Overclock.net banner
9561 - 9580 of 10745 Posts

·
Registered
Joined
·
267 Posts
Does it matter that mine are 16GB dimms?
---
What would that be?

(sorry to ask, kind of a noob on memory fine tuning)
I am not running them right now, and this was really kind of half assed on the timings, but:
2514464


I don't know how well PBO really helps or hurts the numbers, but that is default.

I think generally the effect of bin for Rev E is mostly lower tRCD with higher bin. I also need 20t tRCDRD at 3800 for various 3200 sticks, the 3600 bin should always be lower.

I didn't test tWR down at this speed, tWTR_, or fix tWRRD. So those may not be quite right. I leave tRP looser on these, closer to tRC - tRAS rather than try to push them down.
 

·
Registered
Joined
·
15 Posts
I am not running them right now, and this was really kind of half assed on the timings, but:
View attachment 2514464

I don't know how well PBO really helps or hurts the numbers, but that is default.

I think generally the effect of bin for Rev E is mostly lower tRCD with higher bin. I also need 20t tRCDRD at 3800 for various 3200 sticks, the 3600 bin should always be lower.

I didn't test tWR down at this speed, tWTR_, or fix tWRRD. So those may not be quite right. I leave tRP looser on these, closer to tRC - tRAS rather than try to push them down.
Pretty similar to what I'm running. Some are the same, some lower, some higher. But I have the same test results, except L3 cache, mine are much higher (560 BG/s) but I think that one depends on PBO.

Gonna try to copy this and see if I get better results.
 

·
Registered
B550 gaming edge, 5800x, 5600x, 3600xt, 1600AF, 2070S, 4x8 3200c14, 2tb adata 8200sx PRO, 500gb 970e
Joined
·
569 Posts
It usually means CPPC is disabled in bios or somewhere else.
thats what i thought too, but its turned on (in all three places) well, on auto for DF-Cstates/CPPC cores,
the one thing thats strange is this tho,

2514483


(CTR showing ALL CORES as 131?)
perhaps on my board turning off DF-STATES auto kill CPPC?
"NOTES" running core-cycler has passed 7 iterations now. (all FFT sizes)
my CO is
CORE 1 -15CORE 2 -20
CORE 3 -15CORE 4 +5
CORE 5 -11CORE 6 -10
(NO offset running this time around) if CPPC was turned off would cores idle down still ?
considering ive ran core-cycler for this long and passed, i suppose i may just LEAVE THIS THE HELL ALONE
since im running better than ever. about to run some benchmarks and test with what i did have before hand.


(EDIT) does anyone else have this many options for C-states, CPPC?
2514489

2514490

2514491
 

·
Registered
Joined
·
582 Posts
but its turned on
If so, then acpi seem broken. Try to run CB Rxx single core test and look at the cores, which are under load.
 

·
Overclock the World
Joined
·
2,969 Posts
CTR turns on DF-C States once you run diagnose , or run the tool/use the tool
ZenStates will report it's off, but CTR turns it on temporary till you press the "Exit" button

EDIT:
This is the old Zenstates 2.0.Alpha
irusanovBG doesn't want to publish it yet, but it's not forbidden to me
Don't blame him for bugs, or complain about the tool please (this was also his request)
PBO stuff doesn't work for example
But the flags do work

This has to be off
EDC value will also missreport and the bottom field doesn't function

If bioses lack DF-C States, which this tool on autoruns, you can always disable them to prevent random reboots
@mongoled 8ghz example had DF-C states on, as it overboosted that high
* I publish it, because the lack of bios options from MSI and Gigabyte (closed down AMD CBS) starts to get annoying
 

·
Registered
B550 gaming edge, 5800x, 5600x, 3600xt, 1600AF, 2070S, 4x8 3200c14, 2tb adata 8200sx PRO, 500gb 970e
Joined
·
569 Posts
CTR turns on DF-C States once you run diagnose , or run the tool/use the tool
ZenStates will report it's off, but CTR turns it on temporary till you press the "Exit" button

EDIT:
This is the old Zenstates 2.0.Alpha
irusanovBG doesn't want to publish it yet, but it's not forbidden to me
Don't blame him for bugs, or complain about the tool please (this was also his request)
PBO stuff doesn't work for example
But the flags do work

