Overclock.net - An Overclocking Community - Reply to Topic
Thread: NB frequency max voltage 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)
01-13-2020 10:40 PM
ssateneth your "nb frequency" voltage is directly pulled from vcore. 1.45vcrore = 1.45v cache/nb frequency.
01-12-2020 12:18 AM
8051
Quote: Originally Posted by BroadPwns View Post
Differences here in 1/0.1% comes more from randomness rather than actual cache boost profit. You get better results by upping RAM frequency by one notch, actually. Sure, for min-maxing cache frequency should be raised, but there is barely any gain from doing so.
In the video I linked to the differences in the minimums were more than 10% for GTA V when overclocking the cache/uncore ditto for Shadow of the Tomb Raider. Some of the other results were around 5% which I'd think exceeds random fluctuations.
01-11-2020 05:56 PM
BroadPwns Differences here in 1/0.1% comes more from randomness rather than actual cache boost profit. You get better results by upping RAM frequency by one notch, actually. Sure, for min-maxing cache frequency should be raised, but there is barely any gain from doing so.
01-11-2020 04:16 PM
8051
Quote: Originally Posted by BroadPwns View Post
Northbridge is Cache/uncore. And no, it's not a bottleneck, ever. There's barely ever any gain from raising it, 5% is a dream improvement.
If you're not GPU bound you can see definite increases in min. FPS with uncore/cache overclocks in at least some games:
https://www.youtube.com/watch?time_c...ature=emb_logo
01-02-2020 09:47 AM
Comatosed Thanks for all your input guys, it seems this cpu+mobo is needing around 1.4vcore to be 100% stable for 5Ghz (without sacrificing NB frequency) which is a tad too high for me 24/7, i'll dial it back to 4.9 and fine tune that more which needs around 1.33.

EDIT* So testing SA to 1.25 and VCCIO to 1.24, I did just have it at 1.2/1.19 but still error'd in prime @ 50 mins, went to 1.22/1.21 then lasted 1hr 10 mins, so deffinitely onto something here. BIOS auto wants to set VCCIO @ 1.30, no idea what it sets SA at as it doesn't show. so staying under 1.30 for VCCIO is ideal.

EDIT2* Well it seems 1.25 SA and 1.24 VCCIO had an averse effect and crashed prime 95 within 20 mins, it seems now 1.22/1.21 is the sweetspot and my vcore was way off, what I thought was stable on 4.9 isn't, I do recall not doing a long P95 test but just to see, I set same Vcore as I was testing with 5.0 which is now 4.9 @1.36, spikes to 1.37 and currently 3hrs into P95 blend full AVX no problems so far.
Now i'm thinking if 5.0 @ 1.39 (spikes to 1.4) is not so bad overall, especially being stable 4.8 @1.30 ish, which is another freq I need to retest after these results, so that is more than likely closer to 1.32. The other thing I do dislike is the vcore reading in increments off 0.076
01-02-2020 08:31 AM
Falkentyne
Quote: Originally Posted by Comatosed View Post
Thank you so much for explaining that to me, i'm old school, my overclocking was always done on the BCLK lol I understand that now, I was getting so confused and thrown off because if I raised the SA and VCCIO voltage, it would take longer for the instability to show so I naturally presumed it was related to that.
Now to go back into BIOS ready for another prime95 test, my poor cpu has been at 100% solid past 2 weeks, at least i know i'm not getting any degradation at 1.37. time to go 1.38.


EDIT* Just realised i was at 1.36, currently testing 1.37 but I feel it will go an hour plus on P95 now with your insight. I will leave it till it errors. I really wanted to get stable @1.36 but i guess under 1.37-1.38 can't hurt, temps just barely touch the 80's on full AVX on 2 cores, rest are 70, the lowest core temp is the one that errors :|
Please stop using HWmonitor. It's cringeworthy seeing "VIN3 VIN4 VIN5" and strange random silly voltages no one knows what it's reporting.
use HWinfo64. That's what people use these days.

VCCIO controls the memory controller and shared L3 cache voltage. This is important on hyperthreaded cpus. Yes, your system has L3 cache and that's used, but you're forgetting about the L1 and L2 cache. L1 and L2 are directly controlled by Vcore. Raising VCCIO (VCCSA must be slightly higher than VCCIO) may help slightly, usually up to 1.2v (it helps the memory controller/IMC more) but after this point you gain more by upping vcore, if going past 1.25v VCCIO doesn't do anything anymore.

Upping PLL Termination Voltage (aka VTT, sometimes called VCC VTT; note this is not CPU PLL voltage or CPU PLL OC Voltage (PLL Bandwidth on Asus boards)) to 1.20v may allow you to maximize your cache ratio on ambient cooling. DDR VTT is not the same thing.
01-02-2020 07:07 AM
Comatosed Thank you so much for explaining that to me, i'm old school, my overclocking was always done on the BCLK lol I understand that now, I was getting so confused and thrown off because if I raised the SA and VCCIO voltage, it would take longer for the instability to show so I naturally presumed it was related to that.
Now to go back into BIOS ready for another prime95 test, my poor cpu has been at 100% solid past 2 weeks, at least i know i'm not getting any degradation at 1.37. time to go 1.38.


EDIT* Just realised i was at 1.36, currently testing 1.37 but I feel it will go an hour plus on P95 now with your insight. I will leave it till it errors. I really wanted to get stable @1.36 but i guess under 1.37-1.38 can't hurt, temps just barely touch the 80's on full AVX on 2 cores, rest are 70, the lowest core temp is the one that errors :|
01-02-2020 07:00 AM
acoustic Someone is going to kill you for continuing to call the CPU Cache speed as the NB frequency. Lol.

Your core voltage is also your cache voltage. If you're crashing at higher cache speeds that means the core voltage needs to go up. 1.15v VCCIO/SA is standard unless pushing high DRAM speeds.

Sacrificing your cache speeds down so low for more core speed isn't the best idea, imo. I know a lot of people say it's useless to OC the cache, but balance is best in all things. I'd try to get stability at least at 4300 on the cache.
01-02-2020 06:53 AM
Comatosed I'm pretty certain my vcore is stable, which leaves the cache voltage.
I know my cpu is stable @5.0 1.36vcore (i have even raised it to 1.37 to be sure), currently trying to over clock the RAM but first I'm trying to be stable on the NB frequency so all my roundup errors are related to my RAM o/c but right now they are related to my NB frequency, now you have helped me narrow it down to the cache & voltage. so is my cache voltage controlled by both the SA and VCCIO or neither ?
So far i know the VCCIO is the voltage for the memory controller, i'm still researching what the SA voltage does but right now i have them both at 1.16, I have tested up to 1.2 but still getting roundup error after 45mins prime.
Memory passes memtest86 too even with my oc setting, native 3000 cl15-17-17-35, was in testing to 3200 cl15 18-18-36 by just raising the ram volts +0.010 or at least I thought I was, so back to 3000 for this stability test.
01-02-2020 06:33 AM
BroadPwns Mister. It's not because of its frequency, but because your Core or cache voltage is too low.
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