I checked the handshake bytes and the genesis hash they looked ok, the ports matched too so yeah it looks okay so far, it is getting the blockchain from one of the old nodes now...

-MarkM-

Ah well, maybe its time as a small miner's coin for CPU mining is ending then, since already i was starting to think maybe someone uis using a GPU on it, or at least quite a few CPUs, and with newer code, that includes a daemon, how likely is it that all the CPU miners will be able to upgrade to the new code without some GPU bully or bullies jumping in to drive the CPU miners away?

It'll be nice not to have to run stupid GUI crap across continents from server to the desktop graphics-server though.

Have you checked the ports, IRC channels, and especially things like the addresses-version and magic handshake bytes aren't just what was inherited from litecoin? That is, did the guy actually do the conversion to faribrix or merely clone litecoin itending to then never actually finish the job?

-MarkM-

Has anyone been using 256MB systems successfully? I thought they were using 512 and up...

-MarkM-

Yeah, I had kind of imagined the way to use primes for work would be your job is to prove that numbers that look darn likely to be primes are NOT primes.

-MarkM-

It inherits from bitcoin, eh? Not from litecoin? That might be good because I haven't really noticed any mention of litecoin having bothered with any multi-sig stuff at all... Though presumably when litecoin does eventually get around to rebasing itself from bitcoin again it should hopefully then finally aquire all that stuff that bitcoin has gained since litecoin long ago forked from it.

-MarkM-

Look at /var/log/messages, or maybe it might be in the output of the command "dmesg", to see if there is mention of killing the task.

The make says the thing was killed, using too much memory kills stuff. GeistGeld, that likes to eat itself out of memory, gets killed for example when it uses too much.

I just don't recall offhand whether it was /var/log/messages or the 'dmesg' command, or maybe even both, where it was recorded that the thing had been killed.

(It was something like hmm gotta kill this or its child tasks, hmm, okay think I'll kill it, type of deal it seemed.)

So yeah throw more RAM and/or swap at it. Not sure if in all cases swap is as good (other than slower) as RAM or not, especially in virtual setups, and of course especially if ratio of swap to RAM gets into swap-thrashing territory where the swapping itself needs more RAM and so on in a loop where no real work ends up getting done. A quarter of a gig of RAM seems really tiny, but hopefully the kernel isn't included in that but is outside in the virtuality engine? Or is it?

-MarkM-

How much RAM and swap?

-MarkM-

Right, which probably means there could easily be plenty-enough CPU miners already to easily account for current difficulty without even starting to think about botnets and GPU miners.

Not to mention, but I'll mention it anyway, that in some other thread, about using GPUs, it was recently suggested only two times speedup might be obtained using GPUs.

Which contrasts a lot with earlier posts implying this particular application might get 1000 times speedup using GPUs. i wonder what the real speedup will turn out to be...

If its only twice as fast then using two or four CPU motherboards might be better than diverting GPUs from other things that GPUs are much better at than CPUs.

-MarkM-

You need a pool ready to go at launch, because the only way to come anywhere near your target time per block at launch is for the difficulty to be so high that small miners cannot realistically solo-mine. That is partly due to how huge the large farms are compared to a small miner and partly due to how many large farms will jump on, and, provided a pool is already present at launch, also the number of small miners who will jump on.

Higher difficulty without a pool discourages small miners from jumping on, so mostly it is the large miners a high difficulty needs to be tuned to, so if no small miners try (they wouldn't get blocks anyway at low difficulty their blocks would be orphans) maybe not quite as high would be needed, but still probably at least 0.7 or 0.75 and maybe if a pool is ready then 1.0 might give at least a "we tried, we really did" kind of look to the launch.

(Though likely even 1.0 would still result in blocks coming much faster than your target time because miners of all scales will all be able to mine nicely yet possibly maybe there wouldn't be as massive a horde jumping on your coin as happened for example with Primecoin (for which just one of many cloud providers saw 18,000 people sign up for accounts to mine the thing, and even despite Sunny trying to make it difficult enough still managed to get to a point where blocks came so fast orphans were starting to proliferate for a while.))

Having a pool in place means it does not matter if you start the difficulty too high, because even if it is too high for the largest farms to solo-mine they too can join the pool if it is too high for them.

So, try to make it way to high, thus maybe managing to make it not too insanely too-low. But still likely it will turn out to be too low anyway.

Oh also, maybe, to counter the "over 51% on one pool is bad" argument, release the source code of the pool at launch too, ready for others to simply plug in their domain name or whatever (or maybe they won't even need to do that) to have it run for them on their server out of the box too.

I saw a service post recently that offers wannabe coin-launchers a whole package that includes a pool...

-MarkM-

Primecoin and BBQcoin, the latter for the awesome way it helped small miners by giving them a year or so of CPU mining without being driven away by GPU miners yet still managed to give them a massive profit by eventually attracting the GPU miners back, getting on exchanges, and maintaining a not too shabby price.

-MarkM-

By the way, since, unlike bitcoin, it is specifically designed for the kind of environments where multi-sig would be essential, what kind of features does it provide management for dealing with multi-sig, is its multi-sig user-interface far more user-friendly than bitcoin's for example? Does it maybe default to all addresses being multi-sig, at least when in business mode rather than player mode? Or what?