This has to be off
EDC value will also missreport and the bottom field doesn't function

If bioses lack DF-C States, which this tool on autoruns, you can always disable them to prevent random reboots
@mongoled 8ghz example had DF-C states on, as it overboosted that high
* I publish it, because the lack of bios options from MSI and Gigabyte (closed down AMD CBS) starts to get annoying
completly reset the bios (not reflashed)
ctr showed (default values which showed my worst core as 125 and 129 the rest were ALL 131.
(hadnt ran diag on CTR tho, just opened it to see what core scores were)
also used core cycler tuner as well which shows the values listed below.
(the values below are with a new profile but with all my values entered)
just no turning off CPPC, DF states are off in one area of the bios. the rest are still auto.
(did try all other areas off) but still the same outcome. CPPC order is 123564??? used to be 461235 lol.
if what you shared doesnt work, ill reflash the bios. (only the none modded version)
and see if that changes anything.
2514492
 

·
Registered
B550 gaming edge, 5800x, 5600x, 3600xt, 1600AF, 2070S, 4x8 3200c14, 2tb adata 8200sx PRO, 500gb 970e
Joined
·
569 Posts
HMMM, now the "active cores" that were NOT ACTIVE EVER using the "tool" monitor feature (just opened it for the first time (TODAY)
shows that core 3 and 7 are no longer active. but before it was core 2 and 4 being turned off....?
...i broke something....

2514493
 

·
Overclock the World
Joined
·
2,969 Posts
if what you shared doesnt work, ill reflash the bios. (only the none modded version)
and see if that changes anything.
use AFUWIN, windows flashing utility and override all blocks
else there is a partition aside from AMD CBS, aside from CMOS which brands have no access to
USB dropout issues by "unclean" updates - especially on MSI sit there

It doesn't flash cleanly the normal way
but ignoring that recent 21H1 is buggy with CPPC and 1202 is buggy + AMD took AGESA 1200 down ~ while every boardpartner doesn't have 1203A out

=====================
On CTR RC06
use the Update, CTR thing
There are bugfixes there too, latest version is revision 23
Everything kind of feels broken right now. Still waiting for 1203A for X570 MSI & ASRock boards

CPPC broke for this guy too

AMD Gremlin's , many little gremlin's
 

·
Registered
B550 gaming edge, 5800x, 5600x, 3600xt, 1600AF, 2070S, 4x8 3200c14, 2tb adata 8200sx PRO, 500gb 970e
Joined
·
569 Posts
use AFUWIN, windows flashing utility and override all blocks
else there is a partition aside from AMD CBS, aside from CMOS which brands have no access to
USB dropout issues by "unclean" updates - especially on MSI sit there

It doesn't flash cleanly the normal way
but ignoring that recent 21H1 is buggy with CPPC and 1202 is buggy + AMD took AGESA 1200 down ~ while every boardpartner doesn't have 1203A out

=====================
On CTR RC06
use the Update, CTR thing
There are bugfixes there too, latest version is revision 23
Everything kind of feels broken right now. Still waiting for 1203A for X570 MSI & ASRock boards

CPPC broke for this guy too

AMD Gremlin's , many little gremlin's
i actually figured out what the issue was.
STRANGE ISSUE AT THAT. using the windows power utility (inside your gdrive) the same
power utility editior for unlocking/unhiding windows power stuff that i used that i set 0mhz as my core last time
and made a potato out of my OS,
none the less, i set 5000mhz as max for well pic below.
and thats what was causing this issue?
had already flashed the bios before seeing this reply.
(which did NOT fix this issue, importing my saved power file immediately solved this issue without needing to restart
nor did re-installing chipset drivers "latest version" either)
(i dont have USB dropout issues, which is why i asked if anyone would like to see whats "default"
inside the advanced options tabs, ALL of them a while back.. perhaps something inside the bios on this boards on or off that may affect whea, and USB issues?
(atm default core score (only changing PBO leaving mhz to auto in bios) are
now where they should be, CTR was reading wrong as they were all 131? (before fixing)
now its, 137+133+129+140+125+140
(was never a 131 score period other than my "worst" core with 200mhz auto oc enabled?)
still dont understand why active cores using the "tool" amd Vermeer monitor changed...
i gotta watch myself on programs like power explorer lol else, ill have a case of issues and scratching my head trying to figure out what i changed.
will now attempt to re-run core cycler overnight with DF-STATES off and CPPC ON.

i was smart enough this time around to have made a backup of my power plans using export and saved said backup
to my gdrive. still, setting 5000 should not have affected this should it not?

