Tumen Miner: Difference between revisions
Jump to navigation
Jump to search
created stub and quote of author notes |
No edit summary |
||
| Line 1: | Line 1: | ||
Tumen Miner is a webcl miner originally created from [[Krad Miner]]. It's codebase is located on [[Temujin9]]'s [https://github.com/temujin9/tumen_miner/branches github account]. | Tumen Miner is a webcl miner originally created from [[Krad Miner]]. It's codebase is located on [[Temujin9]]'s [https://github.com/temujin9/tumen_miner/branches github account]. | ||
Original Author Notes, 2011/12/24 | Original Author Notes, 2011/12/24 | ||
Here is the dump, I may have other versions, but this was the first thing I found, it maybe the best I have. I think the important javascript parts were miner_original.js and worker_original.js (which is missing) but you maybe able to de-minfy the worker.js file. | Here is the dump, I may have other versions, but this was the first thing I found, it maybe the best I have. I think the important javascript parts were miner_original.js and worker_original.js (which is missing) but you maybe able to de-minfy the worker.js file. | ||
The whole client side of the system was just the index.html page, and those two javascript files. IIRC the 'web-worker' can't create an opencl context so all the work is done in the main javascript thread. The backend is just a proxy to a bitcoin miner, the reason it exists is just because of javascript security you have to connect to the same url/port, so that can be replaced by a modern mining proxy of some kind. | The whole client side of the system was just the index.html page, and those two javascript files. IIRC the 'web-worker' can't create an opencl context so all the work is done in the main javascript thread. The backend is just a proxy to a bitcoin miner, the reason it exists is just because of javascript security you have to connect to the same url/port, so that can be replaced by a modern mining proxy of some kind. | ||
My advice would be to study how webcl works, and get a demo of that going in your own code, then learn how the mining process works, and try to load the opencl kernel using your own webcl javascript code. then look over the worker.js and miner.js files and pick out the pieces that actually do the work of converting the bytes around to the right formats and talking to the webcl kernel. Looking at the source of poclbm and diablominer should give you an idea of the various byte conversion functions needed, krad miner works in the same ways as these miners do, only the code is in javascript and you have to deal with the strange different world of javascript. | My advice would be to study how webcl works, and get a demo of that going in your own code, then learn how the mining process works, and try to load the opencl kernel using your own webcl javascript code. then look over the worker.js and miner.js files and pick out the pieces that actually do the work of converting the bytes around to the right formats and talking to the webcl kernel. Looking at the source of poclbm and diablominer should give you an idea of the various byte conversion functions needed, krad miner works in the same ways as these miners do, only the code is in javascript and you have to deal with the strange different world of javascript. | ||
Latest revision as of 03:25, 25 January 2013
Tumen Miner is a webcl miner originally created from Krad Miner. It's codebase is located on Temujin9's github account.
Original Author Notes, 2011/12/24
Here is the dump, I may have other versions, but this was the first thing I found, it maybe the best I have. I think the important javascript parts were miner_original.js and worker_original.js (which is missing) but you maybe able to de-minfy the worker.js file. The whole client side of the system was just the index.html page, and those two javascript files. IIRC the 'web-worker' can't create an opencl context so all the work is done in the main javascript thread. The backend is just a proxy to a bitcoin miner, the reason it exists is just because of javascript security you have to connect to the same url/port, so that can be replaced by a modern mining proxy of some kind. My advice would be to study how webcl works, and get a demo of that going in your own code, then learn how the mining process works, and try to load the opencl kernel using your own webcl javascript code. then look over the worker.js and miner.js files and pick out the pieces that actually do the work of converting the bytes around to the right formats and talking to the webcl kernel. Looking at the source of poclbm and diablominer should give you an idea of the various byte conversion functions needed, krad miner works in the same ways as these miners do, only the code is in javascript and you have to deal with the strange different world of javascript.