Topic: [ANN] CureCoin 2.0 is live - Mandatory Update is available now - DEC 2018 - page 49. (Read 696267 times)

Has anyone tried to mix NVidia and AMD GPUs on the same Linux Folding Rig?

It is very likely that our team will be ranked #2 right on Christmas Eve!  
Hope Santa will bring some good Champagne!

OK, here's the problem with your chart of randomness: It provides no definable intel to compare even like items and it raises many questions.

Not the least of these questions is: Where did mbmumford get that magical GTX1070?
11/21/2016   P9676   mbmumford   GTX 1070   1531   7544   1920   0:00:07   729,806   R1, C71, G183   Windows 10 x64   [0x18]   5,879.00   8.1
And how did it so vastly outperform his overclock?
10/31/2016   P13201   mbmumford   GTX 1070   1886   7604   1920   0:01:38   686,432   R17, C4, G0   Windows 10 x64   [0x21]   7,242.20   10.6

Is it the same unit with 2 separate runs?

On a completely different subject - the CureCoin forum is no longer accessable with Chrome, *apparently* due to it's use of an outdated encryption protocal (probably RC4 that was CRACKED a while back).

 Is the 1080 capable of being overclocked as much as the 1070?

 I do concede some of those overclocks (on ALL of the cards) are fairly extreme - but that doesn't negate my point about "real data from real folding".

...

OK, I might concede some points after that if the overclockings were similar/comparable.
Come on, +500 on the 1070 and +300 on the 1080, really?
 


Edit: I was wrong about 1 thing, though, the chart was Maxwell on the Titan

https://docs.google.com/spreadsheets/d/1vcVoSVtamcoGj5sFfvKF_XlvuviWWveJIg_iZ8U2bf0/pub?output=html#

Much more accurate info using REAL folding of REAL units than the benchmark, which only tests specific aspects of folding one at a time.

Depending on WHICH benchmark aspect you look at, the Nano should be directly competative to the GTX 1070 on performance - but in real folding work it's quite a bit worse.


 Also, the COST of those 1080s is what I was talking about - energy efficiency is going to be pretty much a tossup either way, but paying more than 50% more for the cards for perhaps a 20% PPD improvement is NOT cost effective.


 Try finding a 1080 in an ITX form factor short card like the Gigabyte (I build 3 card rigs, but due to limitations of available motherboards I need a short card to allow for decent cooling of all 3 cards).
 I supposed I COULD use 2x 1080 and a single 1070 though, so that point is a strawman argument to a degree.


 The Titan X Pascal is also less cost-effective, though it earns enough more in bonus PPD it might be better on a PPD/watt basis.
 It is NOT cost effective on a PPD/$ basis even vs the 1080, much less the 1070.


 980 ti aren't worth buying any more, unless you can get a used one fairly cheap.
 Dunno why you included THAT in your comparison - that IS a strawman argument.



 I don't do riser rigs - have had too many issues with the bloody things back in the days I DID use them.

 2 x 3 card no riser rigs aren't much more expensive IF at all than a single 6 x riser rig - you can use smaller much less expensive PS (cost per watt on a PS is pretty much even for quality Gold rated supplies 'till you hit the 1000 watt ballpark, at which point it starts climbing pretty fast), you DO take a hit on the HD/SSD/Pen Drive + RAM + CPU doubling up, but you might make that up on the LOW COST very common motherboards for 3 slots vs the rather HIGH price very limited selection of motherboards that can handle 6, and you DEFINITELY make up some on not needing risers.

 It's also a lot easier to get a 3 card rig to work in the first place, though once you get a rig to work at all it's not hard to "clone" it if you're running Ubuntu/Xubuntu - would be trivial if Ubuntu/XUbuntu used LILO instead of that PITA Grub bootloader with it's usage of that stupid majorly irritating UID garbage instead of standard drive designations.

 Non-riser machines also tend to be more reliable, and if you have a riser rig go down you lose ALL of the production from it vs. half if you have 2 x non-riser machines and one goes down.

It comes from the core of my being. 

Thanks for the info. Yet why do you need to be so arrogant in your posts???

Rate: 6 1080s = 6 1070s PLUS 1 980Ti (or even 1 TitanX)
Energy: 7 cards vs 6 to save ~30 watts

Past a hobby rig, whatever "savings" you have are lost in more risers and PSUs.
I'll keep the good stuff, thanks.

 GTX 1070s - they're more cost-efficient.

 Or Titan X Pascals - the ultimate folders regular people can get.

Indeed, you can at least read some. Now the part that you missed learning is that POS is "SHA256 mining"; which means that POW mining receives only a part of that 20%.
When POW hits as "high" as 125TH, then ASIC miners (POW) get about 21-22% of 20%.

Indeed! Life is unfair sometimes. But if you sell your shares whatever you do with the money, you lose your decision power.


According to the very first post by Vorksholk in this thread:

Curecoins go to three main areas: Folders, Miners, and Developers. The folders get 76% of the total coins (80% of the coins distributed per day). SHA256 miners get 19% of the total coins (20% of the coins distributed per day). 2% of the total funds are distributed to people who donated to project development. The other 3% is dedicated to Curecoin developers, and will be used for paying for development costs (such as hiring professional programmers, paying for infrastructure, etc.), and for giving back to the community (folding hardware giveaways, faucets, covering 0% mining pools, etc.).

So, in that theory, the guy that sold his coins to buy a new GTX1080s for dedicated folding gets little/no votes?

Also, if you think ASIC miners get 20%, you have no idea how it all works.  

In my opinion there should be no difference in treatment between miners and folders as Curecoin 1.0 wouldn't exist without SHA256 miners. If they are only 5, good for them: they share the 20% of daily coins. But did they hold or dump their coins?

In the same way there should be no difference between folders and those who just invested their $$$ in buying CURE and thus raising its value.

So probably a vote system based on "shares" i.e. how much CURE one owns would be the fairest from a business point of view. This is the system that FLDC had (VOTE tokens) but if I remember well, they used it to ask more risky questions than which new logo

As well as "how much weight do the only 4 guys that regularly mine CURE hold?" (99% of all ASIC mined blocks are 5 guys) 

I was expecting some remark like that and that's why I added the step 5... so if a rich guy is willing to pay a lot he would at least help a charity

You can also say 1CURE/vote but it would be quite a lot of work to check that a rich guy didn't create a lot of addresses.

And there might of course be the debate about whether a guy who owns only 5 CURE should have as much weight in the decision as Ed Olkkola... In business it's the number of shares that gives the power of decision...


 
 

So, in your theory, basically 1 guy with deep pockets can decide all on his own.  

Maybe the release of CC2.0 would also be The occasion to put a more "professional" or appealing logo on CC2.0?

This wouldn't require too much work for Curecoin Team, and now you have a webmaster (Mr. Clavijo). For example:

1) ask the community designers to propose their designs before January 31,
2) discard all non-serious proposals
3) display on a page on curecoin.info the candidate logos with a different Curecoin payment address under each one
4) give one week to people to vote by doing a Curecoin payment to the address that corresponds to the logo they prefer
5) the winner logo is the one that cumulates the highest amount
6) donate all the received payments to a charity

 

Thanks for the timeline!  

Yet, Curecoin Team, please do the necessary in terms of communication, advertisement and marketing so that CC2.0 release and its new security technology and all their benefits don't get unnoticed by the crypto community and investors!