[dinos22]

there is nothing magic about my superpi except hard work

i do spend a lot of time getting to know platforms and how RAM overclocks on it and how it all affects superpi though but the tweaks are almost universal with some slight system adjustments needed from time to time

tweaking pi has been discussed to death on XS and other forums and it's basically all been laid out many times. Trouble is that these days tweaks do very little to improve your efficiency due to CPUs having large cache internally. This is what for example wazza helped with a lot with older chips. i remember in good old days some chips could have as much as a few minutes imprvement in time with a good wazza. these days we are talking about 1-3 seconds MAX.

what i suggest you do is actually do some searching on XS and sit in front of your system and try things > GIVE THEM A SERIOUS TRY BEFORE YOU GIVE UP. this is the key. i can memtest for days sometimes until my face turns blue almost just tinkerin with different things

dont assume things > test them yourself and find out >> i'll give you a hint > increasing NB doesnt mean faster pi times .....it can make it slower when higher NB if the rest of your subtimings are not balanced. you have to find the right mix so assuming higher NB = faster is wrong if you dont have your other timings in order.

test your tweaked OSes as well. get to know nlite and how it all works and what makes a difference

if you put the work in you will get the best results AND whatever new platform comes around you should be able to adapt much quicker than someone else would who relies on others to give them tips

this is the best advice i think i can give for someone that wants to get more serious about their superpi. Nothing wrong with comparing notes with others but you have to get to that stage where you put a lot of work into it and then tweak more out of the system by comparing notes. Thats why i chat with some of my benching friends to work out kincs in my benching no matter what benchmark.....