Overclock.net banner
Cancel
Create thread New Forums

Overclocking "Raptor Lake refresh" - 14900k,14700k, 14600k etc results, bins and discussion

32 reading
Jump to Latest
841 - 860 of 23,546 Posts
#843 ·
Falkentyne said:
I forgot. I "think" it was 1402. May have been 1303. I was too busy dealing with an issue I had trouble troubleshooting for awhile. I didn't use that CPU in that board for very long. And I was dealing with a Stockfish was freezing the chessbase GUI client issue caused by updating the BIOS to 1402 (or 1303???) and then swapping into the 14900K (which I couldn't fix)--i wasn't sure if the CPU was a problem or the BIOS as this ONLY happened after updating the BIOS (it never happened before). The problem is, i updated the BIOS and then swapped to the new CPU without testing the BIOS in windows. And the Apex Encore wasn't showing this problem. And this only happened with 32 threads loaded, if I set chessbase to use 31 threads, it was perfect, no massive lag of the chessbase GUI.

Then I swapped back to my 13900k QS and.......SAME PROBLEM....
I THEN Switched to the BACKUP BIOS which was on 1303 (or maybe 0038 again I Forgot)......SAME PROBLEM...

I realized then it MUST have been the BIOS update which triggered something in windows (security?? ME?).
I eventually tracked it down to "real time protection" in virus settings.
When I disabled that, everything was back to normal.

I never had to disable that before flashing from 0038 to (either 1303 or 1402, I forgot).
Click to expand...
Tested on 14700K and 14900K on the Z790 Hero and Apex Encore, couldn't replicate a difference in SP score (y)
 
#845 ·
I take it theres still no method for MSI board users to check for SP?

I noticed the testing that falkentyne mentioned about testing everything at 1.25v, I'll definitely do that when I get one, but Amazon can be a month or two slower than everyone else.

My 13600KF does 5.4 all core at 1.25v which is seriously what I should have left it at, but I got greedy for its 5.7 / 5.6 / 5.6 top 3 cores and now lost the 5.7 and 4300 IMC welp.
 
Save
Reply Quote Rep+ Rep+
#846 ·
bhav said:
Is it too early to tell of the KFs are better this time around or not?

Back with the 13900K, the KF had so many golden chips here. Then again might as well go for the KF for the better price.
Click to expand...
The same reason I went for KF. AFAIK KF binning pool isnt in 14900K’s in case of 14900KS

So if intel is binning 14900’s, KF isnt included in that pool therefore more chances of better bin.

MSI have Force 2 system. The lower score is better SP. Like my garbage 13700KF were 163 force

MSI also have unique Ram stick rating called Memory FORCE. The higher the bar, more of your ram reaches its limit.
Font Display device Electronic device Multimedia Advertising
 
#848 ·
imrevoau said:
Quick run on a "lowly" SP 102 P core 14900K @5.8 P. v min of 1.217 to loop AVX2 Small Extreme OCCT with HT off. My 13900KF could not even do 5.7 at 1.24vmin.
View attachment 2631582
Click to expand...
how about ht on? my 13KS needs 1.2v vmin for 5.6 1hr occt avx 2 extreme variable.
 
Save
Reply Quote Rep+ Rep+
#849 ·
bhav said:
I noticed the testing that falkentyne mentioned about testing everything at 1.25v, I'll definitely do that when I get one, but Amazon can be a month or two slower than everyone else.

My 13600KF does 5.4 all core at 1.25v which is seriously what I should have left it at, but I got greedy for its 5.7 / 5.6 / 5.6 top 3 cores and now lost the 5.7 and 4300 IMC welp.
Click to expand...
You are talking about 1.25v under CB23 load and not BIOS set right? So 1.35v BIOS set at LLC6.

If we are to run our CPUs' at max 1.25v under CB23 load, then many would barely run stock speeds. For my 13900K that would be max 5400 and not 5500. For a 14900K that would mean 5600 and not 5700.
 
