Precision Boost Overdrive How Much Over Max Boost You Have - Page 2 - Overclock.net - An Overclocking Community

Forum Jump: 

Precision Boost Overdrive How Much Over Max Boost You Have

Reply
 
Thread Tools
post #11 of 40 (permalink) Old 07-15-2019, 03:58 PM
New to Overclock.net
 
The Stilt's Avatar
 
Join Date: May 2013
Posts: 2,362
Rep: 801 (Unique: 304)
Quote: Originally Posted by AlphaC View Post
Have you tried this on MSI boards? OptimumTech is using a Godlike.


Your theory is that his undervolt dropped performance despite him not realizing it. Power consumption is most definitely down since he dropped 20°C.
No.
The feature isn't motherboard dependant, given were are talking about the same feature (PBO Scalar) and not e.g. full scale current option (cheat) which is available on some MSI boards.

It's not really a theory.
Most likely the CPU was running at 3.6GHz equivalent (stretched), since that's where the 1.00V voltage should allow it to run.
The Stilt is offline  
Sponsored Links
Advertisement
 
post #12 of 40 (permalink) Old 07-15-2019, 04:03 PM
⤷ αC
 
AlphaC's Avatar
 
Join Date: Sep 2012
Posts: 11,111
Rep: 904 (Unique: 590)
Someone should ask him for his Blender render time at 1.0V then.


edit: found this
https://www.legitreviews.com/amd-ryz...o-1-00v_213060
Quote:
I noticed a number of sites and channels are reporting that the AMD Ryzen 9 3900X 12-core, 24-thread processor can be run at full clock speeds with just 1.00V on the CPU VCore. That sounds pretty damn impressive and I just had to give it a try here to see if that is possible and what that means for performance.
The Ryzen 9 3900X CPU just happens to still be in the Gigabyte AORUS X570 Master motherboard, so I restarted our system and went straight to the CPU Vcore setting in UEFI F5g and took it off ‘AUTO’ and dropped it down manually to 1.20V, 1.10V and 1.00V to see what would happen.

And the final score is 5,942 points! Oh wait, that is way lower than the stock score if 7,105 that I got with stock ‘Auto’ voltage settings. Our performance dropped by just over 16% thanks to undervolting the processor. The power draw at the wall outlet dropped from 216 Watts down to 129 Watts. That means the system is using 40% less power and taking just a 16% hit to performance. That actually is impressive when you think about it! Undervolting the 3900X will allow you to put in small form factor system or use near silent thermal solutions a mini case as long as you are willing to give up some system performance.
Click image for larger version

Name:	3900x-1volt-cinebench.png
Views:	27
Size:	25.0 KB
ID:	280666


---------


edit2: also PBO testing by GamersNexus
https://www.gamersnexus.net/guides/3...hmarks-auto-oc

► Recommended GPU Projects: [email protected] , [email protected] (FP64) (AMD moreso) ► Other notable GPU projects: [email protected] (Nvidia), GPUGrid (Nvidia) ► Project list



Last edited by AlphaC; 07-15-2019 at 09:44 PM.
AlphaC is offline  
post #13 of 40 (permalink) Old 07-16-2019, 05:46 AM
New to Overclock.net
 
Join Date: Oct 2017
Posts: 94
Rep: 1 (Unique: 1)
Precision Boost vs Precision Boost OverDrive (PBO) vs Auto-OC

https://www.gamersnexus.net/guides/3...hmarks-auto-oc


Rough Summary, as I understood the article:

Precision Boost:
Default cpu behavior, designed/implimented by AMD, where one or more cores boosts voltage and clockspeed depending on a combination of Temp limit/Current limit/Power limit. Precision Boost generation 2 is used for the Ryzen 3000 series, to differentiate from Precision Boost generation 1 seen in threadripper.


Precision Boost OverDrive (PBO):
Extra Overclocking feature provided by Mobotherboards, is considered overclocking and thus warranty voiding. Primarily raises the power, current, and temp limits with little to no real world effects at above ambient temps.

Auto-OC:
Extra Overclocking feature provided by Mobotherboards, is considered overclocking and thus warranty voiding. Raises the max POTENTIAL boost for single or all core, CPU clockspeed is still bound by the same limiting factors for Precision Boost.

