Overclock.net banner
4,281 - 4,300 of 14,355 Posts
overclocking (memory or otherwise) has always existed in this grey area - back in the day it flat out invalidated your warranty (if they could tell). now the vendors recognise it and cater to it, and obviously they use it to differentiate their products, hence the heavy marketing budget. however the disclaimer has always existed, even if people weren't aware of it.

you can't have your cake and eat it, if you want guaranteed performance and an ironclad warranty, go and buy a dell office box. if you want to play on the margins and see what 'free' performance you can get out the hardware, then understand the risks.

until we see evidence to the contrary, i assume the warranties will continue to operate as they have for the past 2 decades, which is - if you behave sensibly and overclock within reasonable margins they will cover you, partly because they won't be able to tell that you've done any overclocking anyway. people are right to point out that it would be foolish for them to start denying warranties for people using the most 'low-hanging fruit' of the oc features, features that AMD and the mobo makers have heavily marketed - i.e. expo and 'auto overclocking' like PBO.

equally it is foolish for people to start agitating for class actions and the like...if you get AMD/Intel and mobo vendors running scared of liabilities and legal actions, then you will get your cast iron warranties, but they won't be what you want...they will remove all overclocking support and then you can enjoy using your dell office boxes because that is all that will be available. it is to everybody's benefit for it to continue to operate in the "don't ask, don't tell" fashion that it has.
 
Discussion starter · #4,282 ·
overclocking (memory or otherwise) has always existed in this grey area - back in the day it flat out invalidated your warranty (if they could tell). now the vendors recognise it and cater to it, and obviously they use it to differentiate their products, hence the heavy marketing budget. however the disclaimer has always existed, even if people weren't aware of it.

you can't have your cake and eat it, if you want guaranteed performance and an ironclad warranty, go and buy a dell office box. if you want to play on the margins and see what 'free' performance you can get out the hardware, then understand the risks.

until we see evidence to the contrary, i assume the warranties will continue to operate as they have for the past 2 decades, which is - if you behave sensibly and overclock within reasonable margins they will cover you, partly because they won't be able to tell that you've done any overclocking anyway. people are right to point out that it would be foolish for them to start denying warranties for people using the most 'low-hanging fruit' of the oc features, features that AMD and the mobo makers have heavily marketed - i.e. expo and 'auto overclocking' like PBO.

equally it is foolish for people to start agitating for class actions and the like...if you get AMD/Intel and mobo vendors running scared of liabilities and legal actions, then you will get your cast iron warranties, but they won't be what you want...they will remove all overclocking support and then you can enjoy using your dell office boxes because that is all that will be available. it is to everybody's benefit for it to continue to operate in the "don't ask, don't tell" fashion that it has.
Yes, why do people buy enthusiast gear when they are not prepared to do enthusiast things?
 
Creeping closer to Grabibus CB23 highscore, 1410 ain't that bad and also it HAS 24/48-module support!

View attachment 2613205
View attachment 2613206

ROG CROSSHAIR X670E EXTREME BIOS 1410

Version 1410 beta release
9.57 MB 2023/05/02
"1. Update AGESA version to Combo AM5 PI 1.0.0.7.
2. Support 48/24GB high-density DDR5 memory module.
3. SoC voltage for Ryzen 7000 series limited to a maximum of 1.30V to protect the CPU and motherboard.

Before running the USB BIOS Flashback tool, please rename the BIOS file (CX670EE. CAP) using BIOSRenamer.

Please note that this is a beta BIOS version of the motherboard which is still undergoing final testing before its official release. The UEFI, its firmware and all content found on it are provided on an “as is” and “as available” basis. ASUS does not give any warranties, whether express or limited, as to the suitability, compatibility, or usability of the UEFI, its firmware or any of its content. Except as provided in the Product warranty and to the maximum extent permitted by law, ASUS is not responsible for direct, special, incidental or consequential damages resulting from using this beta BIOS."

Rectangle Screenshot Font Software Slope


Still with 24/7 stable settings :
Medium Load Boostit "Enabled" // PBO "Manual" // PPT~TDC~EDC : 750W~500A~650A // Max CPU Boost Clock Override = +200MHz // Per-Core Boost Clock Limit "Enabled" : 5800MHz, 5850MHz, 5850MHz, 5800MHz, 5800MHz, 5750MHz, 5850MHz, 5800MHz, 5650MHz, 5650MHz, 5600MHz, 5650MHz, 5650MHz, 5650MHz, 5650MHz, 5650MHz // CO : -10, -10, -13, -15, -18, -10, -15, -18, -23, -23, -15, -23, -23, -23, -23, -23.
Bios 1410 beta
22°C ambient.