2514495
 

·
Overclock the World
Joined
·
2,969 Posts
i was smart enough this time around to have made a backup of my power plans using export and saved said backup
to my gdrive. still, setting 5000 should not have affected this should it not?
A frequency limit there is helping if DF_C-States enable themself
soo in case of an overboost, there are two more limiters who cut down the spike, preventing a reboot

It does nothing in harm, except hardcap the maximum frequency you can run
 

·
Registered
B550 gaming edge, 5800x, 5600x, 3600xt, 1600AF, 2070S, 4x8 3200c14, 2tb adata 8200sx PRO, 500gb 970e
Joined
·
569 Posts
A frequency limit there is helping if DF_C-States enable themself
soo in case of an overboost, there are two more limiters who cut down the spike, preventing a reboot

It does nothing in harm, except hardcap the maximum frequency you can run
i seen you mentioned this, which is why i set it in the first place.
but removing that seemed to have restored CPPC-CORES back to 4-6-1-2-3-5 order vs 1-2-3-4-5-6 or whatever it was.
going to try again to set this 5000mhz limit (should i set it for max processor freq, and the option below that as well?)

(when i booted into windows it was showing cores being 1-2-3-4-5-6 or whatever for CPPC
which was after re-flashing bios, re-installing drivers, re-enabling my RAM profile and CO tune,
but nothing had changed the order. once i removed the 5000mhz limit (only on the one i highlighted is where it was placed)
did CPPC then begin to show 4-6-1-2-3-5 which is what it used to show.)

perhaps, if your experiencing the same thing im getting, if you changed this, import your saved power file as well.
and see if it changes? (just a thought)
 

·
Registered
B550 gaming edge, 5800x, 5600x, 3600xt, 1600AF, 2070S, 4x8 3200c14, 2tb adata 8200sx PRO, 500gb 970e
Joined
·
569 Posts
CONFIRMED (posted screenshots are on the SAME BOOT meaning i have NOT restarted windows, simply set 5100mhz in power explorer utility
and checked OCCT which said CPPC cores were 123654 then i imported my saved profile, and closed OCCT reopened it, and CPPC cores were
461235
(CPPC cores before changing any values in power explorer was correctly shown, was only after did it change CPPC cores order.
(yes my EDC value is absurd but, im testing some-things) also take note in 6x scaler,
so far this has been the go-to value for this chip.
no overvolting noticed, no crashing, (still need to re-run core-cycler) then y-cruncher or together to see
but the curve offset inside the bios now is giving me -30 +50mv offset with 1A EDC scores. (+ or - 20 points)
12111 r23 4663 r20 (2 runs each) 3rd run dropped R23 to 11998) something i killed started back up...

perhaps, its not the 5000mhz thats causing this issue in power explorer, rather another setting ive not changed thats defaulted
(then again, that doesn't make much sense considering i exported before changing 5000mhz value, changed nothing else only 5000mhz as shown in
previous post.)

confused to whats going on with this, but @Veii maybe you should revert the change if you made one on power explorer and see if you notice a change as well.
shouldnt need to reboot to see it, i didnt anyhow as mentioned above.

if anyone else can speak on what setting inside power explorer is causing this, lmk. would like to be able to use this without fear...

2514498

2514499
 
  • Rep+
Reactions: Veii

·
Overclock the World
Joined
·
2,969 Posts
perhaps, if your experiencing the same thing im getting, if you changed this, import your saved power file as well.
and see if it changes? (just a thought)
I will test
Figured that ManniX-ITA's powerplans break CPPC on 21H1 somehow
But this did not do anything before - will try :)
 
  • Rep+
Reactions: craxton

·
Registered
B550 gaming edge, 5800x, 5600x, 3600xt, 1600AF, 2070S, 4x8 3200c14, 2tb adata 8200sx PRO, 500gb 970e
Joined
·
569 Posts
I will test
Figured that ManniX-ITA's powerplans break CPPC on 21H1 somehow
But this did not do anything before - will try :)
im not running any modified power plans.
i am running 21H1 tho, so the update in windows might be the issue.
 