Save
Reply Quote Rep+ Rep+
#852 ·
imrevoau said:
I'm on an AIO sadly so I think it would throttle :/ I can try and test tomorrow though. I ain't running an hour though, I don't want to degrade my chip.
Click to expand...
OCCT avx 2 extreme can freeze/restart/bsod mid way at the 15-30min mark, or it can show errors on one or more cores. hence 1hr occt is the minimum requirement to test HEAVY AVX 2 stability imo.
10min stability is not really avx 2 stable from my experience. well if u are under intel electrical guidlines, it is fine to run it for couple of hours assuming the temps are <95c preferably.
And if it throttles, just lower the vcore to an amount your AIO can cool and find the frequency you can stabilize. i assume 5.6 to be possible with your cooling? maybe.
 
Save
Reply Quote Rep+ Rep+
#856 ·
So my DDR 4 overclock on a 14700KF appears to be at an end for now. It's now at

3800
15 15 15 28 2t 280 with some more timing tweaks. Voltages are

SA 1.1v
Vram 1.5v
VDDQ 1.3v

Looking at these settings it looks like there is more headroom but it's not to be. It instantly BSOD no matter what settings just after the post screen when trying anything higher.
Maybe a future BIOS will help or if any of you guys have any suggestions.
 
#857 ·
Cracked 42000 one time on CB23 just barely. Usually getting 41700-41800. Window 10, 57/44 ratios. Score stays same even at higher power.
CarSalesman said:
I'm sure some employees read these forums, but we are not the target audience. Their main market audience is those who just want to assemble their PC and get decent framerates in games.
Click to expand...
That's what I mean. Enthusiasts don't care about the power limits, but it makes them a meme/laughingstock in reviews.
yzonker said:
Might try higher VDDQ. You can definitely run more than 1.4v safely.
Click to expand...
How much is safe?
 
#858 ·
raad11 said:
Cracked 42000 one time on CB23 just barely. Usually getting 41700-41800. Window 10, 57/44 ratios. Score stays same even at higher power.
That's what I mean. Enthusiasts don't care about the power limits, but it makes them a meme/laughingstock in reviews.
How much is safe?
Click to expand...
prob max 1.45v set is "okay".
 
Save
Reply Quote Rep+ Rep+
#859 · (Edited)
I was curious, so I mashed the VID's from the various (19 so far) screenshots posted into Excel to get an average, and a crappy graph. Thought I'd share in case it was of any interest.

EDIT: Added SP scores and some rather hideous automatic colouring (Gold = Best, Red = Worst, Green = Average or better)
EDIT2: Sorted by; 6GHz, 5.8GHz, SP, P- SP, E- SP, 5.6GHz
EDIT3: Now with batch numbers
EDIT4: Don't pay too much attention to the averages, quite often people get multiple chips, but only post the best sample
EDIT5: 2024-01-10 - No longer being actively updated. If you really want added, send me a pm with the screenshots

Rectangle Font Parallel Slope Pattern
 

Attachments

#860 ·
raad11 said:
Cracked 42000 one time on CB23 just barely. Usually getting 41700-41800. Window 10, 57/44 ratios. Score stays same even at higher power.
That's what I mean. Enthusiasts don't care about the power limits, but it makes them a meme/laughingstock in reviews.
How much is safe?
Click to expand...
I ran 1.58v for almost a year without degradation. Maybe limit to 1.5v to be safe.
 
841 - 860 of 23,546 Posts
About this Discussion
24K Replies
588 Participants
Last post:
HemuV2
Overclock.net
posts
28M
members
562K
Since
2004
The premier forum for overclocking experts and enthusiasts. Discuss hardware optimization, custom builds, benchmarking, cooling solutions, and pushing the boundaries of computing performance. From beginner guides to extreme overclocking, join our technical community to master system tuning.
Full Forum Listing
Explore Our Forums
Intel CPUs Hardware News NVIDIA PC Gaming Water Cooling