As an example - Antminer's Scrypt router (R4?) - which has never earned even close enough to pay for it's cost premium over other devices with no mining in them that have similar performance, much less it's total cost, and at THIS point I think it might be underwater on cost of running the mining chip vs income unless you have FREE electric.

1300 watt per one rumor, as I recall - my own estimate was closer to 1kw, there is NO WAY they're going to manage 400 watts for that hashrate with current technology.

I'm not sure if they could manage it with 7nn/10nm tech when that gets widely available.

There have been ideas to put mining chips into consumer electronics before.
They keep dying when the costs end up killing the product concept.

There is NO BLOODY WAY you're going to see a mining chip in a smartphone - they eat way too much power, and smartphones use "battery life" as a selling point.
Same issue with ANY mobile device - power consumption is going to kill the battery too fast, if you have enough hashrate to actually DO anything with.


The only "TV" I own has never been used as a TV - I needed a "quick" replacement for a dead monitor and my local Walmart was out of stock on monitors at that point (and there are NO other local options for computer stuff at all).
It also has a nasty irritating tendancy to want to shut itself off if it has no signal for more than about 3 seconds (it's shut itself down occasionally during computer BOOT UP as a result, which is a major PITA).
Stupid excessively "Green" garbage gets irritating that way.

DO try waiting to ask that 'till after the announced shipment date (in JUNE).

21.com had a similar idea a while ago, but seem to have dropped off the map.

I forget who it was that had the idea of a "smart lightbulb" with a miner chip in it, but haven't seen that come to market.

I believe the whole "consumer device with a mining chip in it" concept is a non-starter, due to power usage (forget Green certification) and requirement for Internet access (lots of folks that don't have WiFi of any form or any interest in it) and the additional cost of such an item.

I see no evidence that the L5 is mining away in *large* quantities - note that the L3 and L3++ are still VERY profitable right now if you have low electric cost - but I'd bet Bitmain has at least some engineering test units working and probably have some early production units starting to replace their internal L3 usage, possibly even one full batch worth of such.
I'm still of the opinion that the current pricing on the L3 represents "fire sale old production before the new model is released" - even though that pricing really should not be as high as it is given currently profitablilty levels.

They're going to run into the "limit of production capacity available from TSMC" issue on an L5 though, since it's certainly not 28nm (where plenty of excess capacity exists) - though slowing sales of the S9/S9i should free up some capacity for L5 production.

There is also the possibility that they're going with Samsung for this round of Scrypt miner, since Samsung seems to still have some excess capacity despite Nvidia using Samsung to make a couple of their low-end GPU models.

Problem with LINUX right now is that AMD doesn't support their integrated graphics on recent drivers at all, and current LINUX drivers from different manufacturers don't get along any more.

You CAN get an AMD integrated iGPU to run OpenCL if you're already running a discrete AMD GPU for the graphics output, but that's the best I've managed in many many hours worth of trying to make this stuff play nice together.

*IN THEORY* you should be able to do the same running Nvidia for the actual graphics but I've not gotten that to work to date, and the old Nvidia option for "no-opengl-files" stopped working 2+ years ago.

Because ASIC miners are specifically optimized to do one very specific thing very very well, while graphic cards are designed to handle a much wider workload of varying types and CAN'T be as well optimized with zero "wasted" circuitry for mining with them.

It's like asking MACK to make their truck compete with Formula One race cars while still being able to handle large loads and last for millions of miles across many years to decades.

Baikal ASIC miners.

Please don't necro threads about GPU mining of coins where ASIC miners have existed for them for many months or years.

ASIC "resistant" is not "ASIC proof" is what it means.
But a lot of folks that claim their algorithm is resistant to ASIC don't really understand the capabilities of an ASIC.

equihash does NOT use all that much memory or access it a lot, unlike ethash (which is probably why the E3 for ethash is no better than GPUs on efficiency, VERY hard memory limited algorithm).

They apparently believe the unsupported claims by Marvel2 about the majority of existing ETH hashrate being ASIC, despite all the evidence against those claims and ZERO supporting evidence.

I suspect, given lead time on ASIC development and deployment, that most of the current "big wave of new algorithm ASIC miners" out of Bitmain is due more to the big price jumps LAST year, coupled with saturation of the market for the S9 leaving them with unused capacity for chips to be made at TSMC allowing them to introduce the new stuff now.
Keep in mind that development of a new chip on the current nodes takes at least 6 and more commonly 8-10 months, from "first design work" to "final chip design is verified and production quantity deliveries start", then add in another month or two to get the actual miner production design verified and production up to speed.

I'm quite certain that Bitmain would LIKE to have had more of these new models in production 6 months sooner, but probably weren't able to do so.

The new Z9 model is going to be somewhat like the L3 when it first showed up - there's enough existing hashrate that it'll take a while for them to make GPU mining unprofitable, not like the D3 where a single batch could more-or-less match all existing hashrate or the X3 where that single batch blew away the existing hashrate.
We're still looking at months to get the GPUs switched out into something else for profitability, or retired entirely.

It's not going to be like the E3 though, since the Z9 actually IS a ton more efficient than existing GPUs.

And will remain profitable on GPUs 'till at least somewhere in July and more likely August, given the current total network hashrate vs. how much this miner will introduce at normal Bitmain batch sizes.

Remember, not shipping 'till late JUNE - but probably will kill GPU mining over time, unlike the E3 for Equihash that BARELY matches what a current well-tuned GPU rig can manage.

L3+ was briefly listed at $440 or so a couple weeks back - then Litecoin price started recovering, and the L3+ price rose to follow it.

The big jump came after the ETH core development team said they weren't going to bother forking the coin to put a new algorithm in place.
The rest was small, due to the recent coin price resurgence.

And if you total the network hashrate for all of them *COMBINED*, I'm pretty sure they are well under 20 percent of Ethereum total network hashrate.

There is also the factor that by the time ETH goes full POS, there will be at least one new "even more efficient" generation of cards out from Nvidia, might be a new generation out from AMD, and the E3 is already going to be struggling to compete before the big "profitability falls off a cliff" event happens that the move to full POS will cause.

Not saying it's incorrect on what it's reporting - just pointing out that it doesn't measure WHOLE CARD consumption unlike Nvidia cards, so the numbers are not comparable and tends to make AMD cards look better on a comparison against Nvidia cards or ASIC miners.

Doesn't appear to actually be a miner, but a "runtime" memory strap modifier for GDDR 5x specifically.

Since only the 1080 and 1080 ti (and Titan Pascal) use GDDR 5x on the Nvidia side, there probably won't be a version for anything else - GDDR 5 doesn't NEED it nearly as much.

People are saying this WHERE?

There was a 10 day period in December when Litecoin total network hashrate (and the resulting difficulty increases) nearly DOUBLED.
I've still not seen a good explanation for that - there is almost NO hashrate in anything else Scrypt (ignore the "merge mined" coins like DOGE as their hashrate IS the Litecoin hashrate for the most part).
The price went crazy shortly after that, or that jump ALONE would have killed profitability on anything older than the L3/A4 generation even at super-low electric cost.