I haven't been watching the other ethash coins all that closely, but Ethereum ITSELF has shed a noticeable (but not large) amount of hashrate for weeks now - with a rather large dip on April 9.

 *I* am "in the west, in the Americas" but there's no salt water within 100 miles of me and no issue with salt in the air anywhere in the area of very low cost electric I'm in.

$462?

I smell a "new node" Gigahash Bitmain miner announcement in June for Scrypt.

Guess they got tired of being beat on efficiency by the A4 and matched/beat by the L21 - or more likely slowing sales of the S9 have left them some "available space" to order current-node chips for something else they've been working on for a while.

He seems to have confused "bids" for "final sale price".

There is also the question of how many of the listings are from miners, and how many are from gamers that can "finally get the card I WANTED, now selling off the old card".

I've never used rxOC so I can't compare to it.
I don't have a link to my "work in progress" at this point, possibly soon.

Optimally the next AMDGPU release will make it obsolete, as AMD has CLAIMED they're going to "add undervolt support" to it.

Or it will go unprofitable for a while and shed a bunch of hashrate - a far more likely option IMO for pretty much all GPU-mineable coins when The Next Big Shakeout happens.
Much like what happened to X11 coins when folks GPU mining Litecoin dumped out of it due to the Gridseed GC3355 + coin price dump drove GPU miners out of Litecoin in early 2014.

Or perhaps a geothermal loop to cool the water before it goes into that "coil".

They're not a one trick pony - just somewhat optimized for specific usage.
The P104 and P106 aren't all that different from their "consumer" varients, and I suspect the "reduced core count" on the P102 might be a BIOS controlled option rather than a separate GPU model from the 1080 ti.

Overpriced - that's ALL GPUs to a degree at this time.



Has anyone tried a P102 on Folding@Home? It would be interesting to see if the optimized RAM was enough to make up for the lower core count.

I saw the comment about 1 m^3/sec but thought it was a typo as it's not a normal specification - airflows are normally specified in m^3/min or CFM depending on if you're in a metric area or not.

How bloody big of a farm are you running to need airflow in the millions of CFM?

I'm also curious about where, as I don't know of any area that is close enough to an ocean or a large salt lake to have significant salt water in the air that also has access to low cost electric.

The price won't matter, once ETH goes fully Proof of Stake income from mining it disappears as there will be NO MINING any more on it.

That's also going to dump profitability on ALL GPU mineable coins across the board though as the (currently) 9 million or more GPUs mining on ETH look for new homes, and will hammer profitability even harder on any remaining ethash coins as the E3 units in use by then look for new homes.

You are not dreaming - they kicked Batch 2 up to $1800, apparently Batch 1 sold out fast enough they think they can find enough "suckers" to pay the higher price.

I wish Innosilicon would drop the MOQ 5 requirement, or find a distributor that actually DOES something - those A4s for under $1200 are looking yummy.....

Haven't ever seen it on ethermine, but I've seen it occasionally on nicehash when first starting up - seems to be a temporary "pool overloading on making new connections" issue, as it normally resolves itself in a few minutes.

Dropping the wait time down from the default 20 secs to 10 secs seemed to help a little, but that might have been a coincidence as I only tried that once.


I'm think Claymore dropped support for pre-GCN AMD cards somewhere along the line - he definitely has NOT added and has stated he has no plans TO add -asm cores for older cards so you might just need to reconfigure to "don't use ASM" on those.

Bitcoin.

On a GPU.

*ROFLMAOSC for about 5 minutes*



You have GOT to be joking, right?

Once ETH goes Proof of Stake fully, ETC will be INFINITELY more profitable to mine than ETH will be.

We just don't know "when" that will happen - even the devs can't give a firm date, just an estimated target (that they have missed a FEW times for various reasons).

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

30 amp PDUs are common because many or most data centers traditionally used 30 amp "drop cords" to feed their racks from.
That has changed some in recent years, as "density" has become more important to many data centers but the change has often been to set up *2* or *3* 30-amp drop cords to feed a single rack (like the OpenCompute version that put 45 separate dual-CPU servers into a single custom rack).

Most servers tend to be designed to run on 208 VAC or 234 VAC though, as it's a bit more efficient to run and 208VAC in particular is easy to pull out of a common 440VAC three-phase power feed into a fairly large data center.


I am not aware of ANY NEMA 60-amp outlet of any sort, and I'm not sure if anything larger than 50 amp has ever been available - higher current machinery tends to be large enough to make hard-wiring it practical (and required under the Code at some point).

Even the old Mobile Home 50-amp hookups (NEMA 14-50 in all the cases I've ever seen) are no longer allowed for new installations, though they are "grandfathered" in and acceptable on existing installations, and they seem to still be acceptable for RV usage in campgrounds.

What the heck is that supposed to mean?

Even my Magic 8 ball comes back with "answer hazy, ask again later" about THAT comment.

Why ask here?
The only available information is what Bitmain has posted on their sales site.

I'm still waiting for evidence that any ASIC was mining in quantity on Ethereum before Bitmain announced the E3 (and any evidence that one exists NOW in more than engineering sample quantities).

There were no "sudden huge surges" like Litecoin had in second week of December or even any sudden small surges, the growth was pretty steady and AMD cards in particular were certainly in short enough supply during that growth to explain ALL of the hashrate increase.

If someone was getting enough batches of ETH ASIC chips from a foundry and building them into miners to account for a significant amount of the current ETH hashrate, there should have been a SURGE somewhere, not a fairly steady growth curve.


Note that Bitmain's business model has always been more about "turn the miners, get the profits, so we can buy more chips to make the next batch and turn THAT", which makes them a lot more money SOONER than mining does.

Very minor SLOW factor - what's going to drive GPU mining out of ETH will be the move to Proof of Stake, the E3 is too low performance and Bitmain CAN'T make enough of them to be a major factor anytime soon.