Last edited by BeeDeeEff; 07-16-2019 at 05:51 AM.
BeeDeeEff is offline  
Sponsored Links
Advertisement
 
post #14 of 40 (permalink) Old 07-16-2019, 05:52 AM
New to Overclock.net
 
Duvar's Avatar
 
Join Date: Jul 2013
Posts: 121
Rep: 3 (Unique: 3)
Duvar is offline  
post #15 of 40 (permalink) Old 07-16-2019, 07:18 AM
New to Overclock.net
 
Join Date: Jan 2019
Posts: 81
Rep: 0
Quote: Originally Posted by AlphaC View Post
People are struggling just to get 4.6GHz on the R9 3900X

I think the +200MHz PBO is more of a limit than what is expected clocks

On Zen+ PBO is the limit (i.e. it just allows the cores to turbo to 4.35 on R7 2700X, thermal and EDC/PPT dependent).
This is true, but on 2700X 4.35Ghz all core is achievable quite easily (I have been running that for months).
With the 3900X however, I have yet to see a reviewer get it on a single core let alone an all core boost.
constructorx is offline  
post #16 of 40 (permalink) Old 07-16-2019, 07:59 AM
New to Overclock.net
 
rdr09's Avatar
 
Join Date: Mar 2011
Location: From the US but lives in Africa
Posts: 18,449
Rep: 862 (Unique: 630)
Quote: Originally Posted by constructorx View Post
This is true, but on 2700X 4.35Ghz all core is achievable quite easily (I have been running that for months).
With the 3900X however, I have yet to see a reviewer get it on a single core let alone an all core boost.
check it out . . .

https://www.overclock.net/forum/10-a...l-oc-x370.html

Not the kind of reviewer we are familiar with.

[email protected] 6 http://valid.canardpc.com/show_oc.php?id=2211392 4.6 @ 4 http://valid.canardpc.com/show_oc.php?id=2216580
5.0 @ 8 http://valid.canardpc.com/show_oc.php?id=2511322
2nd AMD Build
(10 items)
CPU
2700
Motherboard
X470
GPU
290
RAM
3200 CL14
Hard Drive
1000
Power Supply
700
Case
212
Operating System
10/64
Monitor
40 1080
Keyboard
M100
▲ hide details ▲
rdr09 is offline  
post #17 of 40 (permalink) Old 07-16-2019, 08:58 AM
New to Overclock.net
 
Join Date: Jan 2019
Posts: 81
Rep: 0
Quote: Originally Posted by rdr09 View Post
check it out . . .

https://www.overclock.net/forum/10-a...l-oc-x370.html

Not the kind of reviewer we are familiar with.
I would like to know just how long these boosts to 4.6Ghz are? A millisecond boost that registers on HWMonitor is not convincing.
A way to estimate this would be as follows.

1) Find the max all core stable clocks get a benchmark score
2) Enable PBO and get another benchmark score
3) Calculate the delta

For instance:

The Cinebench score for a 2700X stock is about 1828 (https://www.guru3d.com/articles-page...-review,9.html).
Say the average frequency was 4-4.1Ghz (say 4.0Ghz average) then we could do this calculation.

1828 / 4.0 = 457

My 2700X PBO Cinebench score (yesterday) was 1978.
My frequencies were observed to be around 4.3Ghz during the run, so as an estimate:

4.3 * 457 = 1965
4.3275 * 457 = 1977.6

So a good estimate of my all core clocks would be 4.3275GHz during the whole run, not just a spike for a fraction of a second.

This can show just how much the PBO is giving in terms of performance gain and not just a spike that is registered on a 'Max' frequency table. This is what I mean.
constructorx is offline  
post #18 of 40 (permalink) Old 07-17-2019, 08:45 AM
Meddling user
 
gupsterg's Avatar
 
Join Date: Jan 2015
Location: Lurking over a keyboard
Posts: 7,085
Rep: 768 (Unique: 351)
Quote: Originally Posted by The Stilt View Post
Scalar does nothing on Matisse.
You can change it to 0.5x (not from the bios thou) and there is no change in the behavior.
I have to disagree respectfully Roger, of course this is based only on my sample size of 1.