:cool:
 

Attachments

There is a power plan from @ManniX-ITA for Zen4 :


Provide power savings and in the mean time good performances and good latency. I use it daily.

Hope we can get from him soon the same one but Ultimate High Power's one for Ultimate perfomances and very good latency :p
 
Thanks! That's what i have been doing and I have been passing OCCT. Unfortunately, and I know this isn't good, but I cant pass CC with YC even without PBO. I been able to pass YC on its own tho and IDK why.... Also, i am runing YC now on core 8 (which is one of the cores that fails on CC as long as everyone after), but HWinfo64 and RM show core 5 getting the work load. Something is wrong and IDK what it is. Any suggestions? Core scheduling in windows?

resource monitor shows CPU 8 getting the workload Im confused lol

View attachment 2613296
View attachment 2613297
It's probably "thread 8" not actually core 8. Thread 8 starting from zero is on core 4 (starting from zero)

Threads 0 and 1 are on core zero, two threads due to hyper threading.
2,3 core 1
4,5 core 2
6,7 core 3
8,9 core 4
 
One question about usb4 if I use one cable usb type c hdmi or usb type c dp in the extreme or hero is possible ?


Inviato dal mio iPhone utilizzando Tapatalk
 
Ok so today after doing some normal RDP related work, and gaming with some Star Citizen in between things, I checked out my HWInfo Temps and get this.

Font Screenshot Electric blue Multimedia Audio equipment


What the heck is going on here? I am hoping just a reporting error but someone else posted on 1410 similar things and I am just confused. We shouldn't be seeing 2V Anywhere on literally anything, ever, outside of LN One time OC achievements.

SOC Is manually set to 1.3v, MC is set to 1.35v, and Misc is I believe Auto but doesn't move from 1.1v 99% of the time.

What gets me to think this isn't just a reporting error is the 109C Die Average, even though nothing else shows anything hotter than 90C for the Hotspot. So something actually got some average to 109C which is pretty nuts. So is this something that is "Fixed" in 1410/Agesa 1.0.0.7 Or not? if I am getting it out of the blue on Bios 1001 even with manual set voltages, and others with manual set voltages still get the occasional spike readings, something else may be up... I doubt it's hot enough to Hurt things as long as it's only a very short time spike, but what the hell and why is the question. Could just a -20/+200 PBO cause this? Is this some ASUS thing even though LLC hasn't been touched and like I said, I haven't pushed really anything even the game SC doesn't push the CPU all that hard 90% of the time.
 
So, today I wanted to try overclocking with the external clock to 105, but by mistake I set Bclk1 to 105 instead if Bclk2. I realised the mistake too late and the system would not even post. I had to clear Cmos and reset everything.
All seems to be working fine now.

Was it really bad? I hope I didn't damage the CPU or the board.
 
Ok so today after doing some normal RDP related work, and gaming with some Star Citizen in between things, I checked out my HWInfo Temps and get this.

View attachment 2613395

What the heck is going on here? I am hoping just a reporting error but someone else posted on 1410 similar things and I am just confused. We shouldn't be seeing 2V Anywhere on literally anything, ever, outside of LN One time OC achievements.

SOC Is manually set to 1.3v, MC is set to 1.35v, and Misc is I believe Auto but doesn't move from 1.1v 99% of the time.

What gets me to think this isn't just a reporting error is the 109C Die Average, even though nothing else shows anything hotter than 90C for the Hotspot. So something actually got some average to 109C which is pretty nuts. So is this something that is "Fixed" in 1410/Agesa 1.0.0.7 Or not? if I am getting it out of the blue on Bios 1001 even with manual set voltages, and others with manual set voltages still get the occasional spike readings, something else may be up... I doubt it's hot enough to Hurt things as long as it's only a very short time spike, but what the hell and why is the question. Could just a -20/+200 PBO cause this? Is this some ASUS thing even though LLC hasn't been touched and like I said, I haven't pushed really anything even the game SC doesn't push the CPU all that hard 90% of the time.
Which bios? NVM i see you said bios 1001 which doesn't exist anymore on asus dark web đź’€...
AMD extremist here will tell you thats normal đź‘€
 
What the heck is going on here? I am hoping just a reporting error but someone else posted on 1410 similar things and I am just confused. We shouldn't be seeing 2V Anywhere on literally anything, ever, outside of LN One time OC achievements.

