The difficulty rise seems to be more attributable to folks adding new rigs - but it's hard to tell for sure without a lot more information than is available.

 More like 2 to 1 ratio when prices ON BOTH get down to normal.
 Still a big difference though.

I would be more worried about the reliability and even availability of the infrastructure for electric and Internet in most of Africa - most of that continent does NOT have reliable power or power at all, and Internet access is VERY spotty as well in most of the continent.

 Worker wages are low but worker technical KNOWLAGE tends to be very very low to nonexistant in those areas - and the areas that have workers with some technical knowlage tend to also have higher wages/cost of living.


 A4 info - check in the altcoin mining area, there are at least 2 threads there one with a fairly current update.
 A3 - appears to have been aborted, given how the A4 specs changed I'm starting to think the A3 was going to end up in the same ballpark as the LK-1401 on specs, competative with the best 28nm designs but NOT really any better and not worth the extra costs involved in producing it on a 14/16nm node.

I had a similar issue with the A2 units I bought from Zoomhash a while back.

 I pulled the SD card, put it on a reader attached to a LINUX system, mounted the card, then poked around in the networking config stuff to find the IP address - which was on a 10.x.x.x subnet as I recall, NOT the 192.168.x.x that Zoomhash CLAIMED the miners were on.

 Default netmask on a 192.168.x x subnet is /24 - you can use /16 under LINUX but that sometimes breaks some "picky" software.

Any R7 card is very LOW END for mining on - I suspect you're not going to do a lot better.
One thing that MIGHT help some is to not run the most recent "Crimson" drivers, they're widely reported to be slow for mining under Win10 - back out to 15.12 or 15.7

 I probably should have said "thermal throttles" not "thermal limits".
 My R9 290 Sapphire Reference cards start throttling at 94, not sure what their shutdown temp is as I don't think they ever reached that.

 When did BitFury move to China?

 Also, GlobalFounderies DOES have at least one of their fabs in the US (ex-IBM as I recall? Or was it ex-Intel?).

 AMDOverdriveCTRL was never updated for the R9 290 or R9 280x, in my reading about it - and it flat out didn't do voltage adjustment on quite a few of the cards it was SUPPOSED to work on (locked BIOS issues was apparently the concensus on those cards).
 I tried to download atitweak from somewhere, but it came up as virus infected, and I never found anywhere else I COULD download it from that seemed like a legit site.

 Academic point on the voltages at this time, since I DID find working copies of TheStilt BIOS series and don't need to worry about LINUX-specific voltage mod tools any more.

 Some cards run hotter, even with massive airflow through the case.

 R9 290, R9 280x, and especially the R9 290x were notorious for running hot - and lasting a long time anyway.

 My triple-box with 3 R9 290s (Stilt bios, dropped temps a LOT then I overclocked and brought them back up some) has 2 Delta 120mm 250CFM fans in front, the 4 stock 200 (220? 180? something crazy-big) fans on the side, and a low-flow 140 in the back I couldn't find a good place for. My 290s run anywhere from 75-85C even with THAT airflow at 90% fan setting and space between them - but they're also pushing 1060-1100 Mhz core clocks and that's in a room with ZERO a/c and decent airflow, but still pushing 95F inside-the-room ambient temps the last couple days (heat wave seeing highs around 90).

 My older HD 7xxx series cards I can generally keep under 70C, the various NVidias I set target and custom fan profiles on to keep them under 75C - the 7750s are an extreme, they all (except the Sapphire with it's GOOD original fan) have Delta 60mm screamers twist-tied to them and never see 60C.


 A R9 280x at 92 when mining is nothing unusual - it's just the card is a hot running card - and it probably thermal limits around 95ish.

The only thing that hits a CPU hard in Ethereum mining is DAG file creation - my machines running Semperon 3000+ need a few minutes for that, but it interleaves with mining so it's not a noticeable slowdown issue even with that OLD low-performnace (by current standards) CPU.


