5800X Boost set to 5050Mhz all cores with AMD Curve Optimizer

woppy101 · 4 Feb 2021 at 10:25

ollies said:
@Towel start at +0 boost, auto scalar, mobo limits and adjust curve first. See how negative you can go all core (usually -30) and then start to adjust based on how each core performs (can also check their peformance score in event viewer, ID 55).

Best cores won't need as much negative curve, but your worst can go quite far (majority of mine are -50). Once you find good curve vs performance you can start to add boost and then after that test out manual PPT, TDC and EDC, which should help reduce temps a bit. Leave scalar on auto.

What settings are you finding your PPT,TDC and EDC are needing?

heartburnron · 4 Feb 2021 at 15:30

woppy101 said:
What settings are you finding your PPT,TDC and EDC are needing?

My chip runs hot at stock with the wraith prism cooler, so reducing temps has been my driving factor when tweaking these stats. I find lowering PPT has the biggest impact on temps. The trade off is that this also reduces the frequency all cores boosts to in multi-core workloads although you can offset this by adding negative values in the curve optimiser, I have all my cores set at -20. This keeps my max temp at 80 and all cores boost to 4550 Mhz. I have PPT, TDC & EDC at 118, 118 and 90.

ollies · 4 Feb 2021 at 19:29

woppy101 said:
What settings are you finding your PPT,TDC and EDC are needing?

It really depends on the individual circumstances honestly, but you can certainly use values from others as a baseline. I am currently on PPT 210, TDC 150 and EDC 185. Majority of cores are set to -50, one is -25 as it kept throwing WHEA-Errors, whilst my best core is -10 and second best -15. As @heartburnron mentioned reducing PPT does seem to correlate with reduced temps as well, but you will sacrifice elsewhere so it's all about finding the right balance for your set up.

PBO can be unreliable (I can pass benches, stress tests and play games without issue, but then I'll get an idle or low load WHEA-Error randomly), but it's also fun and addictive to play with.

mikehhhhhhh · 4 Feb 2021 at 19:43

ollies said:
It really depends on the individual circumstances honestly, but you can certainly use values from others as a baseline. I am currently on PPT 210, TDC 150 and EDC 185. Majority of cores are set to -50, one is -25 as it kept throwing WHEA-Errors, whilst my best core is -10 and second best -15. As @heartburnron mentioned reducing PPT does seem to correlate with reduced temps as well, but you will sacrifice elsewhere so it's all about finding the right balance for your set up.

PBO can be unreliable (I can pass benches, stress tests and play games without issue, but then I'll get an idle or low load WHEA-Error randomly), but it's also fun and addictive to play with.

How did you isolate that single core as the source of your WHEA errors?

ollies · 4 Feb 2021 at 23:26

mikehhhhhhh said:
How did you isolate that single core as the source of your WHEA errors?

Event ID 18 in Event Viewer, check ApicId.

DarkBahamut · 5 Feb 2021 at 00:40

As far as I'm aware +/- 30 is the maximum for curve optimiser. You can enter higher values, but it just applies 30 counts.

Towel · 12 Feb 2021 at 14:08

@ollies

hey, had surgery so havent been checking this,

I will try that tonight and see how it goes, no idea how to use Event ID 18 in Event Viewer, and check ApicId. to isolate errors.

alec · 1 Mar 2021 at 10:15

Was tinkering with PBO, decided to try to increase LLC.
I considered LLC not useful (why would I want increased multicore voltage?) and kept it at lowest setting (maximum vdroop)
Turns out it improves stability.

Previously +150MHz setting with 0 CO was occasionally crashing (LLC Low). Now with LLC Medium same settings with CO -10 are stable.

Towel said:
I will try that tonight and see how it goes, no idea how to use Event ID 18 in Event Viewer, and check ApicId. to isolate errors.

https://www.overclockers.co.uk/foru...ow-to-configure-pbo-curve-optimizer.18920081/

RayRayFPS · 2 Mar 2021 at 21:40

Negative curve of 11 on all cores with PBO set to motherboard and a 125mhz increase gets me the below results. It seems to run pretty stable at this.

