i am sorry to answer only now but i get no any more notification since my last answer... escuse me for that.
yes i succed to boot at 4400 and at 4500mhz (loading if i rember the xmp 1 profile. the 4400mh if i remeber at vccio & sa at 1.25 and 4500 at 1.30v both. but yes they was not stable but probably because trying with tight timing.
2 days ago i tried if i rember well at 4300 and make an test of stability with timers 19-19-42 (not sure for the 42 maybe ws on auto...) and i dont had error on gsat for 2 hours of stability test but each time when i restart widnows is take very long time to boot on bios and sometime i need to press F1 because oc error.
i make alot of test and lost days on that with my weak knowledgs and unfortunatly not realy help "for my level" since the help i get before. of course i understand is realy not easy for people to help on oc memory especialy in this case and about those 4600 who look its a way to make joke.
then after that i back to the 4133mhz with timing 16-18 and some other timing i get help with Jpmboy.
i make you an screenshot of my last oc bench data who is stable for 3 hours under GSAT. it will be kind to tell me what you think about and what values i can tight to optimise the timings.
Jpmboy took me to this step and respond after that. i am thankful he took me to this part but i wish to finish this optimisation to end this oc and give you the last bench with those 4600mhz.
after few days ago i succes to have stability at 4133mhz at 16-18. dimm voltage at 1.45v vccio and vsa at 1.2500v (we dont test under this voltage but propably after the oc will reduce them)...
his last message was to that:
Originally Posted by Jpmboy
okay in bios - to start let's change
tWR to 18
tRFC to 425 (you can try 374)
tRTP to 8
tFAW to 36 (we'll lower this once we see it RDD_s can be lower... later)
try these in sequence, save a asrock TC snip if each will boot and post back (i need to see th eeffect on rtl/iol). Feel free to test stability "at your perference"
run an AID64memory test to see if there is benefit. Honestly, for 24/7 4133c16 is pretty damn good!
i suceed to do it (like you can see on the screenshot) with that tWR to 18 + tRFC to 374 + tRTP to 8 + tFAW to 36) but i get an strange thing when in bios i put tWR to 18 it show me on the soft ASRock Timing Configurator 19... then i put in bios tWR to 17 and on the soft ASRock Timing Configurator 19 it show now tWR to 18. i dont know if it is an bug of the bios (i updated since that but same) or an bug from ASRock Timing Configurator... or its somthing normal when we put an wrong timing...
But i dont get any stability problem aft 3 hours of GSAT with this value line stressapptest -W -M 28672 -s 7200 --pause_delay 7200
Falkentyne said i think an interesting information (but it was before i start the oc with Jpmboy) if i understand and calculate well i should have 5.
DDR4 main timings is CAS, tRCD/TRP (usually set to the same value) and tRAS.
According to Raja @ Asus, to improve stability,
set tRAS equal to CAS + tRCD + tRTP (minimum value for tRAS). So in other words, tRTP should be equal to tRAS - TRCD - Cas.
Then set tWR = 2x (double) tRTP.
tRTP should never be less than 5.
i never fault it will be something so complicated to help someone to oc memory, i learn that with that situation and realy all of you i thank you alot for the time you sepnd on my problem. i cross the finger with your help i will finish this OC
please rember my english is not so advanced
the 1st screenshot i join on this post is an XMP 1 profile i load + the custom oc without xmp of Jpmboy. i see gain in write and if agree i prefer to keep this xmp profile.
the second is the last one i make Jpmboy without xmp he make me load an 4x8gb new samsung memory single side profile (my memory if i am not wrong they are double side, need to be confirmed)