Google Trends Revisited
It's already a year I have published a post on BOINC and Gridcoin in Google trends, I thought it's worth to check it out again.
BOINC is in a slow downtrend, while interest in Gridcoin had some ups in 2017, but quickly diminished in 2018. It is quite interesting to look into second half of 2017. Long term BOINC downtrend seems to be pushed back up by growing interest in Gridcoin. There is correlation and I expect also cause-effect mechanism, I doubt it's only a coincidence. After Gridcoin lost its momentum interest in BOINC went further down to 5 year lows.
Above weakness is confirmed by BOINC site itself.
Date | 6/01/19 | 26/12/18 | 10/01/17 | 14/01/16 | 30/01/15 |
---|---|---|---|---|---|
Active users | 128,143 | 129,021 | 178,224 | 238,559 | 235,315 |
Computing power PetaFLOPS | 36.4 | 34.2 | 21.9 | 11.2 | 8.5 |
There is a positive aspect, too. Network computing power is growing and reached over 30 PetaFLOPS in December 2018.
Interestingly, today on 5/01/19 BOINC reports whooping 169 PetaFLOPS! I wonder whether it's some bug or some powerful party joined some projects? EDIT: Updated table with more consistent data from 6/01/19.
For comparison, until last year most powerful supercomputer was benchmarked at around 125 PetaFLOPS, and Summit supercomputer launched in 2018 dwarfs others at 200 PetaFLOPS. Will you guess what operating system runs on these supercomputers?
Conclusion
Both BOINC and Gridcoin communities need to start thinking out of the box to turn around these underwhelming trends.
It's interesting the increase in computing power despite the reduction in active users. I think that computer hardware performance may continue to outpace the loss of computing power attributed to inactive contributors over time.
It would be interesting to see separately CPU and GPU performance.
In 2000 fastest supercomputer had Rpeak of only 12,288 GFlop/s. Now probably your desktop has similar speed, at least nominal. If the trend will continue, while computing power may still grow, it might become quite irrelevant.
The BOINC client hasn't evolved in the last 10 years other than adding technical functionalities. It looks dated and very much tailored towards techies that can play around e.g. with cc_config.xml files to optimise the output. When BOINC started, people were used to work in a command line environment but today everything needs to be p&p.
In a similar way this applies to the Gridcoin wallet as well. For a broader audience, it is not that obvious to get the combination BOINC & Gridcoin up and running. Obviously people can use the pools, and I truly think they are useful for starters, but that means adding another element to the installation process.
I personally believe we need to remove the complexity of the installation in both clients so that people can start crunching with one push on the button. We need to think about how we can attract people for whom the app installation on an Iphone/Android is the standard.
Yes, now it's an era of plug & play.
As for pools, with current Gridcoin protocol they are necessary, as with 10k + solo crunchers some would have to wait for payouts for years. There is only ~ 960 blocks to stake a day... But this should change with manual reward claims.
BOINC is trying to launch simplified version called Science United (I wonder - is it compatible with Gridcoin?) but forum is quite empty and it seems not ready yet. I also don't feel like it will be a game changer.
Hi @hotbit!
Your post was upvoted by @steem-ua, new Steem dApp, using UserAuthority for algorithmic post curation!
Your UA account score is currently 1.922 which ranks you at #26972 across all Steem accounts.
Your rank has improved 75 places in the last three days (old rank 27047).
In our last Algorithmic Curation Round, consisting of 206 contributions, your post is ranked at #155.
Evaluation of your UA score:
Feel free to join our @steem-ua Discord server