|Topic Review (Newest First)|
|02-07-2019 08:18 PM|
|Fraizer||i dont understood all but look is ok|
|02-07-2019 05:05 PM|
I didn't really mean that as a statement I posed it as a question to you.
I already know the answer I just want to see what you say.
Once again its coming from a place of love brotha.
Also the language barrier is easier than you would think to overcome but the chain of events are critical in what you may be changing.
|02-07-2019 05:03 PM|
you help me at the beging and even you just come now i am realy happy to read you and i know you are here to help me thank you a lot for that
about your question if i remember well i did that but dosent boot at vccio and vsa 1.2500v, only when i put at 1.30 both but when i run an GSAT it failed and my windows freez at 4400mhz it pass at 1.2500v but many error on gsat.
i tried too like i said at 4300 19-19-42 vccio & sa 1.2500v... (i was stupid to not wrot all my tast on a paper...) but i had the issue to restart the windows who take very long time like trying memory speed at boot. but here when i make an bench at this speed and timing i realise it was less better of the 4133 XMP + custom timing.
I have a problem with aida memory bench i dont know if its normal but the bench result are different up to 300 or 400points sometime and the reponse timing in nano seconds too... then not easy to know for every timing or freq if i get real gain.
this is normal ? or i have an problem ?
|02-07-2019 04:29 PM|
If you load the Raw Mhz profile and boot at 4500 and then try to load the XMP profile what is it going to do?
Im only trying to help you or anyone else that wants to understand im not trying to be a dick and Im not always right either.
|02-07-2019 04:24 PM|
i tried with higer timmings like at 4300 19-19... but i get problems and this was with and without xmp (i tried at the end xmp). i spend so many days to make many stability test and trying many stupid timings.. probably all those test was usless... lol
yes you are right about what i said in 1st post but i realy dont know what i did i had an mem ok switch was enable and i understand it was not good and i think like small vccio vsa... at this moment was focusing the 4600mhz and i dont understand at this moment why it dosent work...
i think is better to focus of what i post actualy i know a little more in oc and i am just trying actualy to focus of what i can have the best with this 4 dimm i make run like an quad kit when it is an double kit and no QVL list.
i hope i understand your question and answered right
|02-07-2019 04:13 PM|
Those timings are pretty tight at the frequency your at now let alone increasing it without changing any timings. Thats why your unstable because your relying on profiles that are not made for these high of frequencies except the one that I keep repeating.
In your first post you said you tried XMP1 and 2 and didnt boot, so what was it that got you to boot at 4500 was it perhaps the loose timings you got from the raw frequency memory profile? In the way you responded to me thats what it seemed like.
|02-07-2019 03:52 PM|
here is the chips information of the stick 1 (they are 4 because i have 4 dimm sticks)
i dont know if it is usefull for oc... i get that with Thaiphoon Burner 184.108.40.206 free version.
i can post the 3 others sticks
|02-07-2019 03:34 PM|
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:
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)
|02-07-2019 11:34 AM|
Nice job exposing me there.
|02-07-2019 11:30 AM|
|This thread has more than 10 replies. Click here to review the whole thread.|