Lightning does wierd stuff sometimes.
Or your modem had a GOOD ground that protected past it.


Fibre optics are an insulator, lightning CAN'T travel along them.
Cross-ocean stuff does a "high voltage power line" through the MIDDLE of their fiber lines to be able to power the in-line repeaters, using the fibre optics parts as part of the INSULATION for that HV power line (in effect if not specifically by design).



Anyone else noticed that Phil and I are rated for more "coins" than the probable number of altcoins that exist, with -ck pretty close to that as well?

8-O

But in what quantity?
A few engineering test units? Not at all hard to picture.

Large quantities of production units? Bitmain prefers the rapid high profits of SELLING miners to mining with them as a general rule as they make MORE in a year that way by a wide margin.

I don't have an A5 to compare to, didn't realise Innosilicon had changed their UI to be so much closer to the Antminer S5 UI - which I STILL maintain is more likely to be the ORIGINAL model for that UI (or perhaps something even older out of Bitmain, since I didn't own anything before the S5 from them).

I'm not questioning the possibility of Innosilicon being involved (nothing new in THAT corporate history as I've pointed out), though I'd think they'd be more likely to be involved at the CHIP level than the MINER DESIGN level.


I'm also wondering if the chips involved are a next-generation upgrade from the announced but "never made it into a miner" Innosilicon A3 chips....

Polaris undervolt under AMDGPU is possible NOW (and I think it also can be done for Vega but haven't actually looked at the code for that) - but you have to hack the applicable kernal module and it's an ugly mess - and trying to do it with individual undervolt for each card is going to be a major PITA on the current AMDGPU version.
AMDGPU however does NOT support undervolt OR clock control for R9 series cards (try it, the clocks DO NOT MOVE no matter what settings you use), and STILL offers exactly ZERO support at all for the iGPU on AMD A-series CPUs (even the GCN ones)
OpenCL for SOME APU models under ROCm is supported (works on the 7890k and 7860k, I've not tested on the 5700 as that's pre-GCN, and I don't have any other models to test it on) but forget trying to use APU graphics if you have anything else in the system using AMDGPU.

*SUPPOSEDLY* they are adding undervolt control to "the next version" of AMDGPU - but ghods only know when that will show up, and if it will work with anything OTHER THAN Polaris or Vega.


In my testing, the undervolt DID reduce power consumption some on Polaris cards that are not already BIOS modded, but I don't recommend trying to do so on modded cards due to how the driver code is set up.
The Phoronix "818" figure proved to be rather optimistic across a range of RX 470/480 cards for stability though, I think I ended up at 850 or 860 before I finally got the rig to be stable for more than a couple hours at a time.


Real driver support on a per-card basis would be a LOT better.

Nvidia has no interest in rushing out their next generation cards, as the CURRENT ones already dominate the market and are selling faster than they can make the GPUs.

They also aren't widely used in Mining compared to AMD, so they have a lot less incentive to bother wasting time on "mining specific GPU models" - so far ALL of the "mining specific GPU cards" have originated in 3'd party CARD manufacturers using standard GPUs.


Lead time on a new ASIC miner design is months, not years - but it's still quite a few months.
I suspect Bitmain started work on an ETH ASIC miner probably around mid-year 2017 timeframe, possibly a bit later, when the altcoin "price jump" proved to have some staying power.

What I DON'T understand is their SIA and CN miners - the SIA one swamped the VERY TINY market in no time flat, and the CN miners are going to be pretty much "dead on arrival" given Monero's long-stated OPPOSITION TO ASIC.
I'd have figured they'd target equihash LONG before SIA and that they'd have been smart enough to avoid CN entirely.


Given that AMD was sold out for MONTHS despite very high pricing during the ETH hashrate growth this fall/winter/early spring timeframe, it's VERY EASY to account for that growth without needing some hypothetical and PROBABLY DID NOT EXIST IN QUANTITY ASIC miner.
Also don't forget the Nvidia 1060 and possibly some 1070 cards mining ETH, *AND* some folks switching from other algorithms due to ETH getting more profitable on their EXISTING cards.

