3 of my GTX 950 in it), 74 watts on the 950 75 on each 960).

 The bigger GTXs fall off some from that, more on a $/hash basis than on a W/hash basis though, though past the GTX 970 by report the hashrate doesn't get much if ANY better.

 I never got around to trying X11 on them, the difficulty on that has gone up quite a bit to the point I doubt I'd be break-even at my current electric rate - ASIC deployment seems to be progressing fairly quickly the last month.

 8-(




 BTW sp_ you DO know about Claymore's dual-miner for ETH and ... Decred? I forget what the other coin is called exactly.

 LK-1401 (BW.COM) has reportedly been in production and in miners for a month or two already - just no better efficiency than the Bitmain S7 / BM1385.

 BW.COM has also been paying off on the contracts for those miners for a while now, and has been showing the hashrate that they probably have the miners they claim to have.


 Keep in mind the BW.COM might be new, but LKEtc that part-owns BW.COM is NOT new.


 Bitfury demonstrated their 14nm chip in a series of videos almost 2 months back.


 I also suspect that Innosilicon DID get some "partners" as a result of that offering they made last year - and I'd BET I can name one of those partners.


 Will they meet July to ship miners to non-partner end-users? Dunno. Will they ship before fall? That I'd BET on.


 Oops, just noticed one more thing. *S9* ships mid-June.

 So much for your "no Bitcoin chip has 14nm" claim.

Link to this alleged S9 controller board?

(Edit)

 Nevermind, found it - only 4 connectors populated, but I do see the holes for the other 16. I DO wonder why they populated 4 connectors though if they only need 3 for the hashboards?

There is no such thing as a "low profile" Nano. They are ALL full height double-slot-width cards.

 Short is NOT "loiw profile" - the Nano IS a fairly short in LENGTH card.


http://www.newegg.com/Product/Product.aspx?Item=N82E16814125680

 THAT is a low profile card.

 Once I get my planned move done, I'll be in a land of VERY VERY cheap electricity.

 8-)

RX480 looks interesting, if only due to the HUGE memory bandwidth increase - which *hopefully* would translate to good Ethereum mining hashrate.

 Availability is going to be an issue - NVidia is already having availability issues with the GTX 1080, a "mainstream" card like the RX 480 is likely to have even worse such issue....

And now ETH is heading back up again.

 The whole "Dollar collapse" FUD crap seems to have originated from a comment Putzkin (that Russie ex-KGB type) made - and like MOST of his BS should have been taken with a couple tons of salt.

I have never seen an ATX-type power supply that even THOUGHT about using 3-phase power directly.

 I doubt they exist - 3-phase isn't needed 'till you get into the MANY-KILOWATT range on electrical gea, a standard common 20amp 220 circuit is good for 4 KW after all and higher power 220 circuits aren't UNcommon (ref some electric stoves, some commercial electric driers, and RV-type connections, for a few examples of 30+ amp 220 plug-in circuits).

 For that matter, mobile homes used to COMMONLY use a 50 amp 220v plug-in connection (often set up 2 on one pole feeding from one common meter with a 100 amp feedline, the second circuit usually intended for a utility outbuilding, workshop, etc). Those hookups are no longer permitted under the US NEC, but they're allowed on a "grandfathered" clause for existing setups in most locations.

That's WAY too hot.

You need to fix the overload before insulation starts melting/burning and you get a SERIOUS issue.

As I'm going to need a bit more in funds than I have right now to finance my move, I'm afraid it doesn't really matter what it's at by then - I'll be selling some.

 Optimally it will keep heading up or at least hold at close to where it's at now....

I remember seeing something posted somewhere about "pushed back to mid 2017" - but I might be misremembering the "mid" part.
Definitely 2017 sometime unless it gets delayed more.

I tried it on one of my machines last week.

 It was a ROYAL PITA to get working at all on Ethereum - the interface is VERY non-intuitive tryng to get it to work with GPUs.

 It didn't handle multi-GPU very well at all, 2 of my 3 ran at about 80% of what I see out of qtminer, the 3'd ran at less than 20% - and NOTHING I could find to change made any difference.


 I can't say how this works on anything else, but as an Ethereum miner it is NOT READY FOR RELEASE and should be considered ALPHA-level software - it's not good enough IMO to count as Beta test software.

Some 280x have 3GB of ram (there might be 4GB cards out there too, but I don't have any of THOSE) - those will benefit eventually from changing that 64 bit pointer to =1 instead of =0, when the DAG file gets too big to fit in 2GB.

Keep in mind though, it's the energy use of your ENTIRE SYSTEM that matters, NOT just the use of the card itself.

 Many cards that are quite a bit less efficient AS A CARD than the NVidia Maxwell cards like the 750Ti end up being much more efficient miners when you figure in the overall SYSTEM energy cost.

 Also, it depends a lot on what you try to mine - a $100 750Ti only manages 3.5MH or so on Ethereum, while a less-than-$300 R9 380 manages closer to 20. The total SYSTEM power of that 380-based system will probably be a lower per MH than the 750ti based system even if the 750ti itself is more efficient.



 I've got a couple systems with 750Ti cards in them - only reason is that they are LOW PROFILE systems and the 750ti is the most powerfull card I CAN fit in them (AMD HD 7750 was better for some stuff, but they're very hard to find any more and for some reason AMD never released a comparable R7-designated product).

I run 3 CPU cores on other stuff on my 3 A10-7860K rigs (has 4 CPU cores) with no hashrate drop on Ethereum, *UNLESS* I'm folding on the 7860's GPU as well - then I have to drop to 2 cores doing other stuff or the Eth hashrate gets a noticeable hit some of the time.



 I'm not sure what the firepro D700 equates out to in a "consumer" AMD card, but as long as it has at least 2 GB RAM and 512+ stream units, it SHOULD be profitable.

It depends - if it's a 2GB 750Ti, it's actually profitable at Ethereum but only pulls 3.5MH ballpark (about 40 cents/.day right now) *BUT* the card itself eats less than 20 watts doing so - the REST of your system will eat more than that.

 If it's a 1GB 750Ti, it CAN'T mine Ethereum, you have to look at other options.

Try finding a store with one IN STOCK.

 8-(



 Someone had a review unit they posted some ETH hashrates on - it was..... VERY VERY underwhelming, a 970 outhashes it.

 Not sure how much of that is "new drivers not optimised for OpenCL or CUDA compute yet" and how much of it is "archetecture changes made it work WORSE on Ethereum than NVidia cards already do"....

EVGA PrecisionX seems more appropriate on the stated EVGA 970 cards - and can do the same thing.

I'm sure Afterburner would WORK too though.

 

If you have to COMPILE the miner from source, in some cases you have to install the SDK.

 As a MINIMUM for every miner software I've seen under LINUX, you need some of the header files that are normally included in the SDK (but there are alternate sources for those header files in some cases).


 Installing the SDK should NEVER be a "negative" or "cause issues" with a miner program, unless it's enough to bump you free drive space into "too small" range - but if your drive it THAT marginal on space, you should be using a bigger drive anyway.