hi, i am dangermouse,
let me introduce myself shortly: my gridcoin mining rig is an intel i5-6400 8GB with Nvidia GTX 1080, about 40 of magnitude, it is running in my garage on Ubuntu 16.04. i will build a second mining machine in August.
i am great fan of distributed computing, BOINC, computation in general. i run BOINC for a while in 2005 (water cooling failed), 2009-2010 (server hit by nearby lightning) and in 2016-2017 each time i had some rig out of old computer parts i put together to let it run. i discovered gridcoin in 2016, although i did not start mining back then because i read it was difficult to install and the entry barrier was deemed high. the bitcoin disadventure on mtgox discouraged me as well.
in fact i lost some money in the mtgox crash having some bots who did automatic trading around and all the money in the exchange, so for a while i was cryptocurrency skeptical. but now i am back in the field thanks to gridcoin. even if electricity in bitcoin is not wasted as it is spent to secure transactions in the blockchain, gridcoin does much better with proof of stake and proof of research. proof of stake is promised for ethereum, but it still did not happen. personally, i would crunch also without any reward, e.g. now i am still running climateprediction.net even if it is not giving gridcoins (because the jobs take too long? one took me 9 or 10 days...)
this article is about funding for the Global Processing Unit project (G.P.U), a distributed client similar to BOINC but based on a P2P approch. the idea would be to reward G.P.U. users with gridcoins, like it is done now with BOINC users. The G.P.U. project homepage is at http://gpu.sourceforge.net
I started G.P.U in 2002 when i saw the potential of gnutella and seti@home and then along the years other hackers joined and did remarkable work, nanobit in particular was way better than me. thexa4 did well despite his young age. simbasabi built the best rigs out of the worst spare part crap. we all felt in competition with BOINC (it also started around 2002 on sourceforge). somewhere i have a picture taken from sourceforge statistics showing G.P.U above BOINC... in the distributed computing foundry of sourceforge. i found it!
we tried to made plugins/frontends for each project on BOINC, so we ended having a terragen movie generator (->BURP), a distributed search engine (there was a distributed crawling project not linked to BOINC), a climate simulation (-> climateprediction.net), an orbit reconstruction plugin (code borrowed and translated from orbit@home of Pasquale Tricarico), golomb rulers, discrete logarithm, a whiteboard where it is possible to draw in distributed fashion and share a webcam, a distributed wan database (a bit like MongoDB), a screensaver, an earthquake sensor ...
you can find pictures of the extensions above on the G.P.U. homepage and you can give a try to the client, although it is not developed since 2013. we tried to build a G.P.U. II, but this never went productive.
the G.P.U project still had some recognition as it was downloaded around 85'000 times and we had 5-15 clients kept running between 2002 and 2010.. but maybe we have a better name, like OuterSpace, DeepSpace something like this?
everything was developed just for fun and over the years the project lost users and interest, as in the end our amateurial development could not compete against BOINC. still back in 2002 there were other competitors with sounding names like "Globus Toolkit"... and other approaches like OpenMosix. grid computing was the buzzword of the time, i remember it quite well :-)
my goal would be to connect the G.P.U. project with gridcoin (yes, the name is unfortunate as it can mix with graphics cards, i chose it in 2002 when i was unaware of the GPU acronym for graphics cards), so that people running G.P.U are rewarded with gridcoins. this because i realized that without a financial incentive, the G.P.U network died. The G.P.U. client needs a rewrite anyway. I was planning to move from a P2P topology to a client/server infrastructure as G.P.U had scalability problems when running with more than 20 nodes.
some things i would like to ask (excuse me i am a newbie and i might be off topic):
did you think about implementing a timestamp service on the blockchain, so that one can bring evidence he created a particular document at a particular time? it would be enough to hash the document and store it in the blockchain as a particular transaction. there is some startup doing this on etheruem. i saw also a dao command in gridcoinresearchd, does it refer somehow to the "infamous" DAO of ethereum?
But i still feel that G.P.U. could become for gridcoin what Golem will be for Ethereum: a platform where it is easier for "little fishes" like me to implement some dis.co. projects. As advantage compared to Golem, computation providers in gridcoin/G.P.U. do not need to pay anything, the burden to create something computational intensive and tailored to the platform is big enough.
to come up with something real, i would like to propose to the Gridcoin foundation a "30" deal:
if i would get paid 30 hours à 30$ in GRC i will deliver a 30 page document on October 30 2017 analyzing in which way G.P.U could be connected to BOINC. in the document i will discuss several variants (but not 30 ;-).
One variant will be having G.P.U. to mimic a BOINC client: G.P.U would download and upload tasks and issue scheduling requests to the server in the same manner BOINC client does
Another variant would be to set up a compliant BOINC server in respect to reporting work done, but different in the way in how it distributes tasks to the G.P.U clients attached.
Even another approach would be to have a G.P.U. server which acts like a BOINC mining pool.
And so on, for sure you have some ideas, too?
The payment should be done half on July 30, where i will start working on it and the other half on October 30 upon satisfaction of the Gridcoin foundation. In positive reception case the document could serve as request for further fundings for the development of the bridge between gridcoin and G.P.U. for the variant chosen by the community. of course, it will be for me a duty and an honour to invest much more than the 30 hours to get some sound document...
If you would like to contribute to this endeavour, the funding address is SEtxQ4SePHSP2xSfjy4MATvCWGoFMCcahn
thank you for reading through this, what do you think?