Hello there, everyone,
For the past few days, I have been working on a new project. Because I wanted to learn more about how the Bitcoin works internally, I decided to give it a go and create my own miner. But not just an ordinary miner, no, something slightly different.
I present you:
{CurlyBrace}, the first PHP Hypertext Preprocessor Miner (with a little bit of JavaScript as well)
(Disclaimer: This is a proof-of-concept. But it mines. For Real!)
Some basic info:Language: PHP, and JavaScript for the front-end(steering wheel) of the miner.
Type: CurlyBrace uses the getblocktemplate standard.
Speed: Slow(0.1Mhash/s), but will hopefully allow paralellization in the future by talking to multiple webservers at once.
Why this odd language choice?Using PHP, a serverside web programming language has certain interesting characteristics, and I was interested to see how well it would perform.
Pros:-You now have some use for the normally unused processing power you have in your monthly server subscription.
-You don't have to pay for electricity costs during hashing on a remote computer.
-Opens up the possibility of setting up many servers (e.g. using free web hosting) in parallel, without much effort/costs.
Cons:-Hashing is of course not fast. Tests thus far return about 0.1Mhash/s per slave webserver.
-Your worker username and password are visible by websites you are using. When using a cross-site AJAX solution, this data is even more visible.
What does the code do right now?-You visit the webpage with the javascript client on it.
-Client asks the pool for a getblocktemplate reply.
-Client sends the returned block data to one or multiple server slaves, asking them to hash until they finds a hash below the target.
-Server hashes until it finds a hash below the target or it has tried 1000000 hashes(0.1Mhash). Returns results back to client.
-Client shows the debug info the server returned.
-If no correct solution is found yet, client sends a new request, but with an incremented nonce, to ensure that each server is looking at a different part of the possible hashes at one time.
-This continues on until a solution is found.
-Client submits solution back as a share to the pool. (this part might not be fully functional yet)
What still needs to be done?-Support to increment the extra nonce (change the coinbase) when all 32int nonce values have been tried. (Right now the code is so slow that this never really happens before other parts die in one way or another, but its something to consider in the future.)
-Possibly change the way the final solution is padded together and submitted as a share (Might not work yet).
-Stop searching and load the next block when the current block has timed out. (Right now the script will continue searching on the current block until a solution has been found by it)
-Enforcing that the correct getblocktemplate and block version is used.
-Support for the optional parts of the getblocktemplate specification.
-Bugfixing. Lots of it.
-Possiblility implementing a scrypt algorithm as well(started, but currently disabled.)
Why the need for a client webpage to operate the miner?1. PHP is executed in a synchronious fashion, JavaScript on a web page is not. This allows parallellization(talking to multiple servers at once) when using JavaScript in the client side.
2. Very easy to start and stop.
3. I've had trouble getting cURL to work in PHP with ports different than port 80(which is the default HTTP web page port, but is almost never used for a bitcoin mining address).
Also, many PHP hosts don't allow cURL to be used at all.
Thus, using Javascript together with an Ajax Cross Domain service (such as
YQL, which I'm using right now) seems like a better option.
Can I try this myself?Yes, you can
download the source here, upload it to your favourite webserver and direct your browser to
http://yourserver.com/bitfrontend.htmlAlternatively, you can currently see the script working
here. Note that you shouldn't keep this version running for a long time as it doesn't use your own computing power but a few free webservers. I don't want to choke them. ;-)
Why do you post this here in such an early stage of development?Basically I would love it if some people took the time to look through my code, and tell me if my calculations of the bitcoin hashing is correct. The main reason I started this project was to learn the inner workings of the bitcoin, so this would mean a lot for me.
I will be developing this project further during the next couple of weeks. But if there is anyone interested in contributing, feel free. There's a lot of optimizing and cleaning up to do ^_^'.
Thank you very much for your time. If you'd like to donate, you can thow a few coins in my direction here:
1MUZQsCSifF27XX98W13JLz6E9YaU3mVqG .
Have a great day!
~W-M