Dreamic · 3 Mar 2021 at 03:27

Grim5 · 3 Mar 2021 at 04:36

Dreamic said:
Another month of 24/7 power and -19 All Core just failed with a random reboot finally. I have to be 100% stable at like -15 right? Would take some months to find out for sure.
Was just idle browsing in Chrome for this one.
-15 All Core assuming that wouldn't reboot months down the road seems like a poor result... gonna wonder if a new BIOS will allow me more whenever a new comes out too, but it'll literally take more than a month to find out if it did or not each time lol... this is so dumb.
My i5 4690k overclocked to 4.5ghz with static vcore for like 6 years never had a problem reboot or BSOD and 4.6ghz wasnt stable with the same vcore, Intel was so much simpler if it didn't work it would let you know pretty fast not be fine for 700+ hours straight and then have an issue

I had issues like that in my i7 4790k - it run fine overclocked at 1.3v for 8 months and then one day just started crashing every 5 minutes until I pushed voltage to 1.325v

LtMatt · 6 Mar 2021 at 08:57

Dreamic said:
Another month of 24/7 power and -19 All Core just failed with a random reboot finally. I have to be 100% stable at like -15 right? Would take some months to find out for sure.
Was just idle browsing in Chrome for this one.
-15 All Core assuming that wouldn't reboot months down the road seems like a poor result... gonna wonder if a new BIOS will allow me more whenever a new comes out too, but it'll literally take more than a month to find out if it did or not each time lol... this is so dumb.
My i5 4690k overclocked to 4.5ghz with static vcore for like 6 years never had a problem reboot or BSOD and 4.6ghz wasnt stable with the same vcore, Intel was so much simpler if it didn't work it would let you know pretty fast not be fine for 700+ hours straight and then have an issue

You probably just have one core that is one of your two Best cores that does not like an undervolt much. My best core of my 5950x only likes -7 negative offset, yet it is the fastest cpu core on these forums. Look at the Cinebench R23 ST resultsin the bench thread.

gerardfraser · 7 Mar 2021 at 23:11

Dreamic said:
Another month of 24/7 power and -19 All Core just failed with a random reboot finally. I have to be 100% stable at like -15 right? Would take some months to find out for sure.
Was just idle browsing in Chrome for this one.
-15 All Core assuming that wouldn't reboot months down the road seems like a poor result... gonna wonder if a new BIOS will allow me more whenever a new comes out too, but it'll literally take more than a month to find out if it did or not each time lol... this is so dumb.
My i5 4690k overclocked to 4.5ghz with static vcore for like 6 years never had a problem reboot or BSOD and 4.6ghz wasnt stable with the same vcore, Intel was so much simpler if it didn't work it would let you know pretty fast not be fine for 700+ hours straight and then have an issue

Aren't you Canadian, stop making us look bad with S....D.

Dreamic · 8 Mar 2021 at 00:15

alec · 8 Mar 2021 at 14:40

Ooh, exciting
OCCT 8.0.0 has new features that will surely help find hidden instability when using CO.
Ability to cycle load between cores and run intermittent loads
https://www.ocbase.com/

leveluprob · 9 Mar 2021 at 22:50

I'm having some major issues with my 5800x

Being "cooled" with a deep cool 240aio and I've not touched anything and In the bios at idle it's reporting 67C yet ryzen master is reporting 92C

I've re set the cooler twice now and I've

ordered another cooler just to see if that's the issue. Motherboard is a crosshair 8 hero running the latest bios ( the one needed for ryzen 3rd gen)

Never had this issue before installing CPUs over the years.

I've seen people mention the default volts are higher but I'm a little lost when it comes to undervolting.

Just to add I tired cinebench at stock settings and it shot up to 105C and switched off. So something definitely is not right

Joxeon · 9 Mar 2021 at 23:16

leveluprob said:
I'm having some major issues with my 5800x

Being "cooled" with a deep cool 240aio and I've not touched anything and In the bios at idle it's reporting 67C yet ryzen master is reporting 92C

I've re set the cooler twice now and I've

ordered another cooler just to see if that's the issue. Motherboard is a crosshair 8 hero running the latest bios ( the one needed for ryzen 3rd gen)

Never had this issue before installing CPUs over the years.

I've seen people mention the default volts are higher but I'm a little lost when it comes to undervolting.

Just to add I tired cinebench at stock settings and it shot up to 105C and switched off. So something definitely is not right

Sounds like a contact issue. Either that or the coolers faulty.

leveluprob · 10 Mar 2021 at 17:40

Joxeon said:
Sounds like a contact issue. Either that or the coolers faulty.

I'm hoping that's the case. I'll try the new cooler and hopefully that fixes it

Dreamic · 10 Mar 2021 at 23:19

leveluprob · 11 Mar 2021 at 19:50

It definitely was the cooler got the freezer 240 installed and after messing around with voltages "default was like 1.4??? I got it to this