SOC Is manually set to 1.3v, MC is set to 1.35v, and Misc is I believe Auto but doesn't move from 1.1v 99% of the time.

What gets me to think this isn't just a reporting error is the 109C Die Average, even though nothing else shows anything hotter than 90C for the Hotspot. So something actually got some average to 109C which is pretty nuts. So is this something that is "Fixed" in 1410/Agesa 1.0.0.7 Or not?
Overboost bug still showing its existence :)
Its not a readout-bug. 10 - 50 ghz spikes exist since 3-4 years now.

The issue is the slingshot boost , bypassing P-States on a full Package C6 suspension.
Deactivating DF-C States and ontop putting Powersuply Current Control to "Constant current"
Yet still enabling Global C-States ~ mitigates/bandaids this problem

To make matters worse, this goes hand in hand with BAR overboost, that can slingshot RDNA2/3 likely too - to 7ghz and crash.
Optimally its just an annoyance with too harsh powerplans and broken boosting system.
But its a bug that is put under the carpet since many years now. First discovered by me ~ but as sadly to say, nobody listens.

This is a global issue plaguing 3 generations now, and its ignored ~ because i refuse to believe its "not found".
So far i haven't noticed it on 1206++ (1410) ~ it doesnt mean they finally took care of it, but i haven't seen it.
It was another suspicion why they could die but that is just momentarily & could, but shouldn't fry a chip ~ yet, its a thing and not an SMU reporting mistake.
Long enough in Zen "industry" to tell you that.
// Reminder question, why VRM even allows 3v on non LN2 mode although was playing devils advocate there. 🤭

Disable DF-C States and its ok for now. Loss of boost, but oh well.
I do fine too, soo no reason why you can't. :)
You wouldnt even need MediumBoostIT here, but could use it if need arrises & you get capped at c-state limiter.

Actually outside of even having build a powerplan that can force trigger this (by accident ~ soo never released followup of promise, too dangerous)
Its not such a big deal for you to not have Package C6 states. Normal C6 are ok which RM still will report as sleeping or hibernated.
==================================
Honestly i've never seen VRM follow those requests.
That i doubt a bit, else sample would be burned/dead.
But clock hold for a ms and VID requests over 1.62 to 1.68v ~ were a thing and are a thing.
It's becoming "normal" , and once again ~ i believe its not ignored. But i will not say this is resolved yet.
Saw 7800X3D also spike to 5.8ghz on early bioses, tho without tracked VID request, funnily.
Not that HWInfo is fast enough, but you get my point :)
Is this some ASUS thing
Its an AMD thing.

ASUS userbase is just "bigger" , soo oddness is found easier and complains are harsher
~ as pricetag is higher and expectation is bigger with this Boardpartner.
Most are simply AMDs fault, remain ~ oh well. This here is an AMD overworked, thing~
 
Overboost bug still showing its existence :)
Its not a readout-bug. 10 - 50 ghz spikes exist since 3-4 years now.

The issue is the slingshot boost , bypassing P-States on a full Package C6 suspension.
Deactivating DF-C States and ontop putting Powersuply Current Control to "Constant current"
Yet still enabling Global C-States ~ mitigates/bandaids this problem

To make matters worse, this goes hand in hand with BAR overboost, that can slingshot RDNA2/3 likely too - to 7ghz and crash.
Optimally its just an annoyance with too harsh powerplans and broken boosting system.
But its a bug that is put under the carpet since many years now. First discovered by me ~ but as sadly to say, nobody listens.

This is a global issue plaguing 3 generations now, and its ignored ~ because i refuse to believe its "not found".
So far i haven't noticed it on 1206++ (1410) ~ it doesnt mean they finally took care of it, but i haven't seen it.
It was another suspicion why they could die but that is just momentarily & could, but shouldn't fry a chip ~ yet, its a thing and not an SMU reporting mistake.
Long enough in Zen "industry" to tell you that.
// Reminder question, why VRM even allows 3v on non LN2 mode although was playing devils advocate there. 🤭

Disable DF-C States and its ok for now. Loss of boost, but oh well.
I do fine too, soo no reason why you can't. :)
You wouldnt even need MediumBoostIT here, but could use it if need arrises & you get capped at c-state limiter.

