Topic: These new EFFICIENT x11 algos everyone is talking about ?? BULLSHIT or real? - page 14. (Read 16283 times)

I'm not convinced either way on this question yet and I'm open to all info.

That said, my current opinion is this... I don't think it's as simple as "less heat/power than scrypt = inefficient". I think that's a totally reasonable and logical conclusion for a person with limited knowledge (like me, and most of us) But that's the problem. We don't know enough about any of this to make an educated assessment. We're just speculating based on... well, not much.

So here's my speculation - As a person who has mined X11 for about a week now, I can tell you my GPUs are getting stressed. They are also getting hot. One overheated last night and I had to shutdown for a bit. My desktop is slow, the same as when using high intensity in scrypt. The only difference I've noticed is my power use has gone from 700+watt to around 410watt. In a country with high electricity prices, that's great for me. That said, it varies. Sometimes goes as high as 420.

At the end of the day I think SpeedDemon is right...


If scrypt did'nt generate so much heat and use so much power would we be questioning X11? No.

Thats a rumor spread by the ASIC MAFIA.

If you'd bother to have a look at what your GPU is doing with the ATI utility, you'd see that its running at 100% with X11.
The fact that it doesn't get so warm is due to the lower memory footprint and the overall lower cache utilization.

1) The algorithm is probably less ram intensive and it probably has plenty of room for optimization. That's understandable because the GPU miner was launched as a necessity in a short timetable, after DRK community gave a bounty to have a GPU miner developed quickly to counteract the possible existence of secret GPU miners.

Personally I tried doing #pragma unroll and #pragma unroll 1 (disable unrolling) tests to check the performance on my 5830 on the 15 sets of "for i =" of darkcoin.cl.

I could get around 5% more hashrate by #pragma unroll 1 on the fourth "for i" (947kh vs 939) and a #pragma unroll 1 with the 8th "for i" (979 vs 939kh). Interestingly if I #pragma unroll 1 on both the fourth and the 8th, speed goes lower than 939, so I kept the 8th.

Thing is, the unrollings don't always work the same on series 5xxx and, say, series 7 cards. I've also gained a couple percent on scrypt too: https://bitcointalksearch.org/topic/m.5863333 by playing with unrollings but it doesn't translate to other cards well, so I wouldn't say scrypt was intentionally unoptimized.

2) So far as everyone is using the same program and only a few people (<1%) use something more optimized, everyone is better off since everyone is burning less electricity. Given the extremely low profitability right now of most altcoins, electric cost is a problem.

3) Having >1 hashes is better for security in the long term. If say a hash is broken / proven unreliable, you have another 10 to go on. If however your ONLY hash is broken = game over for a coin.

It is not difficult to find out that the thread starter cryptohunter is really negative all the time. Just look around at his previous posts. I agree that the topic is interesting, but it makes me ill to just see him trying to spread fear and shit all around this forum. His tone is not neutral, indicating that he want people to make panic moves.


I agree with you. If you start a thread using a neutral tone and asking a question, fair. In this case the thread starter already have his opinion about x11 and just want others to share it with him. He is not asking a question, he just just trying to make it fail. Who knows why he is doing that.

Anyways I'm still interested in neutral answers with good fact, not speculations like the thread starter.

good read but still no answers...  

i have read that forking a scrypt coin other to another algo (eg. n-scrpyt, x11) is a very risky procedure and
could cause mass confusion and easily cripple a coin. devs of scrypt coins know this and probably would not
fork over to another algo. if this is true could they be spreading rumors themselves? knowing that there
are other algos much better than scrypt they could be purposely down playing x11 or whatever other algo out
there. just a thought.

'Conspiracy Theory with Jesse Ventura' will continue after these messages.

Hmmm

I think you are somewhat confused, i never said one was better than the other they are all cPoW;  X11 6hash X15 X "whatever" it actually is all a similar system.

The same miner will basically mine all of them.

but topics like this help stop some of the FUD

Believe or not, you will not get the real answer from this question because there are lots of X11 lover.

Probably better to focus on the question of X11's value rather than the Glen Beck-ish "I'm just asking questions here" conspiracy stuff.  X11 is open source, as far as I know.  Pick it apart and see for yourselves.  That's the whole point of open source, yes?

The DRK dev has never given any indication that he's anything but straight up and even makes his real identity known.  That's not proof he's not doing something nefarious with X11, of course, but really, what proof can he offer that would be conclusive enough?

If there is any evidence of malfeasance, by all means bring it to the table.  Have someone build a more efficient miner, if it's possible.  But until there's one shred of evidence to cast doubt, the conspiracy bit doesn't add anything to the discussion IMHO.  

I hate that stupid ass acronym "FUD", but this is pretty much the definition.

QRK is better than X11?are you sure?just for 3 random algo round?
however, "X11" is a fashion name, and more and more people accept it, that's the truth.but IMO, it's not very hard to make X11 ASIC,but will it worth doing?Need 11 diff asic to a single algo? and I think It's easy to change by a hard fork, X12 X13 X14 etc?
X11 may be the new beginning of GPU mining.
and ,  i am a miner ,  I like X11 cuz it's very cool on my card!

They are all cPoW Complex Proof of work, and yes its probably the closest to " fair" as you can get.

