I won’t say it’s impossible, but I would be really genuinely surprised. 32MB / hash of total bandwidth (read + write) is needed, and 2MB or so of stashes per hashcore.

You have 1280 URAM blocks of 288kb by 72 bit interface dual ported in the biggest configuration .That’s an incredible amount of internal bandwidth but you can only store 23 or so simultaneous Cryptonight7 2MB blocks in that. The absolute biggest part (which isn’t on the 1525 board) has 360Mbit URAM, 96Mbit BRAM, and 48Mbit Distributed RAM, holding a theoretical 63 MB of pipelines, assuming you didn’t need a single bit of that for the rest of your logic (you do).

The external memory at say 4x64 DIMMs @2666 is only 85GB/s, or 2.6 KH worth of bandwidth with a perfect access pattern.

Even if you could imaginarily use all 2000+ balls on the FPGA for 2666 MT/s DDR style  speeds you’d still only clear 20KH against external memory and that isn’t even real bandwidth.

 Even if you took the biggest part with 128x32 Gbps transceivers to SERDES memory you’d only have 16kH limit from bandwidth.

Unless you break the algorithm itself, there’s no where to find the bandwidth + storage space for 64khs on a single FPGA.

at least an order of magnitude.

When I get these on 240 VAC the power supplies can supply 1200 watts each so I will switch to "High Efficiency Mode" which will pull power from only one supply and have the other in redundant very low power mode.

I was about to pull the trigger before all the coins fell off a cliff in value, hah...

I am pricing out a full set of UPS units, though, through some distributors... since I'll need 100+ of them. Gonna just have to suck it up I suppose!

Apparently this IS, in fact, a thing :

https://community.spiceworks.com/topic/1447342-dell-r820-slow-after-power-outages

So... that being said I'll have to work on them tomorrow when I have some time.

I did discover that if I shut the server all the way down, pull the plugs, wait 10-15 seconds, then plug it back in. After rebooting, it seems to be acting normal again. So what is it about pulling the power OR changing the performance profile that would fix whatever issue is happening?

Well... I have to take that back a bit now.

My power flickered last night thus turning off all my hardware. Came in, turned all the Dells back on... started XMR-STAK on them all & walked off.

Noticed my pool hash-rate never returned to how it was... every one of my R815s is now ~400 H/S slower than they were before the power flickered (1600 H/S vs. 2000 H/S).

I've tried re-booting and even re-compiling XMR-STAK... seems to be permanent ?!

Any ideas ?

These are our assumptions and they are  our scripture. We will stand by our assumptions until we get smacked in the face by asic sales.

I see day by day, ASICS of many coins are launching, individually or multiple. Chinese companies are trying to kill the GPU mining market once again cause they are greedy, just like they did to Bitcoin and Litecoin. GPU Mining died in 2013 and was resurrected in 2015 with the Cryptocoins Generation 2.0, although many of you miners just started mining on 2016 due to huge pumps on those 2.0 cryptocoins.

As Cryptocoins got more popular than in 2011 - 2013. The new generation of Cryptocoins 2015 and afterwards are endangered by ASICS once again, and 2018 looks to be the year of per the title "GPU mining will die in 2018!".

Keep in mind that there are possible ASICS for ethash already for Q1, and even if ETH goes POS, ETC and other ethash coins will be only ASICS and if that happens only Equihash coins will be minable for a time and I have a word that there is already an ASIC in development for Equihash already and will start delivering in Q4 2018. As GPU mining keep gets scarce, other popular coins will have pumps and so ASICS will be born for them.

I want to know your thoughts about it and I wonder the limits of --> "where, there is money to be made, there is greedy"

Keep in mind that this price crashing we are seeing is due also to the fear of ASICS for most GPU mining coins.

News from Baikal, i recive this tweet.....
Is possibile?
They are in Panic!!!
https://twitter.com/baikalminer/status/983278127504769026?s=09

If you look at my stats, post count is a LOT higher than activity.

I was misremembering details, but it's not "post count" that determines activity as such, it's "consistent posting" that does.

https://bitcointalk.org/index.php?topic=2658590.0

Activity isn't a measure of post COUNT - it is a measure of how many weeks in a row you've made at least one post.
I find the claim of $20+ a day per $1 in electric usage hard to believe, even if you're in one of the 2 counties "next door" to me where electric is available for 3c/KWH or a hair less even for for small miners.
$10 to $1 ratio, sure with a well optimized rig, but I don't see any way RIGHT NOW to get to $20 to $1.

A month ago, $20 to $1 would have been believable.

How do I know you're not Santa Claus and don't live at the North Pole?

There is no proof, and they would have no desire for knowledge of their "testing" to be publicly available.

But when you have a situation like what happened with Cryptonight/Monero, when despite GPUs being sold out across the board and Coin price dropping yet the network continues to grow exponentially in hashrate, something else in driving that increase. Ergo, probably ASICs.

The real question isn't how do you know they are, it's why wouldn't they be mining it? Wouldn't you? Following money, motives, and incentives usually paints a clearer picture than the limited "factual" evidence you have available.

I know a lot of people speculate about that and I do believe it to be true.  I mean, why wouldn't they do it?  I think I've come full circle on my opinion of Bitmain at this point ... Really doesn't matter to them if they sell this unit at $800 or $400, because they've already accumulated enough income off of mining with them privately to sell at whatever price they can get.  Jihan is a very smart man and I'm sure they carefully calculate the max price they can sell their used hardware at.  Getting money out of the used hardware is just the icing on the cake for them, if they have been mining with them for months.  You'd be doing it that way too, if it was your business that spent millions in R&D on something like that.  

Jihan doesn't even care about crypto, from what I've heard.  All that I've been hearing as of late is that he is using this as a catalyst into some really scary AI shit.  

The problem is $20+/day at $1 electrical cost at today's current situation is a very debatable claim. If you'd like, i'm offering you the chance to PM the details of what you are mining and i can guarantee to vouch for you if it checks out. I will stake my account and swear that I will never tell anyone else about it.

Interested?

How can you be better than the average joe when it's all about hardwares and stuffs? Do you mean that one should have resources which they can able to buy those things that they need? I'm at 50/50 point when it comes to mining, especially that there is a scarcity in GPU's. Wish that I can turn back the clock and never invest into ICO's or trading new coins. I was able to purchase 3x 1080 Ti's when it was cheaper @ $1k/unit and put half of my investment to those things that I mentioned. If you have faith in what you do and have a better decision capacity than I am, mining will still be profitable.

Bitmain looks like an American company where only money really matters, hard to say it's Chinese. I wonder where is the socialism and equality. That jihan guy maybe spent most of his life with evil american manufactures. Learnt from them.

[/i]

For once I am in a closed garage, so using fans will no exhaust the heat, at all.

Using fans in a closed spaced is mimicking an oven, that is why i need to vent to the outside.