FalloutBoy wrote:
I have a rampage V extreme Edition 10 with Corsair Memory CPU-Z identifies the Die type as Samsung, now I know to run it at it's rated speed of 3000Mhz I have to run it at 1.35V and NP it works .... but ....

I have been playing with MemTweakit and discovered that I can get the timings down from 15,17,17,35,1T @ 3000Mhz to 14,14,14,21,1T @3000Mhz now my testing was very basic I ran up a game looked for artifacts etc and saw the fact that it would run - so this is not a 24/7 usage scenario yet.

But I would like to know is if 14,14,14,21,1T @ 3000 is a pipe dream or can this actually have some basis in reality?

Note pushing CAS Latency down to 13 does cause it to crash at 1.35V as soon as the apply button is hit.

hey there nice over clock

i have g.skill trident z rgb 2x8 gb of 3000 mhz cl-14-14-14-34 @1.35v xmp
@ 4000 mhz cl-17-19-19-39 1.4v vccio and vcssa both @ 1.21v stabe in hci /prime / real bench / mem test 86

i would stress test your ram extensivly after a manual oc games and assuming its stable is not a good idea .

after any ram change you "should"

run mem test 86 for a quik basic test to deem stable to move to next stress test to deem "stable"
then
run hci mem test for et least 1000 % coverage try for 1500 - 2000 . - no error pop ups
to run hci mem test right you want to test about 85 -90 % of your ram .
and divivde the ram ammount by how many threads your cpu has ( this includes ht cores as well )

Example 1 instance of hci " per cpu thread " x how many threads you have
i have 16 gb ram windows reports i have abot 14gb free so
14,000 mb ( ish ) divded by 12 threads ( 8700k 6 core / 12 thread ) = 1166 mb per hci window x 12 hci windows
thats how u use it let it run for 1000 %

also i run prime 95 26.6 blend for a good few hrs after hci
and asus real bench to test my cpu oc , ram oc , gpu oc all stressed at the same time .

report back you could try higher ram mhz like 3466 / 3600 ?