Based on a video claiming to show the new Bitmain ethereum ASIC, it's not all that impressive.

1500 Mhash (about 50 1070 or BIOS modded RX 470/480/570/580) but no power number quoted.
At 1000 watts (LOW for a Bitmain miner) it would only be 5x or so as efficient as current GPU mining.
At 1500 watts, it would be LESS THAN 3x more efficient.

Factor in the probably efficiency improvements of the next-gen Nvidia cards (due later this year SOMETIME) and the numbers get even less impressive.

Given the HUGE amount of ETH hashrate alone, even at Bitmain's typical "3000 or so" miners per batch it would take quite a few months for their ASIC to capture *HALF* of the ETH total network hashrate.
Then add in the existing total network hashrate from the other ethash coins like ETC.....

Yes, it's a threat - but it's not going to kill ETH mining soon, and unless they price it CHEAP it's likely to never make enough to pay for itself.
Unlike SIA and X11, it's NOT going to kick the hashrate multiple times higher in less than a month - it's going to be closer to the move in Scrypt where the L3+ and Innosilicon miners are ONLY NOW finally getting deployed in enough numbers to replace the "previous generation" like the A2 and the Titan and make them generally unprofitable.

IMO if Vitalik and crew keep their current target timeframe for POS deployment and actually MEET it for a change, it'll be a waste of time doing a "algo change" fork.


I don't class this miner as a "big mistake" like the Cryptonight ones from Baikal and Bitmain, but it's not looking a TON better.


https://etherscan.io/chart/hashrate shows an all-time high spike on March 26, but seems to have been a "day of the week" or "luck" thing as the trend has been a gradual DROP for about 2 weeks now.

I suspect they have some engineering and "early production" test units up and running - but a few hundred at most.

I can picture that, the rig I just re-assembled with a mix of HD 7750 and R9 280x and a single 7850 is profitable but not by a lot - even at MY low electric rate.
I suspect it's going to switch over to Dnet RC5-72 work fairly soon if coin prices keep sliding.

Up side - the 7750 and 7850 cards were paid for via Litecoin mining YEARS ago, while the R9 280x were paid for sometime last year early on.

That looks almost IDENTICAL to the Bitmain UI on the S5 miners I used to have and almost NOTHING like the interface on my Innosilicon A2 miners.

Not saying the HARDWARE does or does not have a lot of Innosilicon input or influence, but that UI don't look like Innosilicon at all.



As far as the Dragonmint unit managing over 16 TH in their video, my first guess would be "cherry-picked unit".

Most gold/platinum/titanium rated power supplies don't lose much efficiency from 50% to full load - 3% ballpark is pretty common, sometimes LESS sometimes a bit more.

https://www.techpowerup.com/reviews/EVGA/SuperNOVA_G2_1300/5.html (table at the bottom of the page) for an example - drops from almost 93% at 40% load to a hair over 90 at full rated load or a hair under 90 at 110% of rated load.

373.06 is probably the best overall driver for anything before the 1070 ti - not sure on Windows for that GPU or the 1050/1050ti but LINUX mining likes 384.98 very well indeed and that version works stable with older cards as well.

Intel and AMD have been reported - MONTHS ago - to have a partnership on some laptop chip line, though it might have been more than Intel licensed AMD "Vega" iGPU technology instead, the reports weren't 100% clear.

NOT an April Fools joke.

https://arstechnica.com/gadgets/2018/01/kaby-lake-g-unveiled-intel-cpu-amd-gpu-nvidia-beating-performance/
https://www.anandtech.com/show/12003/intel-to-create-new-8th-generation-cpus-with-amd-radeon-graphics-with-hbm2-using-emib

Among MANY other reports.


However, that Project Canis thing MIGHT be April Fools.

I'm afraid you'll have to drop out of the Internet for a few years for failing your "Understand Typo" skill check.

9-)

I'm afraid the coins part is badly broken.
I'm pretty sure I'm being shown as having invested in more cryptocoins than have ever existed.

9-)

https://bitcointalk.org/index.php?topic=3232693.0

I had to do a bit of digging to find it.

