Topic: [ANN][YAC] YACoin ongoing development - page 178. (Read 380060 times)

Windmaster, are you willing to share the code you developed for opencl ?
Maybe we can work together to improve it or test anyway.

This thread is worth already a 20% jump.

I suspect a lot of people had the wrong impression about what would happen when N changed from 32 to 64.  My impression from the official YACoin thread is that people were expecting an instantaneous drop of difficulty and a simultaneous instantaneous increase in the block reward.  The reality is that difficulty is based on block solving rate of the network at large, and only indirectly affected by N as the result of N affecting hash rate of the network at large.  And block reward is derived from difficulty.  The change from N=32 to N=64 caused about a ~20% drop in hash rates for the available CPU miner implementations.  I note that difficulty has indeed been dropping.  I've looked at it a few times today and it's down to about 8.1 now (and was 8.25 a few hours ago).

The rate of difficulty recalc after each block certainly could've been faster though.  It seems to move at rather glacial speeds.  We can't change that now though.


This one I'm unsure about.  I thought for sure that if there were a significant number of GPU implementations floating around, someone was going to post one by now (as a claim to fame) as soon as GPU mining would have become unprofitable (which by my calculations is probably getting close to becoming unprofitable as YAC prices drop).

Anyone that has followed my posts through the days after the coin launched knows that I experimented with GPU implementation of scrypt+chacha20/8(N,1,1), the hashing algorithm used by YAC.  I never got as far as making it work with cgminer or mining with it.  I just created a basic OpenCL kernel (no cgminer interface), performed some benchmarks at various values of N, didn't bother doing any major optimizations, and decided I needed something that would go significantly faster (my GPU farm only consists of 24 Radeon 5850's and 24 Radeon 6950's), as it took me about 1.5 days to do it and difficulty was already skyrocketing.  At that point I dropped it and went straight to writing a Verilog implementation for FPGA's, knowing that I could probably only pull that off up through N=32 and would have a limited time to make it happen (in the end, I got about 1.5 days of FPGA mining in before N=64).

We won't know for sure until someone posts a modified cgminer what their hash rates ended up.  I stopped mining altogether at N=64, as that basically killed viability of my FPGA implementation.  I can still place'n'route an FPGA design at N=64 but the projected clock rates from the Xilinx tools are dramatically slower than for N=32 with nearly 4x the logic area (meaning only 1/4 as many parallel computation cores are possible in the available logic area) and it just doesn't look like it would be worth the effort.  N=128 won't even place'n'route, the design is just too dense.  So at this point I'm going to state that I believe N=64 kills profitability of any FPGA attempt (with current FPGA technology) and N=128 isn't even routable and placeable on current FPGA technology (at least without going to off-chip memory, DDR3 or otherwise).

Currently I get far more YAC by buying it up on BTER than I would by mining, with any of the technologies I have available.  My GPU farm is mining LTC right now.  If there are indeed GPU farms mining YAC, I'd be expecting them to start switching back to LTC or BTC for profitability reasons either now or in the near future (well, unless YAC value goes way way up).

So, that's about all the answers I have at the moment.

Yes, plus YAC is listed already and it might confuse possible users if there is a second one. A relaunch only makes sense if there is a need to change the concept.


If you are talking about me: don't worry. I haven't mined that much with my i7. I was also extremely lucky. I opened the browser in the morning and thought, hey, why not try to compile sth. in Linux? I just had installed VirtualBox a day ago...
edit: I have mined far less YACs than people bought on the exchange.

And you are absolutely right: it is and was very easy to get a significant number of YACs on the exchange. The money supply of YAC is worth ~1200 BTC right now. Maybe that's a fair price for a successful launch and listing at an exchange and some developers + pools.

I am also happy to contribute, but I do not have much time.

This is true.  I've bought more YAC on BTER than I mined.

I am not a tech guy so cant help with any technical issues. However happy to pay bounties (1K now and 1K in future) for any worthy cause. PM me when the time comes. I know its not a lot but I don't have a lot YAC too.

I think we should re-launch a memcoin, based on scrypt-jane, init N=2048.

I've not noticed any difference on may 14th when the N++ event happened.

1) difficulty was supposed to drop - I left nothing
2) GPUs were supposed to take a hit - AFAIK they still own the network.

Answers?

PS. YAC has been so cheap these last days that anyone who missed the launch could overtake any early adopter just throwing a few BTCs on bter.com

nope all fair points...
My issue was one that if the orig dev comes back, he may change something and make all your work go to crap....

Creating yet another new alt-coin is outside the scope of what I'm trying to do here.  YAC is already underway with an economy (however small) built around it.  While the launch was not perfect, and I myself got screwed on the launch as I wasn't expecting it 1.5 days late (and thus, I got an 8.5 hour late start before starting mining), it appears to have been the closest to legit coin launch in the recent past.

