[MikeS3000]

@MikeS3000 Yeah probably 1.55v for that 4000c15 to 4133c15 range. But some stick just wont do it. lotto just gotta try it and see.

Or try 4000c15-16-16 will be easier to run or 4133c15-16-16 if you can clock higher.

temps are crucial when memtesting tighter subs in that 45c plus range more likely to error.

I'll have to man up and push 1.55v then. I can boot 4100 and 4133 but it fails Y-cruncher in seconds every time. Not sure how to stabilize it. I could drop in my old trusty Hynix sticks to see if they can boot 4100 and 4133 and not error in y-cruncher. That would rule out the IMC as the culprit. I'm getting nearly identical Y-cruncher results at CL 15 and 16. Maybe that's just a result of instability at too low of voltage. I've got a 140mm fan on the RAM fulltime that keeps me between 41 and 45 deg when memtesting so should be good with that. My b-die is XMP 3600 1.45v 14-15-15-35 so a higher bin but not the top bin like your 4000 c14.

 

[hawkslayr]

This is what I have been able to do so far, it passes 3 cycles of tm5 anta extreme, 30 mins occt (both sse and avx2) and y-cruncher pi-2.5b. Basically just copied secondaries and tertiaires from people I saw in this thread and set primaries myself after not being able to copy flat 15s.

However I'm wondering if it's normal that I couldn't get tRCD below 16 and if there is any way I could get it to 15 atleast? I'm also wondering if it actually makes a difference to have set tRP to 14 or I will still get the same performance as if it was set to 16 like tRCD?

I tried setting tRFC to 300 but got error 12 in tm5.

If someone has any suggestions on what I could tighten I'd be glad to hear them.

VDIMM 1.55 and SA/VDDQ 1.35

 

[bscool]

This is what I have been able to do so far, it passes 3 cycles of tm5 anta extreme, 30 mins occt (both sse and avx2) and y-cruncher pi-2.5b. Basically just copied secondaries and tertiaires from people I saw in this thread and set primaries myself after not being able to copy flat 15s.

However I'm wondering if it's normal that I couldn't get tRCD below 16 and if there is any way I could get it to 15 atleast? I'm also wondering if it actually makes a difference to have set tRP to 14 or I will still get the same performance as if it was set to 16 like tRCD?

I tried setting tRFC to 300 but got error 12 in tm5.

If someone has any suggestions on what I could tighten I'd be glad to hear them.

VDIMM 1.55 and SA/VDDQ 1.35

I have never tried tRP lower. No idea how it acts. I run it same as tRCD.

Did you try 4133c16-16-16, sometimes 133 dividers are easier to get stable.

Or 4000c15-15-15 might be better. Maybe you already tried. either way no big deal outside of benchmarks.

 

[Ichirou]

I have never tried tRP lower. No idea how it acts. I run it same as tRCD.

Did you try 4133c16-16-16, sometimes 133 dividers are easier to get stable.

Or 4000c15-15-15 might be better. Maybe you already tried. either way no big deal outside of benchmarks.

tRP lower than tRCD was only ever possible on AMD, but maybe Z690 now allows it?

 

[FeelsBadMan]

I'm on the Strix D4 (bios 0901) with a 12700K and 2x8 SR Patriot Viper 4400C19 sticks and I can be completely stable passing anta777 Extreme 3 cycles multiple times at 4000MT/s with the settings below with just 1.25 Vccsa and 1.35 VDDQ (maybe even lower haven't tried for both), 1.54 VDIMM. I'm trying to do 4100, but even trying 1.43 Vccsa, 1.45 VDDQ, 1.5 VDIMM at 4100-18-18-18-36 CR2 500trfc auto secondaries-tertiaries I get errors. Is there anything I'm missing or something else I could try? Also I've noticed my latency on aida with the stable settings isn't as low compared to other people with same primaries and same or even looser secondaries-tertiaries.

 

[Gsen999]

I'm on the Strix D4 (bios 0901) with a 12700K and 2x8 SR Patriot Viper 4400C19 sticks and I can be completely stable passing anta777 Extreme 3 cycles multiple times at 4000MT/s with the settings below with just 1.25 Vccsa and 1.35 VDDQ (maybe even lower haven't tried for both), 1.54 VDIMM. I'm trying to do 4100, but even trying 1.43 Vccsa, 1.45 VDDQ, 1.5 VDIMM at 4100-18-18-18-36 CR2 500trfc auto secondaries-tertiaries I get errors. Is there anything I'm missing or something else I could try? Also I've noticed my latency on aida with the stable settings isn't as low compared to other people with same primaries and same or even looser secondaries-tertiaries.

