random read iops improvements above 400k will not be cost effective https://www.anandtech.com/print/13951/the-samsung-983-zet-znand-ssd-review The Samsung 983 ZET (Z-NAND) SSD Review: How Fast Can Flash Memory Get?
and cheap low power consumer drives are starting to hit that same 3GB/s sequential read rate
the surfminer is a good idea
I hope so
we shall see
if you do put some effort in at this point, also consider desktop gui stuff
yeah
I was thinking of making a desktop UI mode where one tab would run a node, with just dumping the nodes output to a text box (that can mostly be ignored) and then other tabs supporting normal client things.
also throw in a tab for the miner
I'm not sure that is a good idea. I'm not sure it is a bad idea either.
and yes, no need to be pretty yet, that’ll increase the user base already
people find a gui very useful to configure things
and the tiered config panel approach to hide most stuff under ’advanced’ should do for the miner
address pool where is your snow field thread count
memfield or precache are pretty much for users whom can manage anyway
and the users whom for the gui would be trip up on concepts like ’comment line’ or ’path to directory’
I can absolutely see the value, I'm just thinking that it will be a lot of work
i guess i could do a config file generator
online, in javascript
the file chooser dialog of a file upload input field could potentially be abused for this
and can default threads to a multiple of the core count, if there is an api for that
webassembly should have something
Right, can keep it simple and do things like launch an OS editor on a default config file with explanations inline
saves a ton of UI work
and then you can run the node (or miner) in a tab with some stats at the bottom about how it is doing (block high, peer count for node), (hash rates for miner)
yep, and with a bit of forethought in the architecture, everything can be improved in baby steps over time