I participated in the Elacoin launch the other day and my personal opinion is that one really takes the cake for launch shenanigans and multiple botched launch attempts.  Complete with the usual disappearing developer, the developer adding checkpoints that rewrote the blockchain and invalidated large numbers of legitimately mined blocks, and further, by the remaining miners proposing to hard fork their blockchain and rewrite the client to increase the mining rewards to themselves and jack up their rate of inflation.

In comparison, I'd say YAC had significantly closer to a clean launch.  Still not perfect, but I'd say it was the cleanest launch so far among the recent alt-coins.  I think the primary complaints about the YAC launch were the missed launch deadline catching everyone off-guard, low difficulty, and pocopoco failed to get the Windows client optimized for acceptable solo mining hash rates and get UPnP working properly.  In comparison, Elacoin launched with no Windows client at all, developer had nearly zero experience and should never have contemplated starting an alt-coin, was reportedly (if believable) trying to mine the genesis block on a netbook with Intel Atom CPU, it was basically a Litecoin copy'n'paste, and difficulty was 0 (which resulted in the client actually segfaulting and crashing on each REORG, much to my dismay as I was running one of the Elacoin seed nodes everyone was using).

I suspect the only reason to start a new coin, with no attempt to innovate, is as another pump'n'dump.  There's already enough of those already.  YAC was based on a new twist on NovaCoin that (at least temporarily) leveled the playing field between CPU and GPU mining, and thus I believe this coin launch actually had merit.  Forking it, renaming it and starting over without making significant improvements would produce just yet another copy'n'paste coin with no purpose other than to enrich whoever is launching the new coin.  However, anyone is free to fork any of the coins and start their own.  It's open-source afterall.  Is it a good idea to unleash yet another alt-coin solely to use it as a pump'n'dump?  Not really, in my opinion.

If someone is disappointed they missed the launch, I'm afraid I probably have nothing to offer.  Remember, I missed the launch too, by a whopping 8.5 hours.  I had to throw massive amounts of CPU power at it (800x dual-processor Xeon servers, 760x Amazon AWS instances) and even went as far as custom modifying a set of four LTC FPGA prototype boards (which were significantly slower than GPU's for LTC mining so did not evolve into a successful product) with 20 Xilinx Artix-7 FPGA's each to mine YAC instead while it was at N=32 (and very easy to optimize for FPGA's).  Yet with all that, I apparently mined less than people that got a single i7 PC in play right at launch.  I'm not sure there's anyone here that would benefit more from a complete YAC relaunch than I.  If I were pushing for a relaunch of YAC, it would definitely just be another pump'n'dump "make the developer rich" scheme, as I could completely rape the sh*t out of the new coin with the amount of processing power I fired up last time around (around 3x more than someone could pull off with a single Amazon AWS account).  It comes down to motivation, I believe the first YAC launch was actually valid, but do not believe forking it for the purpose of relaunching it as a brand new alt-coin would be valid.  As much as I'd like to relaunch YAC as a new coin, and as much as it would probably result in me mining the lion's share of that coin, that's just not a valid reason to do it.

Hopefully everyone complaining about being late to the YAC launch will realize that I probably have more reason to be upset about being late than anyone.  Can anyone here claim to have been more butt-hurt than I about missing the launch?  Considering the amount of resources I had to bring to bear to mine less than people with a single PC at launch time?

this is good, and ill help contribute were I can....but....

Can we reset the coin, get rid of yac and all association....
ie this is a new coin forked from yac, with a new name etc etc

GREAT STUFF

Good! We need an active dev for YAC.

Good to hear someone willing to share is also working on the problem.  I'd be happy to throw you my measly hashrate to help test if needed.  I've setup pushpool/mmcfe for ltc in the past (mmcfe is indeed messy in a bunch of ways).  Look forward to throwing up a pool when things are stable.

Let me know (pm) if I can be of any help.

Awesome News!!! Go Go YAC!!

nice!!

Agree, coins should be currency and people should sell and buy with it for coin to be anything other than speculative commodity. It's the only way for long term success.

First of all a big thanks to WindMaster for stepping up, YAC is indeed different from the other me-too coins and it's a real shame pocopoco pulled a Satoshi. I can and will help with code.

Now as to pools, I am test running a pool (pushpoold and mmcfe), both modified to work with YAC. Pushpoold hit a snag when N increased so all shares got refused. I have a temporary fix in place, but due to low hashrate (+-400 KH/s) it might take some time before I actually mine a block on my pool to confirm it really works. As soon as I know it's actually working I will put my changes on github. Aside from pushpoold there's also mmcfe which frankly is an abomination, horrible horrible code full of bugs and nasty shortcuts. Also once I'm sure all my changes work as intended I will publish them. I share your opinion that having properly working pools is critical to the success of the coin in the long run.

Of course if some of the pool operators who have confirmed working implementations already would be so kind to publish their changes instead of riding the greed wave it would be appreciated.

Regards,

cK'