I haven't been able to come up with tangible ideas which could be translate into code in order to get average sync times due all the variables involved... I could make a super gross average given average connection speed, average computer used in nodes... but that's just not realistic.
I think there's no other way to do this but to make a database and manually record every different device used to sync, the constant would be speed connection and then you can extrapolate into your speed own speed connection.
Like Ronnie Coleman say, everyone wants to be a bodybuilder but no one wants to lift heavy ass weights.
Topic: Benchmark mode (Read 191 times)
Once you have an up-to-date blockchain, then synchronisation and validation are the important factors. Running a slow node can have an impact on the Bitcoin infrastructure, and I suspect that this will lead to some peers dropping you. Enabling both inbound and outbound connectivity will make a difference. The way your ISP connects to the Internet backbone will become more important in the future, especially if we enter into an era of attempted government control. Given a reasonably modern machine with adequate ram and disk space, I don't think hardware makes a great deal of difference, but the use of a large SSD seems to speed things up. Firewalls and ISP throttles and bandwidth caps seem to be significant, and these are things I am investigating at the moment.
I'd be interested to hear of any tinkering that can be done to the core setup to improve peer interaction.
I'd be interested to hear of any tinkering that can be done to the core setup to improve peer interaction.
I was thinking about developing a benchmark mode for Bitcoin Core which would extrapolate synch time in order to make a database with different devices, without having to manually download and validate the entire blockchain.
Would it be possible to develop this with extrapolation techniques? You see benchmark modes in various software most commonly in videogames, by taking averages of FPS during different workloads. Because of the nature of blockchain the difficulty is that speed decreases and increases depending on where you are at (the first years go pretty fast, then we have the spam years which lower the average and so on). If you have any ideas in order to extrapolate averages with given data then post it here.
PS: Maybe move this to development subforum.
Would it be possible to develop this with extrapolation techniques? You see benchmark modes in various software most commonly in videogames, by taking averages of FPS during different workloads. Because of the nature of blockchain the difficulty is that speed decreases and increases depending on where you are at (the first years go pretty fast, then we have the spam years which lower the average and so on). If you have any ideas in order to extrapolate averages with given data then post it here.
PS: Maybe move this to development subforum.
Jump to: