Overclock.net - An Overclocking Community - Reply to Topic
Thread: *Official* Intel DDR4 24/7 Memory Stability Thread Reply to Thread
Title:
Message:

Register Now

In order to be able to post messages on the Overclock.net - An Overclocking Community forums, you must first register.
Please enter your desired user name, your email address and other required details in the form below.
User Name:
If you do not want to register, fill this field only and the name will be used as user name for your post.
Password
Please enter a password for your user account. Note that passwords are case-sensitive.
Password:
Confirm Password:
Email Address
Please enter a valid email address for yourself.
Email Address:

Log-in


  Additional Options
Miscellaneous Options

  Topic Review (Newest First)
07-12-2020 03:25 PM
Gen. Welcome friends! 4600 17-17-1T 1.52V with 8700K 5200/5000.
07-12-2020 12:56 PM
nick name
Quote: Originally Posted by XGS-Duplicity View Post
Ahhh interesting. For 4600, I think I had to boot 4533 first or maybe start with 4400 then goto 4500 then 4533 or something, then fastboot from there with 4600 strap to sneak in a benchmark. Did it a couple times, not in awhile though. In terms of voltages, all i can say is that its really going to depend on the dimms. I probably don't have much to offer in terms of advice, i don't know much about amd's architecture, i just know they got chicklets.
That's one of things that I wonder about too. Whether my kit would need the same voltage on an Intel platform so I was hoping to find some folks here that have played around at 4400 and 4600.
07-12-2020 12:50 PM
XGS-Duplicity
Quote: Originally Posted by nick name View Post
Yeah, I've found that I can't get to 4600 reliably without booting 4400 first. Then 4600 works like a charm.

Playing with high speed is actually what brought me here to you Intel guys. Nobody really plays that high on their AMD systems so I don't have anyone to ask about DRAM voltages at those speeds.
Ahhh interesting. For 4600, I think I had to boot 4533 first or maybe start with 4400 then goto 4500 then 4533 or something, then fastboot from there with 4600 strap to sneak in a benchmark. Did it a couple times, not in awhile though. In terms of voltages, all i can say is that its really going to depend on the dimms. I probably don't have much to offer in terms of advice, i don't know much about amd's architecture, i just know they got chicklets.
07-12-2020 12:43 PM
nick name
Quote: Originally Posted by XGS-Duplicity View Post
ohhhh. Well, now that you mention that. That's what I was doing lately this week. 4200 memory strap does not train at all with this kit on this board(even if i tried training it after training 4133), in order for me to get there, I have to train 4133 memory strap + 101.5 busclock and then go back into bios to remove the busclock, select 4200 memory strap and then enable memory fast boot. If i just leave the busclock, its only memtest stable until I shut the pc off. I think I tried jumping from 4200(post training/fastboot) to 4300 strap with just voltage increases last night or this morning using memory fastboot. I can't figure out for the life of me how to train 4266/4300/4333/4400 with this kit so i'm just trying all sorts of stuff.
Yeah, I've found that I can't get to 4600 reliably without booting 4400 first. Then 4600 works like a charm.

Playing with high speed is actually what brought me here to you Intel guys. Nobody really plays that high on their AMD systems so I don't have anyone to ask about DRAM voltages at those speeds.
07-12-2020 12:37 PM
XGS-Duplicity
Quote: Originally Posted by nick name View Post
Honestly, I never found anything that explained it to my satisfaction. In my encounters -- folks were talking about the usefulness of certain voltages on the Ryzen platform in regards to "moving around memory holes".

Other instances that seem to correlate are when folks need to use one strap to get to another strap when simply trying that second strap first won't work.

But I properly have an improper understanding of it and everything I am saying is utter cr*p. So I'd wait for a second opinion.
ohhhh. Well, now that you mention that. That's what I was doing lately this week. 4200 memory strap does not train at all with this kit on this board(even if i tried training it after training 4133), in order for me to get there, I have to train 4133 memory strap + 101.5 busclock and then go back into bios to remove the busclock, select 4200 memory strap and then enable memory fast boot. If i just leave the busclock, its only memtest stable until I shut the pc off. I think I tried jumping from 4200(post training/fastboot) to 4300 strap with just voltage increases last night or this morning using memory fastboot. I can't figure out for the life of me how to train 4266/4300/4333/4400 with this kit so i'm just trying all sorts of stuff. What voltages did they increase to get around memory holes? I've already tried increasing dmi voltage all the way up to 1.35v and i've also tried increasing vccpl OC voltage up to 1.35v(xmp usually sets at 1.31v according to gigabyte eztune software). Also tried all the way up to 1.4v sa/io. Tried tons of vdimm too, didn't seem to make a difference.
07-12-2020 12:32 PM
nick name
Quote: Originally Posted by XGS-Duplicity View Post
hmmmmmm this term is new to me. I don't quite follow. I googled ram memory holes and pci hole/memory remapping comes up. Is this what you are referring to? I don't understand any of it, my bios doesn't have memory remap feature or anything like that. To add, all my ram is usable to my knowledge, windows shows 32gb.
Honestly, I never found anything that explained it to my satisfaction. In my encounters -- folks were talking about the usefulness of certain voltages on the Ryzen platform in regards to "moving around memory holes".

