First for AMD but not first period - Intel Skylark series is 14 or 16 nm.

 I also don't see the new AMD CPUs helping on GPU mining at all, though they might get competative with recent Intel CPUs on CPU mining.

 As a literal "home miner" for now, I'll just point out that some of US also pay very low electric costs - and have a lot less overhead than the "big farms" so, though not so much on a "per hash" basis.
 Up side, my "labor cost" is free, for some ways of accounting....

 There were 2 folks that wrote A2 software upgrades (This is NOT Firmware, it's SOFTware).

 I'm sure one or both will be looking at A4 software "fixes" as soon as they get an A4.

It has made sense most of the time 'cause the price keeps dropping, though there have been a couple of small spikes where it made sense to hold on to it for a short while.

 I'm STILL boggled that it ever was even close to 1 Bitcoin for a price, much less the "very very EARLY adopters" pricing in the 3 digits for a very short bit.

 I'm amazed the price ever WAS above 1 BTC.
 

They help a little with heat dissipation on systems with only one GPU in the system, or where the GPUs are WIDELY spaced apart.

 They're a MINUS in a system that's crowded as they reduce airflow.

 They offer some physical protection to the card and make it less likely you will damage the card via ESD while you are handling the card.


 Caps shorting out can be VERY dangerous.
 Had one go the length of a LONG room (40 feet or a bit more) and imbed the wire in a concrete block wall once when I was a power supply test tech - pure LUCK it didn't hit anyone on the way, it could easily have put someone in a hospital if it had hit them in the wrong place.
 That cap was somewhat larger than anything on a current GPU card (600v 10uf or some such as I recall, 2 inches long almost + axial leads and almost half inch diameter) but the principle is the same....

Google Drive appears to have exceeded your file transfer limit.

Mega.nz linked worked (a FIRST for me, usually mega just hangs but I recently moved so might have been an issue with my old ISP).


 Most of my Window machines are NVidia but I DID get it working (for a brief while) on an AMD A10-7890k GPU side - at a whopping almost 3 sol/s (I didn't expect much, but that seems a BIT low.).
 Next step - try it out on one of my win machines with a HD 7750.

 (edit) which achieved slightly LESS than 1 sol/s while the A10-7860k in that machine was at about 1.5 sol/s, definitely not worth it on that old of a card and it seems VERY ODD that the APU in that machine with the same core count, slightly slower clock speed and MUCH inferior memory speed was faster. I guess OpenCL is a lot better optimised for the newer GCN versions? (/edit)


 All of my serious AMD GPU cards are on LINUX boxes though.
 Since they're all running Ubuntu 14.04 and Catalyst 15.12 they *should* be prime candidates to run the LINUX version of this "soon".




 Depending on your rig, ETH might BE more profitable right now - but that might change too, as the ZCash slow ramp keeps climbing.
 flypool (and presumably nicehash as well) released updated miners today too, the flypool one added about 10% to my 1070 cards on my dual + HD7750 hybrid rigs and the trio rig, and added about 20% to my 950s and 30+% to my 960s on my older rigs. It as JUST enough to get me to swing them all back over to ZCash right now.

HD 6xxx and earlier, and a the low-number HD 7xxx series (7600 and lower IIRC) are Terrascale, NOT GCN, and require completely different coding.

 Same issue affects the A10 5800 and older APUs.

 HD 7750, 78xx, and 79xx were all GCN to my specific knowlage.

 The 15.12 drivers PREdate the release of the RX 4xx series, and won't work at all with those cards.
 Unfortunately, the newer 16.xx drivers do NOT work real well or real reliably with older cards.
 I'd recommend NOT mixing those 2 generations of card, tends to have low stability even when you CAN get it to work at all.

 The equihash algorythm that ZEC uses also seems to be a lot more CPU/main memory intensive than ETH was, I've found it's a lot more stable to NOT run CPU and GPU mining on the same ZEC rig, or at least don't try to mine on ALL the cores of a multi-core CPU that also has multiple GPUs installed in the system.

 Reminds me a lot of building Folding@Home rigs, where you NEED to reserve some CPU or it just don't work for beans (which may be why most of the Folding rigs I moved over to ZEC have been stable).



 GTX 650 might not have enough RAM to run ZEC, or it might not have a up-to-date enough CUDA version. That's a pretty old card.

One issue with trying to write a calculator for ZCash (ZEC) right now is the "block reward ramp up period" has almost a month left to go.

 Sol(utions)/s pretty much *IS* the same as hash/sec for other algorythms.

 May I offer some Navy Sludge? 9-)

 Amazing level of irony there, newbie.

There never was an ETH ASIC - just a propriatary multi-GPU rig design - even asicminermarket never claimed "ASIC" for their ETH rig.

A4 and A900 are legitimate products - but they are NOT from asicminermarket - at BEST that site is reselling the units.


 On the other hand, asicminermarket *IS* frequently cited as a scam site - just because the products they claim to be selling are legit does not make THEM automatically legit.

 Nope, it WASTES a lot of money.
 Big mining farms have the people and proven design in place to just clone their existing rigs INEXPENSIVELY (with an occasional update to that proven design when new major hardware shows up).

 Odd, SHA256 is the oldest algo in the cryptocoin world yet it's still in widespread use and Bitcoin that uses it certainly isn't dead.

 ETH profitability jumped a fair bit when ZEC went live.


 In the absence of a decent LINUX/AMD miner, I've still got my 3 primary rigs mining ETH at the moment.
 (No, I WON'T support xxternal and his propriatary pool/miner BS - and given the MASSIVE number of comments about held or non-existant payouts, I'm VERY HAPPY with my stance).


 I prefer ethmine over ethpool, but the only major difference between the two is the payout method, and I don't have THAT massive of hashrate available.

 Wrong direction, per what folks are saying - should have gone UP a few mv.

 Undervolting tends to make solid state hardware LESS reliable, as long as you're not overheating.



 As much as  I like the reliability on my A2 units, the software was never particularly impressive.

 The flypool version of the nicehash miner is stand-alone on Windows - though if you're looking for AMD support it's not an option.

 It seems to be more stable in mixed GPU/CPU mining if you don't try to use all the cores on the CPU, but that might be 'cause most of my mining rigs are "dedicated to one type of mining" and therefore have other stuff running on them.

None of my 1070s are running more than about 65% of rated TDP when doing normal mining - if the VRMs are overheating at THAT low of power consumption, I'm sure I'd have noticed when stress-testing them earlier.

AND efficiency. Don't forget the MUCH lower power usage.

 No, the OP was proposing to design an entire board sized to fit in existing S5 miners (using the frame and fan, possibly the heatsinks, and perhaps the microcontroller).
 You can NOT just put Bitfury chips into anything else, they're not compatable that way at all.

 I'm sure if you wave $1 million under Bitfury's nose, they'll be glad to sell you chips.