I wanted to test the new '60 times faster' bitcoin 0.10.0 wallet against Darkcoins update times, I have created a dimensionless ratio of the amount of seconds of actual 'real time' it takes to update each hour behind of the wallet measured in (seconds)/(hour of update time behind). [kind of like a 'Q' factor to measure quality in physics]
And my weekend benchmark results are averaging out nicely
Connection speed 100 megabits/second.
Machine 1 : SSD (500 megabytes/s up+down, SATA 3) 4.4 gigahertz i7 4820k, windows 7 64 bit.
Machine 2 : HDD (3 gigabits, SATA 2), 2.4 gigahertz Intel Duo E4600, Windows 7 64 bit.
Machine 1,SSD: Darkcoin (0.11.1.25) Best update ratio = 1.346 seconds per hour of update
Machine 1,SSD: Bitcoin (0.10.0) = Best update ratio = 3.304 seconds per hour of update
Darkcoin is 2.4 times quicker than bitcoinMachine 2,HDD : Darkcoin (0.11.1.25) = Best update ratio = 3.817 seconds per hour of update
Machine 2,HDD : Bitcoin (0.10.0) = best update ratio = 26.14 seconds per hour of update
Darkcoin is 6.9 times quicker than bitcoin on traditional 3 gig sata hard drives of average cpu spec.Observations:The cpu of the HDD machine maxs out at 100% with bitcoin, indexing goes very slow especially on the first run from cold start and gives spurious update times (Darkcoin is totally stable at 40% CPU)
Darkcoin update times were statistically closer together and more stable especially on the HDD.
Conclusions:a) Fast modern machines and solid state drives work well with the bitcoin faster hop times but older machines do not like it at all, and use too many resources bottle necking the data transfer.
b) Bitcoin blockchain length is too large for older,slower machines (still most of the world, as not everyone's 'old' machine is a 2012 iMac like Gavin Andresens!)
So does Darkcoin now hold the world coin record for update times? The excellent Dev's have nearly broken the
1 second per hour of update barrier Ironically Mr Andresen is campaigning to have a larger blockchain!
“The old code would spend about 600ms CPU time per hop propagating blocks with already-seen transactions, this new code will spend under 10ms.” --Gavin Andresen--