Windows 7 has issues with anything more than 4 GPUs - sometimes 5 if you're lucky, and there was a site that showed how to "patch" the drivers to work with more but that site has been down for a while.

I believe the same limit applied to 8/8.1 but I gave up on 8 a couple years ago when the UNUSABLE INTERFACE wasted 6 hours of my time trying to do one SIMPLE thing (FIND the pre-installed mining software and configure it and RUN it) and wouldn't let me do so no matter what I tried, causing me to blow 8 off the system entirely and replace it with XUbuntu (took me perhaps ONE HOUR to do a "from scratch" install, do all the needed updates, install drivers, install mining software, and have the system MINING - and I could have done it faster had I been paying 100% attention to the system).

I recommend a move to LINUX, which is currently running 7 various mixed Nvidia cards on my ASUS B250 Mining board under the 384.98 drivers with no issues (other than a KNOWN "doesn't report the power limit data correctly" bug with 1050 and/or 1050 ti cards but you can still SET the power limit and it works).

Tell AMD to add support for clock control and undervolting in the AMDGPU drivers.
At this point, fglrx is INFINITELY better for R9 and older series cards than their AMDGPU stuff.

Also, tell AMD to SUPPORT THEIR BLOODY APU iGPUs in something newer than fglrx - 2+ years and STILL NO BLEEDING SUPPORT UNDER LINUX FOR APUS on the AMDGPU stack.

I'm sure they have "engineering sample" and test units working by now, possibly hundreds of them given the widespread "release sometime in the next 2 months" reports from more than one source.

1500 Mhash/sec is equal to a bit less than 50 GTX 1070 or a bit more than 50 "modded BIOS" RX 470/480/570/580 cards - if this thing eats 1400-1500 watt ballpark at the wall like the T9 and S9 often do, that makes the efficiency less than 3 TIMES a well-optimized 1070 or RX 470/480/570/580 rig (and would be pretty close to a prediction I made somewhere).
At 1000 watts, it would be about 5 times as efficient - but I'd have to see it run at THAT low of a power level to believe it.
At anywhere between those power levels, it's going to run quite a bit hotter than a good "efficient" riser rig, NOT a win from a cooling aspect though definitely one from a footprint aspect.

A GPU killer, but a SLOW one even at Bitmain normal batch sizes - and they better price them pretty CHEAP for miners to have a reasonable chance to pay them off before ETH goes Proof of Stake, $10k would be a deal KILLER for them.

This is all assuming that production units can actually reach that 1500 Mhash mentioned in the video or close to it.


(edit seems I was off some IF the figures in the video stand up, I was predicting 800 Mhash more or less at 1300-1400 watts at the wall, earlier post in this thread)

That is NOT an ASIC rig, that is a GPU rig using "laptop" type GPUs.

It's not rxOC (I haven't figured out how to get it working on a USB flash drive yet) but I've got a working XUbuntu 16.04 setup running with AMDGPU drivers.
It does some undervolt and substantial underclock with Polaris cards, but the R9 290 cards I've tried on it have no clock control at all (limitation of the bloody AMDGPU driver stack) and I'm not sure of the BIOS mods on those cards (TheStilt) is letting them undervolt.
I've not tested it on Vega as my only Vega card is in my gaming machine.

It does NOT allow use of an AMD iGPU as video output (MORE AMDGPU limitations, over 2 YEARS since AMD has supported any APU at all under LINUX) but the ROCm stack for OpenCL support DOES work on the iGPU of an A10 (for what that's worth).
It DOES allow fan control on everything I've thrown at it.
It's also not a fancy "one place setup" sort of thing, more of a "minimal install standard XUbuntu" type setup.


The main issue is that the AMDGPU driver stack is still a very poorly done and VERY incomplete "work in progress", to the point I actually recommend using Windows 7 instead of LINUX on Polaris cards BECAUSE of the junk drivers for LINUX.

I just dropped 5500 watts or so a couple days ago - Scrypt mining profitability got low enough I turned off the A2 farm.

Still going to be ballpark 10 KW or so on mining + evaps + misc usage though.