 BTW - your GPU from what I've looked at about it is about a tossup to a little WEAKER than the on-chip GPU of the current A10-78xx series - which is considered to be ENTRY LEVEL at best and does not handle many higher-end games well if at all. Works fine on stuff that's 3+ years old, or stuff that isn't heavy graphics-intensive though.

 I wouldn't try to run current high-end FPS games or Ashes of the Singularity on it....

The "software" voltage configuration on the S5 never actually worked - it's a string design with NO ability to configure the voltage.

You have some other issue if the miner isn't working, NOT an issue with the voltage setting.

 Size, sure - but power consumption on BOTH of those units was quite high for a "home" miner.
 I'd aim for under 1KW and preferably WELL under for a "consumer" miner - 500-700 watt range is a lot more manageable for non-BIG FARM miners, IE the S5 and SP20.

 WiFi - definitely should be OPTIONAL.
 A lot of us don't have any use for it nor any interest in paying for it.
 Available AS AN OPTION, sure - some folks DO want it.

Forget mining Litecoin on ANY GPU - the best are very far inferior to the ORIGINAL Scypt-capable GC3355 ASIC, and that early ASIC is already unprofitable unless you have free electric.

 Your GPU is VERY VERY weak for Ethereum, even if it's a 2GB model - for perspective, the GTX 750ti 2GB only manages 3-4 MH/s ballpark on Ethereum despite being quite a bit more powerfull than the GT730.

 Even for gaming that's a pretty weak card - I'd go to at LEAST a GTX 950 (pretty inexpensive and viable for ETH mining at 10-11 MH/s, though more on the basis of decent hash/W ratio not so great on hash/$).

 Wish I could, but your insistance on using broken github makes it impossible for me.

 Still waiting on a reply to my earlier question of "where can I get it that does NOT use github".

It's not a good card for Ethereum, my GTX 960s see about 11-12 MH/sec at best (barely more than my GTX 950s get though the power usage seems to also be just barely more).

I've not seen ANY LINUX tool for AMD cards that allows voltage control at all.

UC is also problematic - ATICONFIG will let you set an underclock, but the card then IGNORES it in all of my experience with it on dozens of different cards.
Ditto CCC for LINUX.

OC on the other hand works just fine.

To achieve 0.1 LTC/day, the card would have to mine about 10 MH/s - which is *AN ORDER OF MAGNITUDE* greater than any GPU has ever managed.

 NOT POSSIBLE.

 I seriously doubt the card will be able to manage 2 MH/s, much less 10.
1.something it MIGHT manage - but that's entirely due to it clocking quite a bit higher than any other card ever used for Litecoin, and DESPITE it having fewer stream units than some cards that HAVE been used for Litecoin in the past.

 Keep in mind that Scrypt wasn't a "memory intensive" protocal to anywhere near the same degree that Ethereum is - it was more limited by the processing units x clock rate.

 I stand by my original reply, it won't be as efficient as ANY Scrypt-related ASIC has managed - closer than anything that came before, but STILL inferior to the original-generation GC3355 much less anything more current - and those GC3355 units are NOT PROFITABLE any more unless you have free electric.

 Dang, you must have EXPEN$IVE power or a very inefficient mining setup.

I doubt that the miners would have sold at $4000 on the original spec - probably more like $6-7000.

Even with the reduced "final" spec, the price is a LOT more reasonable than I was anticipating - especially if the miner is configurable to be able to use the "more efficient" settings the chip has available, and *IF* it can use the less-efficient but higher-hashrate setting the chip ALSO has available.


 Based on the final spec, they appear to be running the chip at the second-highest hashrate setting in their table, or close to it - which is also the second LOWEST efficiency.

 At least one site I found that actually had the BIOS files has the numbers mixed up for the 290 bios and the 290x BIOS versions - watch when you flash that you're flashing the RIGHT BIOS type, flashing a 290x BIOS on a 290 card CAN cause the issues you have seen.


 The main difference between the versions, other than the 290/290x split, is the default voltage setpoint - the higher number versions undervolt more, and IN THEORY should reduce power consumption vs higher number versions - with the tradeoff being the potential that the card won't be able to clock as high.