Other instances that seem to correlate are when folks need to use one strap to get to another strap when simply trying that second strap first won't work.

But I properly have an improper understanding of it and everything I am saying is utter cr*p. So I'd wait for a second opinion.
07-12-2020 12:29 PM
XGS-Duplicity
Quote: Originally Posted by Falkentyne View Post
Yeah. Changing RAM timings back and forth too much and voltages around sometimes requires a 'deep' cmos clear to reset everything.
A fast cmos clear resets the UEFI only. But some things are still left unchanged in NVRAM and god knows what other low level stuff.

I'm just glad *someone* trusts my hard work around here still.
Ahhhhhhhh ok that makes sense. Here i thought my imc degraded LOL. thankfully you exist. Gave you a +1 for the helpful tip. Yeah I still follow most instructions you give me. I just really don't want to lower my cache lol.
07-12-2020 12:21 PM
Falkentyne
Quote: Originally Posted by XGS-Duplicity View Post
Ok so i got c14 3900 back up and running. Your instructions worked, thank you I reflashed F9 after trying c14-3900 on f11e(failed to train), trained c17-4200 on f9, then reset cmos again and trained c14-3900.
What should I chalk this up to? Just sometimes it gets finnicky after training many different profiles or something? I guess i must of tried like 30-50 times to train c14-3900, c15-4000 and c16-4200 in the last 24-36 hours, either fail to train on c14 3900 or wrong/slower rtls on 4000+ for the most part. c14 3900 is tough to reach on 4 dimms on this board, it trains about once out of every 4 or 5 attempts at training. I guess for good measure every once in a while I should just do a full powerdown and hold cmos in for 10 seconds instead of a quick press i guess?

I haven't used a hex editor or anything before, not familiar with that stuff. I'll have to get a cheap second usb stick for modded bioses.
Yeah. Changing RAM timings back and forth too much and voltages around sometimes requires a 'deep' cmos clear to reset everything.
A fast cmos clear resets the UEFI only. But some things are still left unchanged in NVRAM and god knows what other low level stuff.

I'm just glad *someone* trusts my hard work around here still.
07-12-2020 12:08 PM
XGS-Duplicity
Quote: Originally Posted by nick name View Post
I'm not very familiar with memory holes, but from little I do know it sounds like what you were moving around.
hmmmmmm this term is new to me. I don't quite follow. I googled ram memory holes and pci hole/memory remapping comes up. Is this what you are referring to? I don't understand any of it, my bios doesn't have memory remap feature or anything like that. To add, all my ram is usable to my knowledge, windows shows 32gb.
07-12-2020 12:06 PM
nick name
Quote: Originally Posted by XGS-Duplicity View Post
Ok so i got c14 3900 back up and running. Your instructions worked, thank you I reflashed F9 after trying c14-3900 on f11e(failed to train), trained c17-4200 on f9, then reset cmos again and trained c14-3900.
What should I chalk this up to? Just sometimes it gets finnicky after training many different profiles or something? I guess i must of tried like 30-50 times to train c14-3900, c15-4000 and c16-4200 in the last 24-36 hours, either fail to train on c14 3900 or wrong/slower rtls on 4000+ for the most part. c14 3900 is tough to reach on 4 dimms on this board, it trains about once out of every 4 or 5 attempts at training. I guess for good measure every once in a while I should just do a full powerdown and hold cmos in for 10 seconds instead of a quick press i guess?

I haven't used a hex editor or anything before, not familiar with that stuff. I'll have to get a cheap second usb stick for modded bioses.
I'm not very familiar with memory holes, but from little I do know it sounds like what you were moving around.
This thread has more than 10 replies. Click here to review the whole thread.

Posting Rules  
You may post new threads
You may post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off