cancel
Showing results for 
Search instead for 
Did you mean: 

LLC Quickie

Theagg
Level 7
Older bios had LLC levels described as 'normal', 'high, 'extreme' etc. Which was self explanatory

This current UEFI has them simply listed as 'level 1', 'level 2' etc.

So, which level is the highest (or extreme etc ). Is it level 1, or level 5. Browsing around I find conflicting answers. Some seem to suggest level 1 is highest, others level 5 ?!
3,754 Views
7 REPLIES 7

MNMadman
Level 8
For some boards, the highest is Level 1.

For the C6H board, the highest is Level 5.

Most people recommend that you stay away from the highest levels of LLC as they tend to over-volt by quite a bit.

MNMadman wrote:
For some boards, the highest is Level 1.

For the C6H board, the highest is Level 5.

Most people recommend that you stay away from the highest levels of LLC as they tend to over-volt by quite a bit.


Thanks for clarifying that. Level 2 or 3 seem to hold the voltage fine at ~1.375v on my OC (4015 Mhz with a +0.025 offset)

setting it to 5 is major overkill. I need slightly more than 1.4v to get 4.1ghz stable so I adjusted my offset to give me 1.395v in bios while setting my LLC at 4 which gives me the tiny boost to 1.417v under heavy loads.

AfterShock wrote:
setting it to 5 is major overkill. I need slightly more than 1.4v to get 4.1ghz stable so I adjusted my offset to give me 1.395v in bios while setting my LLC at 4 which gives me the tiny boost to 1.417v under heavy loads.


I haven't tried 4.1 yet but at 4.015 I'm pretty close, which just goes to show the variance of the Ryzen I guess, as my offset is +0.01875v and the system is stable, with LLC at 3 and the load volatage (Cinebench) staying put at 1.352v

I would also recommend grabbing real bench for a stability test, it stresses the system as a whole. I would do a run of cinebench and if that passed I would do 15 min of realbench using the 8 gig memory preset if that passed I pushed a little further, rinse and repeat, my final stability test was a 4 hour run of realbench.

I'm at the brick wall of my chip, any further meaningful increase requires too much voltage. it took 1.375 at 4ghz to 1.4+ at 4.1 which I was still ok with but thats about my limit with this chip.

AfterShock wrote:
I would also recommend grabbing real bench for a stability test, it stresses the system as a whole. I would do a run of cinebench and if that passed I would do 15 min of realbench using the 8 gig memory preset if that passed I pushed a little further, rinse and repeat, my final stability test was a 4 hour run of realbench.

I'm at the brick wall of my chip, any further meaningful increase requires too much voltage. it took 1.375 at 4ghz to 1.4+ at 4.1 which I was still ok with but thats about my limit with this chip.



My Swiftech H220x AIO is drying up (resevoir now just above the minimum mark) so although it's working fine at the moment, I might wait til thats sorted out before trying lengthy and thus consistently hot stress tests. But yes, will locate real bench.

Other than that, rather than trying to push from 4015 to 4100, with the small percentage gain that will give, I'm more interested in getting my RAM over the 3000mhz mark as from numerous testers that looks to be where higher gains will be made, rather than an extra 85mzh of CPU speed. I'm currently stuck on 2933mhz with my 32gb of Vengeance LPX 3200, but since someone else has posted they got their 32gb of Dominator to 3200mhz, I'm thinking the Vengeance just doesn't have what it takes at the moment, at least not on this bios (1002)

Raja
Level 13
LLC 1 works quite well for everyday use. Just a stock capture here. That said, the CPUs don't pull much current at all, so I don't expect big changes. Small FFT Prime load hitting the VRM.

63870

63871


Blue trace = Vcore
Yellow trace = current (peak is around 130W)

1.40 VID set in UEFI gives around 1.38V "idle", dipping no more than 20mV when hit with load.