Can I boot 4133+ ? DDR4 OC 3200 --> 4000+ with 10900k

[gabeomatic]

Would love tips on if any timings can be adjusted or look glaringly off or even quick or dirty timings to try and boot 4133-4200 with a 4 dimm SR setup?

This kit is 4x8gb SR 3200 Dominator Platinum (white) @ 16-18-18-36 (two kits of 2x8gb) @ 1.35v XMP running at 4000 16-20-20-40 at 1.44v IO/SA 1.3/1.35v

tRFC 583-585 @ 4000, 555-560 @ 3800 (I know CJR can do ~260-270ns but Micron rev E cannot)

trefi 25-35k for now

TRDDS 4, TRDDL 6, Tfaw 16, RTP 6, TWR 13 (with 12 entered in bios, should I try 10?)

CPU is currently 1.425v bios 52/49 0avx HT on with only LLC4 for now (runs quite less under load)

I ran some rough tm5 anta777 extreme profile and fortnite/ warzone for some quick testing. Do my primaries look too tight and may be giving me false stability and actually diminishing performance? No hwinfo errors etc. I know CJR is a bit more temp sensitive then rev E so should I keep going to 1.45v+? Anything look quite off from initial glance? I am getting similar latency as my 2x16 CJR Dual Rank kit got 44-45ns at 3800 with tuned subtimings so just want more input. I know people even got 4200+ plus I haven't been able to boot at 4133 etc just yet, even at C18.

My 10900k is SP78 on the newest bios with an h150i elite (166-174 cooler rating) - tested up to 53/50 HT on 0avx, and 54/50 HT off but it is stock while ram testing. Will post some aida / cpu-z/ asrock timing config screenshots when I get home from work, but would love opinions on where to go from here. Thanks reddit fam!

Best I've gotten so far (aside from 59500s read speeds)

After fiddling with a bunch of different setups this is where I'm at. Wasn't sure if anything looked glaringly off, got a bit of help from an OCN member but not sure if its even worth trying for 4133+ Cause I've got the itch haha. 1.43v IO/SA 1.26/1.3 - Pretty stoked where I'm at so far from a stock 3200 16-18-18-36 4 dimm kit.

Cpu is 52/49 1.425v bios LLC4, runs way less under load SP78 Chip h150i elite capellix, tested up to 53/50 Ht on 54/50 HT off. Anywhere interesting to go from here?
After days and days of messing around, pretty happy where I'm at for non bdie!

 

[Deleted member 185158]

This is certainly well within average of this system.

Yes B-Die would be the way to go honestly.
16gb A0 or A1 PCB, for benching Cas 12 at 4000mhz.
That's sompetitive benchmarking stuff right there.

You're at the limit of the cooling for 14nm.

How much itch do you have?
Does it get as bad as mine does??

 

[gabeomatic]

So much that I’ve managed to get this far at least. Yeah those two dim setups can scream at over 4400 but this setup Ali already had got me going. What do you think? Anything look glaringly off?

And wow, nice 8700k! How is it cooled?

 

[Deleted member 185158]

So much that I’ve managed to get this far at least. Yeah those two dim setups can scream at over 4400 but this setup Ali already had got me going. What do you think? Anything look glaringly off? View attachment 287694

And wow, nice 8700k! How is it cooled?

That was DryIce cooled.

Your overclock looks exactly how it should. I see nothing to improve that stands out honestly.

 

[gabeomatic]

That was DryIce cooled.

Your overclock looks exactly how it should. I see nothing to improve that stands out honestly.

Cpu is 52/49 1.425v bios LLC4, runs way less under load SP78 Chip h150i elite capellix, tested up to 53/50 Ht on 54/50 HT off. Anywhere interesting to go from here?

Updated benches (doing further testing with PPD 0 tXP 4-6) Does PPD 0 not like keeping c states on?

Max tREFI 1.44vdimm IO/SA 1.3/1.35

Maybe will try lowering tCWL to 14, tCKE to 7,possibly tRDD_L to 4, tWTR L/S to 7/3, 6/2 respectively but would need some more input probably. Can try the tRAS down a notch or two again but these all might be pushing it.

Also got advice to try the 4 timings under tCKE timing to 7/4/6/6 or even 6/4/6/6 and the next 4 at 11/11/11/11 - maybe still get tRDWR_dr down to 10, maybe 8? tRDWR_dd might do 10?

Can tWRRD_dg go to 25? Man I opened a can of worms as I'm still fairly a noob with ram, seems like over time I keep pushing vdimm and IO/SA with all these tweaks and end up taking things a tick further. Still shocked where I've gotten though and glad I fell down the rabbit hole! Thanks all

 

[Deleted member 185158]

Cpu is 52/49 1.425v bios LLC4, runs way less under load SP78 Chip h150i elite capellix, tested up to 53/50 Ht on 54/50 HT off. Anywhere interesting to go from here?

Updated benches (doing further testing with PPD 0 tXP 4-6) Does PPD 0 not like keeping c states on?

Max tREFI 1.44vdimm IO/SA 1.3/1.35

Maybe will try lowering tCWL to 14, tCKE to 7,possibly tRDD_L to 4, tWTR L/S to 7/3, 6/2 respectively but would need some more input probably. Can try the tRAS down a notch or two again but these all might be pushing it.

Also got advice to try the 4 timings under tCKE timing to 7/4/6/6 or even 6/4/6/6 and the next 4 at 11/11/11/11 - maybe still get tRDWR_dr down to 10, maybe 8? tRDWR_dd might do 10?

Can tWRRD_dg go to 25? Man I opened a can of worms as I'm still fairly a noob with ram, seems like over time I keep pushing vdimm and IO/SA with all these tweaks and end up taking things a tick further. Still shocked where I've gotten though and glad I fell down the rabbit hole! Thanks all

My Maximus X hero was great for pumping 1.45v system agent if left auto. Lil fyi.

But you can spend a LOT of time tweaking this. As you've noticed already I'm sure. In short, no one can give you specific numbFYI.

Cstates I'm accustomed to disabling. I like a max performance system, that means no core parking either.

Also for memory efficiency, not just latency, use other benchmarks. PiFast, Pimod, pyprime and Ycruncher are all memory sensitive.