nothing so "complex" abt it
end of the day it is lowest cost per hash + best efficiency be it an ori or modded bios. no matter how it is tuned or adjusted, 480 will consume more power than the 470's.
simple, now show/tell me:
470 4gb vs 480 4gb perf vs watts so on and so forth according to list you listed out.
50-100W more is a RUBBISH comment & spreading FUD !
obviously i do have actual results mate as i am also sharing it with citronick live, i will also test SMOS shortly using the RX 5xx series. all i can say atm, the driver is pretty unstable. there are some new features too.
depends how "best" is categorized for a miner in terms of a kiddo making rubbish comments (2-3 mths ROI?) or in terms of a commercial/investment operations.
tbh, best is, least $ spent/invested for the best returns in shortest period of time, well that's for me for a start.
numbers NEVER lie
of course but many of us don't do a 100 gpu plus setup.
for me in the garage 3 or 4 rigs filled with 1080 ti's will be better then 5 rigs filled with rx 480 8gb.
For a guy with 100 rx 470 4gb using custom flashed cards they make sense due to the extra power save.
I have seen claims of a 6 card zec with rx 470 8gb doing 1500 h using 600 watts at the k-watt meter. that is a nice rig and x it by 50 since the custom rom cost money it makes more sense for that miner.
It is why smaller miners are disadvantaged since custom flashed cards are not cost effective for one 5 card. So people decide to diy a flash with good and bad results.
I have a thread with 5 links to I fucked my card up flashing it. If I spent a day I would find 100 links here on this site.
When you do a calculation or statement about rx 470 4gb you naturally don't have this happen to your gear and you don't factor it in.
I am not attacking you statement I am saying it is based on your setups'
So what is your dead card or damaged card due to flashing loss under 2% correct ?
Do your figures with a 10% fail rate for flashed cards. This alters the numbers.
100 flashed 2 bad is 98 good
100 flashed 10 bad is 90 good.
that alters the statement. you made.
now lower it to a guy 1 rig six cards 1 dead due to a bad flash 5 good he has a 16 ⅔ % fail rate
and the next guy 1 rig six cards 0 dead due to all good flash he has a 0 % flash rate
first guy hates flashing
second guy loves flashing
first guy was better off with all rx 480 8gb
second guy is like you happy with rx 470 4gb
over all fail rate is 8% but for one guy it was 16 % the other 8%