When LTC was released it was common knowledge that a GPU miner was out a lot earlier than people knew, this time around we have a GPU miner with a more optimization that has to be done.

The Power usage will probably still be good at "full" optimization.

Making cPoW the likely next contender for the largest part of the market , the Quark algo acts in exactly the same way as X11 as they are both " cPoW" .

im sure there will be more as well , and thats the idea.

That is relevant in what way to a discussion on x11 you silly girl?  when i see people trying to distract from the core of the conversation i know they feel uncomfortable ...... although how you've made it this far with your obvious brain damage i can only imagine.

I support all FAIR coins that just happened to be the fairest/cleanest launch of any coin i've seen yet. Anyway i am not hear to talk about beecoin with you, if you want to pm me i can perhaps give you a few pointers on starting a fair coin if that's what you're after. If you want to check my history you will notice i support a lot of fair coins i am not against all new coins. Sadly the fair ones do worse than the scams since the scams have motivated devs that want to cash out their instamine/premine

Get back to cleaning those toilets and quit wasting time here.

However if you're a real slapper i'd certainly be interested in meeting up and furthering your education in person.

Of course there are miners that mine twice as fast but not available to public. Maybe one day they will sell the license for the fast miner, right now they are using it themselves. That's why CGMiner is such a great piece of software, it was optimized to the max and open source. Too bad it stopped updating for Scrypt and The develer Never came back to update it for Scrypt-N.

One way to check for this is to run two, three or four instances and see if the sum of all four exceeds the single miner running. Can someone do that?

Whatever happened to BeeCoin? I heard that was THE shit.

To much hype with this not as good as they say.

Honestly, the car analogy is probably a bad example. Cars and computer hardware are two different things. For example, you have a high mhz cpu dual core vs a 4 core low mhz cpu quad core. Obviously, the high mhz dual core will win most single thread processes, but the quad core will win most multi threads. In a car's case, a high revving 4 cylinder car has to do high rpm to match the same horsepower as a lower revving V8.

I'm not trying to prove anything. Truly, if you wanted to know if it was efficient or not, you would need a programmer and computer engineer to debunk it or verify it.

Thanks that's what i thought.

If x11 is really a superior chained algo let's use it, but really we need to find out if it is or not. I was suggesting using it on a few threads and even was talking with a dev about switching from scrypt to x11 until he mentioned a few of the points i brought up here. I had just though there would have been a lot more questioning on this algo before everyone just jumped on it claiming it was the best thing out, so i just assumed it must be as great as everyone says..... then i was made aware there could be some serious issues with it.

Let's see if some super boffins turn up here to tell us straight if x11 is amazing or is a waste of time and nothing more than a marketing gimmick that seems to have gone viral with the noobs. All i hear is x11 x11 x11... be terrible if it turned out to be shit.

Yes, basically what I mean is that there's actually nothing useful to be done on parts of the chip. Scrypt is stressing more parts of the videocard, that's what I know for sure. Regarding if some things could be done to get more hash out of a card, if there could be better load distribution or work around existing bottlenecks, that I don't know.

To give you accurate not half-assed answers though, I would need to dig deeper into OpenCL, hardware articles and ATI/AMD manuals. I'm very tempted to do so, when I have some free time and peace of mind

we are not here to discuss your particular coin....it is the x11 chain we are here to discuss.  However if you wish to directly discuss your coin...

so far we have already called into question

1. that x11 is innovative.... it is not, it is less innovative than qrks own chain.... the algos are not even random and the block times are slower making it probably easier to attack too.
chaining algos is not innovative ( perhaps you are talking about the dark aspect of the coin but then again as yet it is talk - i guess we await zero coin)

2. that is it secure ? seem post 1 it seems less secure than qrk too with regards attacks, again untested so a bold claim.

3. that it efficient - this is open to question, the chain is not efficient - the miner is possibly intentionally crippled, or if not intentionally then it has probably been vastly improved so that most of x11 miners are getting raped by highly efficient optimised miners.

4. asic resistant ... again already mentioned it would be less costly to create asic for x11 than scrypt. It could already have happened. Far less costly then scryptN and scrypt jane with high N.


The thing about scrypt is in a few months everyone will have access to scrypt asics cheaply in some form or another.

Right now it is entirely possible there are those running highly efficient x11 miners that give the same advantage to a select few as having asics really. Even then there could be x11 asics too since the algos in that chain are not asic resistant. X11 saying it is asic resistant is a lot less likely than scrypt jane with high N or scryptN being asic resistant.

So you need to change your sig because it could be potentially be spreading misinformation and blatant lies.

So really relying on drks algo is looking pretty weak, it is looking like marketing hype over anything else. Show me some real advantages over existing high N factor algos or chained algos.
Also x11 is getting cloned by every single new coin coming out now so again this is no longer a real selling point for drk.

If i was a drk coin supporter i would be touting it's drk properties that could be end up making it a rival to a real dark coin like zero coin, not pushing the x11 algo because as yet i see NO evidence to suggest it was even worth coming up with.

Drk has a reasonable community and dedicated dev so i am not knocking it really, however let's discuss x11 not your coin.

e.g. with a p2pool node (http://drk.poolhash.org/) :


offtopic: off