I can only crudely show that changing scalar within in AMD section of UEFI works for me on UEFI 2406 & UEFI 0068 of C7H on R5 3600.

UEFI 2406 has no PBO menu on Extreme Tweaker, so only option is AMD menu. UEFI 0068 has PBO on Extreme Tweaker, but does not function.

See WMV in this ZIP, only scalar is changed, as I had FastStone Capture running the CPU would not go to idle like it does without, which presents a more clearer picture.
gupsterg is offline  
post #19 of 40 (permalink) Old 07-17-2019, 11:00 AM
New to Overclock.net
 
The Stilt's Avatar
 
Join Date: May 2013
Posts: 2,362
Rep: 801 (Unique: 304)
Quote: Originally Posted by gupsterg View Post
I have to disagree respectfully Roger, of course this is based only on my sample size of 1.

I can only crudely show that changing scalar within in AMD section of UEFI works for me on UEFI 2406 & UEFI 0068 of C7H on R5 3600.

UEFI 2406 has no PBO menu on Extreme Tweaker, so only option is AMD menu. UEFI 0068 has PBO on Extreme Tweaker, but does not function.

See WMV in this ZIP, only scalar is changed, as I had FastStone Capture running the CPU would not go to idle like it does without, which presents a more clearer picture.
Obviously you are entitled to have an opinion
Note that the scenario below is power limited (PPT), but I've tested adjusting the scalar value under several different conditions.
Even when the CPU is completely unlimited in terms of PPT, TDC and EDC, increasing PBO Scalar (FIT ceiling) does not increase the allowed voltage and therefore the maximum achievable frequency.
For example on Pinnacle Ridge, increasing the PBO scalar allowed significantly higher voltage in all conditions. On Matisse the allowed voltage is the same, regardless if the PBO is either 1x or 10x.
Because of the voltage limit, this CPU won't boost higher than 4.175GHz on all cores, regardless of the removal of all limits and despite the fact that the tDie is with well within the spec.

Maybe it depends on the silicon characteristics, but that's the case on this 3900X specimen at least.
IIRC I also checked it on the 3700X and the results were the same.

Spoiler!
The Stilt is offline  
post #20 of 40 (permalink) Old 07-17-2019, 11:28 AM
Meddling user
 
gupsterg's Avatar
 
Join Date: Jan 2015
Location: Lurking over a keyboard
Posts: 7,085
Rep: 768 (Unique: 351)
As you know, avid fan of your shares and search on differing forums for your posts, so thank you for engaging further .

I do not doubt what you see , or know , I am showing what I see and trying to gain understanding . Maybe it is just my CPU, maybe as R5 3600, dunno, but wanted to share what I see .

I have done retake, pretty much same setup, but VDDG is 0.950 and VDIMM 1.355V. This time I did screen capture at end with more length, 3rd run still going but defo not gonna reach what 6x 2x does.

ZIP with 6x 2x, link. I think the 2x over longer time is showing similar frequency probably due to clock stretch or something else, defo not getting same average VCORE SVI2 TFN as 6x. The shorter RT runs show better difference between scalar change IMO. 4x is pretty much same as 6x for me, I use 6x just to give it room (in mind) for another load, leaving [Auto] scalar does not gain me same average MHz/voltage.

Sample of 3rd run 0x = scalar is [Auto]

Spoiler!


Without changing scalar, past PBO+100MHz, I do not gain improved average ACB. As I tweak scalar from say 1x upto 4x I see improved average MHz and slightly higher average voltage.

*** edit ***

0x ie scalar auto never reached clocks or voltage, near end only when CPU go from loaded to idle did it peak to PBO+150MHz.

WMV link.

Last edited by gupsterg; 07-17-2019 at 11:53 AM.
gupsterg is offline  
Reply

Quick Reply
Message:
Options

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



Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)
 
Thread Tools
Show Printable Version Show Printable Version
Email this Page Email this Page


Forum Jump: 

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
Trackbacks are Off
Pingbacks are Off
Refbacks are Off