Topic: [ANN] [XMG] MAGI | CPU mining | mPoW | mPoS | [MagiPay] - page 815. (Read 2375972 times)

I'm a computer science professor.  Much as you were pondering, I occasionally hack crypto miners as a diversion from my real work -- much like a lot of the other miner hackers around here.

You can see a bunch of my prior explorations on my blog - http://da-data.blogspot.com/

Nice! Naah, I love what I do. Coding I do for work is different to say Pebble or game development or general tinkering I do in my own time. 50% speed up? Wow, I take my hat off to you - that is a very impressive speed up! What's your background out of interest?

btw, just to keep you on your toes, I went ahead and did the next easy increment.  It's now about 50% faster than the public one.

Do it!  It's tons of fun the first few times (assuming you're not feeling burned out on coding from the day job).  It gets a little repetitive after a while, but m7m's an amusing enough hash function that there's some meat to learn on in there:  multiple hash functions and a cute little bigint mixing function implemented using GMP.

Funny you mention this, I'm a developer and was planning on doing a hotspot analysis on the code to see where the bottlenecks are. I've optimised code in the past for various companies and I work with C/C++ every day so I was tempted to take a look to see if I could eek out some extra performance from the miner. Now I know someone else has had success it makes me even more eager ;-)

Just a quick pointer - if folks are interested in putting together a small bounty for open-sourcing (or if someone wants to buy it, privately and exclusively) a 20% 30% faster version of the miner.

https://bitcointalksearch.org/topic/m.11788473

I put it in the mining discussion thread, but that one's pretty low traffic, so I'm pointing to it here.  2 BTC gets it, public or private, to cover the time I put into it;  10% goes back upstream to Wolf as a thank you for open-sourcing the version he released that I built on (saying that publicly here so you can hold me to it. grin.)

That's right.
But the pool admins might have some tools for that!

That's great.. but can I ask a potentially silly question? Limiting the hashrate won't stop people from creating multiple user accounts and splitting their hash power between them will it?

I will give a respons on your long post later Lightsplasher. I just need to share this great info about http://minerclaim.net/ our newest Magi mining pool added the max hashrate to their rules.
Thats really awesome! Special thanks to minerclaim.net!

The Magi pay site looks very interesting.  I've been thinking about trying to integrate this into the web application that I've been thinking about building.  (Nobody hold their breath waiting for me to complete it please, I've a slight tendency to plan on things that I've troubles getting done)

Just a couple of thoughts on clarification and potential enhancements that might be useful.  It states that it takes an hour for the payment to be received by the merchant; is this so that the block chain can be confirmed?  Is there a way for the merchant to be informed that the customer has started the payment process at the time they make payment so that they could perhaps be granted limited access to resources on good faith?  Is there any type of arbitration or escrow type service planned for the customer perhaps so that web merchants that accept XMG through the merchant services can be trusted?  (Maybe some customer feedback or other indicators that the merchants are legitimate?)  Perhaps even requiring a certain amount in an escrow balance from the merchant would be enough to establish good faith on the part of the customer and not burden the merchant services (acting as a neutral third party) with this.

Example: Lightplasher to offer a jar of blackberry preserves through his website at $5.00 USD or equivalent XMG.  It would be nice to receive payment in XMG and not worry about how customer originated the payment (USD, Greek coins, using Paypal on Magi Pay, etc.).  Once customer starts payment they are granted access to the special web customers only access page.  After further inspection and reading reviews the customer decides that Lightplasher is far too esoteric to order blackberry preserves from and requests a refund through the merchant services.  If the dispute arises prior to shipment the customer easily gets refunded through merchant services.

Perhaps you all can see why I mine cryptocoins instead of working for clients or canning blackberries.  (I should get 20 or 30 jars of blackberries this fall out of my backyard – any takers?) 

Yea notice the link sometimes without the ...s and than its not working!

Yes that was the right one

There are some wrong links for Magi standalone payment gateway.
This is the right one: https://www.m-epays.com/

Don't be too late! Register if you would like to join!!

Pay per last N shares ... just google for PPLNS to get details how it works. Not very useful for Magi, but I think some pools use it.

Aha, ok.. what is PPLNS out of interest?

If PPLNS is OFF then when a new block is found by your pool you get BlockReward/TotalNumberOfShares*NumberOfYourShares-PoolFee%-Donation% If PPLNS is ON then it is calculated little bit differently to discourage pool jumpers.

"expected shares" is just a reference value.

So.. I have a general question about mining pools, specifically what happens when the number of shares is considerably higher or lower than expected.

For example, let's say for a block a pool expects 100,000 shares but it only took 10,000 shares in order to generate the block (10% of expected work). Does this mean that the rewards per share are greater for each of the 10,000 shares? If the share reward was the same there would be a surplus of block reward as only 10,000 or the 100,000 expected shares would be paid.

Similarly, let's say a pool takes a long time to find a block and there are 1,000,000 shares when 100,000 were expected (10x or 1000x expected work). In this case each share has to be less valuable because there won't be enough coin to honor the expected share payout.

Is my understanding of this correct ?

Maybe because then the infected botnet open it's offices...

I've noticed a pattern.. at about 10am each day (UK) time the bulldowzers join the party