Actually outside of even having build a powerplan that can force trigger this (by accident ~ soo never released followup of promise, too dangerous)
Its not such a big deal for you to not have Package C6 states. Normal C6 are ok which RM still will report as sleeping or hibernated.
==================================
Honestly i've never seen VRM follow those requests.
That i doubt a bit, else sample would be burned/dead.
But clock hold for a ms and VID requests over 1.62 to 1.68v ~ were a thing and are a thing.
It's becoming "normal" , and once again ~ i believe its not ignored. But i will not say this is resolved yet.
Saw 7800X3D also spike to 5.8ghz on early bioses, tho without tracked VID request, funnily.
Not that HWInfo is fast enough, but you get my point :)

Its an AMD thing.

ASUS userbase is just "bigger" , soo oddness is found easier and complains are harsher
~ as pricetag is higher and expectation is bigger with this Boardpartner.
Most are simply AMDs fault, remain ~ oh well. This here is an AMD overworked, thing~
I Knew I saw you post something about this not too far back in this thread. Damn what a goofy bug to just leave in there. If it was just the ultra momentary voltage blips alone it probably isn't much to worry about, but to see HWinfo catch the 109C Die Average reading means something got hot for a split second above the 109C itself even. Grr
 
Somebody please help me with an issue that I can't seem to figure out on x670 hero board. The training of memory on 1410 bios seems wonky. No matter the MBIST\DFE\MCR settings. I set-up very good tune settings that gives me 7.9 pyprime runs, SOTR 715FPS, consistently for the whole time, until I restart my PC. The re-training happens, and next time I come to Windows its 9 pyprime and SOTR 706 fps consistently. Then I will need again like 10-15 re-trainings, fiddling with settings, to again catch up one good training it seems, which will again yield me the same best results for the time until next restart. Does anybody know how to fix this stuff? Or what could be the issue?
 
One question about usb4 if I use one cable usb type c hdmi or usb type c dp in the extreme or hero is this possible @Veii?
 
Somebody please help me with an issue that I can't seem to figure out on x670 hero board. The training of memory on 1410 bios seems wonky. No matter the MBIST\DFE\MCR settings. I set-up very good tune settings that gives me 7.9 pyprime runs, SOTR 715FPS, consistently for the whole time, until I restart my PC. The re-training happens, and next time I come to Windows its 9 pyprime and SOTR 706 fps consistently. Then I will need again like 10-15 re-trainings, fiddling with settings, to again catch up one good training it seems, which will again yield me the same best results for the time until next restart. Does anybody know how to fix this stuff? Or what could be the issue?
You are probably on the edge of stability. Soo there is more variable factors during the training.
You should leave Memory context restore to Disabled, and try to adjust Voltages or ODT or RTT´s or whatever to have consistent training.
 
You are probably on the edge of stability. Soo there is more variable factors during the training.
You should leave Memory context restore to Disabled, and try to adjust Voltages or ODT or RTT´s or whatever to have consistent training.
I can't seem to find what would be on the edge. All tests pass and the tune is not that special or extreme. It's a rather low setup, but when the training is good, then it yields very good results. My ODT\RTT's are at auto. I should probably try to tune them, but whatever I have set there doesn't seem to have much effect. I've usually gone with procODT 40\48ohm and Dram drive stregth at 40. Rest I am leaving at auto. VDD is 1.41v, 6000MT\cl28 and vddq is -100mv like it should be. Rest is pretty much auto.
 
Discussion starter · #4,299 · (Edited)
View attachment 2613344

Still with 24/7 stable settings :
Medium Load Boostit "Enabled" // PBO "Manual" // PPT~TDC~EDC : 750W~500A~650A // Max CPU Boost Clock Override = +200MHz // Per-Core Boost Clock Limit "Enabled" : 5800MHz, 5850MHz, 5850MHz, 5800MHz, 5800MHz, 5750MHz, 5850MHz, 5800MHz, 5650MHz, 5650MHz, 5600MHz, 5650MHz, 5650MHz, 5650MHz, 5650MHz, 5650MHz // CO : -10, -10, -13, -15, -18, -10, -15, -18, -23, -23, -15, -23, -23, -23, -23, -23.
Bios 1410 beta
22°C ambient.

:cool:
Bah! I'm just 89 points behind now :p

Only change using the MANNIX Zen 4 powerplan

Product Font Window Screenshot Software

Font Technology Screenshot Electric blue Multimedia


Mannix did have better latency, as you said:

Font Electronic device Screenshot Software Multimedia
 
  • Haha
Reactions: GRABibus
4,281 - 4,300 of 14,355 Posts