View attachment 2542865

12700k imc is worse than that of the 12900k. Most likely the imc

 

[drkrffxx]

I have managed to set my primary timings to 14-14-14-20 at 3733 T1 G1, does it make sense to continue pursuing tRAS lower than 20? Does it even make a difference being at 20 instead of the usual 2x CL?

 

[Ichirou]

I have managed to set my primary timings to 14-14-14-20 at 3733 T1 G1, does it make sense to continue pursuing tRAS lower than 20? Does it even make a difference being at 20 instead of the usual 2x CL?

tRAS can be as low as it can go as long as it is stable and performance isn't regressive

 

[drkrffxx]

tRAS can be as low as it can go as long as it is stable and performance isn't regressive

Good, thank you. 14-14-14-14 here I go haha

EDIT: I was kidding, but it seems off to a good start.

I cannot get 3800+ without errors at any timing, so it's probably my IMC being ****. But I feel like I have a decent kit on my hands going to waste by the IMC or BIOS, not sure.

 

[X909]

On 12700k I guess it's the IMC. I had two and their IMC where way worse than that of both 12900k I tested later.

But with 4 DIMMs and the 12900k I struggle in the same clock range than you

 

[donald24]

There is some saying about never to trust a POST memspeed before even benchmarking it, if not following req's are met:

MCH Full check enabled
MRC Fast boot disabled

Before even posting my experiences with it, what do you guys think about this statement?

 

[bscool]

There is some saying about never to trust a POST memspeed before even benchmarking it, if not following req's are met:

MCH Full check enabled
MRC Fast boot disabled

Before even posting my experiences with it, what do you guys think about this statement?

I say do what works for you. Test and verify for yourself what someone says or tells you. Try it, if it works for you do it.

I dont follow that "rule" and I dont have issues on z690. On older gens like z490 I did that, didnt on z590 because training takes much longer and never had issues.

Some things that you/people did on older gens you dont have to continue to do on new gens.

 

[FeelsBadMan]

12700k imc is worse than that of the 12900k. Most likely the imc

oh ok, thought IMC is the same on both cpus and it's just silicon lottery, if I had known that a few months ago and how much of a better bin 12900K P-cores are, maybe would have gone with that one. Now waiting for the 13900K I guess

 

[bscool]

Edit...........nothing to see here

 

[truehighroller1]

I will be keeping her at 4133 I guess. I managed to drop my llc to 4 and get everything stable which I couldn't do before. That dropped my temps a bit.

 

[Balaned]

I can't locate a reason for this but my TWR settings won't stick even though changes are accurately reported in the BIOS. For example, I'll set TWR to 12 but in WIndows it will show as 24 via Timing Configurator. I'll cold reboot into the BIOS and the setting still shows as 12 there but 24 in Windows. I double checked the value in MemTweakIt, also 24. So which is accurate?

Asus TUF Z690 D4 BIOS 0807, 12900K, 2x16GB DR B-Die at 4133/15/16/16/32 1.54v, 1.35v SA & VDDQ.

 

[bscool]

I can't locate a reason for this but my TWR settings won't stick even though changes are accurately reported in the BIOS. For example, I'll set TWR to 12 but in WIndows it will show as 24 via Timing Configurator. I'll cold reboot into the BIOS and the setting still shows as 12 there but 24 in Windows. I double checked the value in MemTweakIt, also 24. So which is accurate?

Asus TUF Z690 D4 BIOS 0807, 12900K, 2x16GB DR B-Die at 4133/15/16/16/32 1.54v, 1.35v SA & VDDQ.

Post your memtweakit first page timings.

tWR is also controlled by tWRPE and tWRPEN are you setting those in the bios?

I believe when I tested it only tWRPEN had an effect on tWR. I dont touch tWRPE or tWRPEN on z690 only tWR.

Edit I am on z690 Asus MB no idea if MSI or gigybyte etc are the same or differ on this.

I know on older gens this rule above is slightly different as tWRPE will effect tWR also. From what I remember been a while since I messed with z490/z590 hard to keep it all straight

 

[truehighroller1]

Post your memtweakit first page timings.

tWR is also controlled by tWRPE and tWRPEN are you setting those in the bios?

I believe when I tested it only tWRPEN had an effect on tWR. I dont touch tWRPE or tWRPEN on z690 only tWR.

TWRPDEN controls it.

 

[bscool]

TWRPDEN controls it.

So do you use TWRPEN or just set tWR? The end result is the same I believe, but could be wrong.


Edit on z690 I edited my post above as on older gens slightly different from what I remember. TWRPE controlled it or could.

 

[truehighroller1]

So do you use TWRPE or just set tWR? The end result is the same I believe, but could be wrong.


Edit on z690 I edit my post above as on older gens slightly different from what I remember.

I noticed that you just set yours. I just use the twrpden to set it and then set the twrpe to the same value and it seems to make a difference speed wise for me.