·
Registered
Joined
·
1,779 Posts
11950-11970. I have a CLC280, runs CB23 at 80°.
Thats running CB23 at priority "Realtime" ??

If the above run is "Realtime" can you do another run but with normal priority with HWInfo64 open and note what all core frequency is ?

I would have expected your score to be higher, but it looks like you are thermal limited.

See below pic from my setup of CB23 where CPU temp is at 73-74C while frequency is at 4660 mhz

cb23-pbo-freq-temp.jpg

I wonder if I can somehow simulate an 80C CB23 run to see how much frequency drops by.

My estimate of being close to 123xx for your CPU is based on CB23 not exceeding 75C.

CTR turns on DF-C States once you run diagnose , or run the tool/use the tool
ZenStates will report it's off, but CTR turns it on temporary till you press the "Exit" button

EDIT:
This is the old Zenstates 2.0.Alpha
irusanovBG doesn't want to publish it yet, but it's not forbidden to me
Don't blame him for bugs, or complain about the tool please (this was also his request)
PBO stuff doesn't work for example
But the flags do work

This has to be off
EDC value will also missreport and the bottom field doesn't function

If bioses lack DF-C States, which this tool on autoruns, you can always disable them to prevent random reboots
@mongoled 8ghz example had DF-C states on, as it overboosted that high
* I publish it, because the lack of bios options from MSI and Gigabyte (closed down AMD CBS) starts to get annoying
Ugghhhh, ughhhhhh and ughhhhhh

"DF-C States" was/is set to disabled in the BIOS

:mad::mad:

And ive just ran the tool you kindly shared and look whats its showing and yes, its set to disabled in BIOS, no wonder I cant get consistent results when attempting to tweak PBO with telemetry, no consistency across reboots ......

2514504


AMD Gremlin's , many little gremlin's
Yes toooooooo many, I am on Eder modifed A.A1 BIOS that uses agesa 1.2.3.0a, maybe I am going to have to do a flash with a different utility than the default MSI flash tool .........
 
  • Rep+
Reactions: Veii

·
Registered
Joined
·
1,779 Posts
So I re-flashed with the AfuWin64 tool, "all blocks"

Still no change with setting "DF-CStates" to disabled as per "ZenStates 2.0.0" debug tool.

So looks like @Eder unhidden options are not taking effect ??

:/
 

·
Enthusiast
Joined
·
876 Posts

Attachments

·
Registered
Joined
·
582 Posts
I would have expected your score to be higher
Why all of a sudden? His score is well inline with others.

I wonder if I can somehow simulate an 80C CB23 run to see how much frequency drops by
You may expect roughly 100 pts down per 2°C+
Attached my 77°C run for the reference )) and earlier tests summary (winter BIOS).

May I ask you, what makes your cores run @ such high Vcore and ~ 15W / core ?
Is it some +global offset?
 

Attachments

·
Registered
B550 gaming edge, 5800x, 5600x, 3600xt, 1600AF, 2070S, 4x8 3200c14, 2tb adata 8200sx PRO, 500gb 970e
Joined
·
569 Posts
one last thing, you mentioned that turning off DF-CTATES in bios is best.
no doubt, but you also showed zen-states and a picture showing what to turn off there.
but once i restarted the pc it somehow has it turned back on again.
(as everyone else is showing)
C6 state is off, but C6 package stays on...



2514501
 

·
Registered
B550 gaming edge, 5800x, 5600x, 3600xt, 1600AF, 2070S, 4x8 3200c14, 2tb adata 8200sx PRO, 500gb 970e
Joined
·
569 Posts
Why all of a sudden? His score is well inline with others.


You may expect roughly 100 pts down per 2°C+
Attached my 77°C run for the reference )) and earlier tests summary (winter BIOS).

May I ask you, what makes your cores run @ such high Vcore and ~ 15W / core ?
Is it some +global offset?
2514522

almost 20 watts a core for me....?
(edit) i do have to readjust core 4 on the 9th iteration it had a fail (stated core 3 which is on positive offset) gonna try neg offset
but none the less...
 
9561 - 9580 of 10745 Posts
Top