-MarkM-

If a GPU is only twice as fast as a CPU, maybe putting more CPUs on one's motherboard might be more cost-effective than putting one or more GPUs in a machine?

Or think blade servers, two or more CPUs per blade, would GPUs really be more cost-effective?

-MarkM-

Ah, instamine. Figured.

Even just adult-sites need massive security compared to most sites, casinos likely are worse, with that difficulty it seems kind of totally bonkers batshit insane to even touch this thing for a casino, heck have they even yet really decided whether bitcoin, with its massive force of mining gear securing it, is secure enough for them to adopt it?

-MarkM-

Ha ha ha what a roundabout way of bringing up the likelihood that they haven't actually tested this thing with an actual casino, let alone even actually have a casino, so really any stuff they did to make it easily integrate with casino software and easy for casino users to use to interact with casinos with is purely imaginary, no real ergonomics / interfaces / etc testing having been done...

Maybe it'd be interesting to also know how many person-years of actually running online casinos their design team had before finalising this design?

-MarkM-

Starting difficulty?

-MarkM-

Maybe, if you only count the extra CPU-usage, that is, if you already have the motherboard and internet connection and whatever else you leave attached (disk? mouse? keyboard? whatever) running 24/7 anyway.

However a lot of "normal people" seem not to normally leave their machines on 24/7, so for them it might be more like "a normal PC uses 200 watts, which is one kilowatt-hour per five hours, which comes close to five kilowatt-hours per day, assuming your machine really does not go over 200 watts at the wall when cranking all its cores full blast".

I used to figure turning on a machine cost about $30/month at $0.14 to $0.15 per kilowatt-hour until I got a kill-a-watt meter and learned my box with a 5870 in it was eating 240 to 250 watts before even firing up a miner to mine with that GPU, and more like 340 to (once warmed up) 370 with it mining. Admittedly that was before taking it in to have massive amounts of dust removed from inside and new thermal paste for its CPU heatsink so its CPU would not keep throttling from overheating and its fan control wouldn't keep blasting the fans as loud as they would go.

So when its a matter of whether to plug in or turn on yet another of the machines the air conditioner is trying to make you willing to put up with the noise and heat of, I think your "at most 2.5 kwh/day" is somewhat optimistic.

I have a 12000 BTU air conditioner outside in the minivan waiting for the guy who is making us new windows to come by and install because the 8700 BTU unit currently in my current window cannot handle the heat if I turn on (in addition to what I have on 24/7 already) the four Dell Optiplex 755 boxes. So depending on the season/weather even just the turning on another computer estimate is a little low...

(Not to mention, oh heck with it lets mention it, that I also had to provide an air conditioner (I got a 6000 BTU) at the other end of the apartment so others here don't get too overheated.)

-MarkM-

Maybe there are just too many regular CPUs all fighting for a piece of the same pie?

Pools maybe help miners figure out faster whether they are losing money, whereas with solo mining high variance could easily lead to people hanging on a few more days, lucking into maybe even a few blocks one day so hanging on for even more days before lucking into another block which might make them wait longer hoping that was the first block of a lucky burst of blocks and so on.

Intermittent reinforcement is illegal as a pay method in some jurisdictions, and even illegal in the form of slot machines (aka gambling) in some jurisdictions, because it is pretty much the best reinforcement schedule to use when you want to wean the workers out of being paid without having them quit...

So it could very easily be that too many people are running too many ordinary CPUs on this with too many of them thinking that if they hold out a little longer while others quit they will start getting lucky again. Meanwhile how many more thousands of new people are starting up, taking a few days to see for themselves whether maybe their build and/or their CPU might be a little luckier than average, and so on... all the usual stories gamblers tell themselves to convince themselves to put in another slug and pull the lever once again... maybe just once... well maybe just once more... well heck its got to work this next time... etc

And of course all along more and more people who have some amount of electricity they consider basically "free" (already part of rent, whatever) are probably getting involved, they can keep going "forever"...

(I have servers out on the net I already paid for by the year in advance, got anything better for me to have their CPUs work on 24/7?)

-MarkM-

By using enough RAM, maybe? How much RAM you got and how much swap space?

-MarkM-

That is why the SHA256 coins created merged mining; since miners do not need to stop mining one coin in order to mine another, but can instead mine all of them at once, they don't really have much if any reason to attack each other, instead they all pool their hashing power to secure all the chains at once. (In theory; in practice a lot of miners just don't care about a few percent more income so leave that income on the table for those who do find a few extra percent a worthwhile bonus, and of course by doing so they also increase the percent those extra chains add up to for those who do like more income.)

-MarkM-

Oh yeah that happened with DeVCoin too though we thought it was to do with the curl stuff DeVCoin used to go out and get its lists of who to give coins to. Dunno what Tenebrix would have needed that would not be thread-safe and that other coins don't use though.

For DeVCoin we just use the daemon for mining, people who need GUIs probably aren't the geeky techie types likely to get into setting up p2pool to merged mine for them or stuff like that so for DeVCoin the GUI not being able to mine isn't much (if any) of an issue.

-MarkM-