[Official] Ryzen 2700 and 2700X Owners Club - Page 195 - Overclock.net - An Overclocking Community
Forum Jump: 

[Official] Ryzen 2700 and 2700X Owners Club

Reply
 
Thread Tools
post #1941 of 1946 (permalink) Old 02-06-2020, 04:40 PM - Thread Starter
Tactical Moron
 
Minotaurtoo's Avatar
 
Join Date: Jun 2013
Location: Florence, AL
Posts: 2,435
Rep: 81 (Unique: 47)
I too used that trick lol... I have a "curing" type thermal paste that you really have to warm up to get it to cure, and to remove the cooler either takes a small pry bar or a really warm cpu...

I've swapped to a 3900x.... I was able to get all core 4.3ghz at only 1.256v...eventually failed a prime95 test, but still the improvement frequency/voltage is real... fully stable at 1.23v for 4.2ghz all core and uses less power than stock settings.

Offical Ryzen 9 3900x Owner!


Minotaurtoo is offline  
Sponsored Links
Advertisement
 
post #1942 of 1946 (permalink) Old 02-11-2020, 07:34 AM
New to Overclock.net
 
chevy350's Avatar
 
Join Date: May 2012
Location: Ohio
Posts: 49
Rep: 1 (Unique: 1)
I've noticed these values and am thinking they would be caused by c-states?? I have gd and pd disabled for memory, but I believe low-power idle is enabled. Anyone seens HWInfo reporting weird readings like this lol
Attached Thumbnails
Click image for larger version

Name:	Capture045.jpg
Views:	17
Size:	110.8 KB
ID:	326056  


Just my toy
(13 items)
CPU
2700
Motherboard
Asus ROG Strix X470-F
GPU
XFX GTS Black Edition RX580
RAM
Samsung
Hard Drive
Samsung 950 Pro
Power Supply
Seasonic Platinum 860W
Cooling
NZXT Kraken X62
Case
Corsair 750D
Operating System
Windows 10 Pro
Monitor
Acer 24"
Monitor
Samsung 24"
Keyboard
Razer
Mouse
J-Tech Vertical Mouse
▲ hide details ▲
chevy350 is offline  
post #1943 of 1946 (permalink) Old 02-11-2020, 09:33 AM
New to Overclock.net
 
cssorkinman's Avatar
 
Join Date: Apr 2009
Posts: 9,777
Rep: 491 (Unique: 276)
Quote: Originally Posted by chevy350 View Post
I've noticed these values and am thinking they would be caused by c-states?? I have gd and pd disabled for memory, but I believe low-power idle is enabled. Anyone seens HWInfo reporting weird readings like this lol

I usually only get those kind of errors if the machine goes into some kind of power saving mode.


cssorkinman is offline  
Sponsored Links
Advertisement
 
post #1944 of 1946 (permalink) Old 02-12-2020, 02:53 PM
New to Overclock.net
 
hazium233's Avatar
 
Join Date: Jul 2017
Posts: 99
Rep: 0
Quote: Originally Posted by deepor View Post

PPT = 141
TDC = 95
EDC = 140

That's the values that are used at stock settings, when PBO is disabled. With these defaults, the "EDC" value is the one that limits everything. If you increase it, the CPU will start boosting higher. After increasing it enough, you will then start getting limited by TDC instead (if your cooling is strong enough).

What I like to do is increase EDC, then lower TDC to limit things to something where my cooler and normal fan speed setup can still keep things below 80°C.

You can use AMD's "Ryzen Master" software to check on the EDC, TDC, PPT limits. That Ryzen Master tool will show how close the CPU is to the limits.

There's also another limit for the temperature. When boosting, the CPU will try to stay below 85°C. It will limit the boost when you hit that temperature. That's why when your goal is to find the limit you want to use for EDC and TDC, make sure you are still below 85°C while you experiment with stress tests. This 85°C temperature can also be customized somewhere in the BIOS if you want, on my board it's one of the options on the "NBIO" page of the "AMD CBS" section.
Thanks for this info. I still haven't altered my core parameters off of default, and I had wondered about the behavior. Does temperature for Tdie (or Tctl) affect the frequency below 85C though? Or is it only indirectly by somehow affecting EDC or TDC?

I ask because from Prism to TS140P it seems that I was able to boost Cinebench R20 by 39 points all core. Granted, that is only about 1%, but even on the Prism it seemed an aggressive curve and cooler ambient was necessary for my best scores.

COL1
(3 items)
CPU
Ryzen 7 2700X
Motherboard
Asus Strix X370-F (4012)
RAM
Ballistix Sport AT 2666 (D9TZV)
▲ hide details ▲
hazium233 is offline  
post #1945 of 1946 (permalink) Old 02-12-2020, 05:22 PM
New to Overclock.net
 
deepor's Avatar
 
Join Date: Feb 2013
Posts: 4,896
Rep: 483 (Unique: 328)
Quote: Originally Posted by hazium233 View Post
Thanks for this info. I still haven't altered my core parameters off of default, and I had wondered about the behavior. Does temperature for Tdie (or Tctl) affect the frequency below 85C though? Or is it only indirectly by somehow affecting EDC or TDC?

I ask because from Prism to TS140P it seems that I was able to boost Cinebench R20 by 39 points all core. Granted, that is only about 1%, but even on the Prism it seemed an aggressive curve and cooler ambient was necessary for my best scores.

In my experience, at default settings, only the "EDC" value will matter. This will be the only limit the CPU will hit. The default EDC value makes it so the voltage will be very low when the CPU is under load. The power usage is always far away from the PPT and TDC limits. This then makes it so the temperature will not matter if it's below 85°C.

Where the temperature matters is if you hit the "PPT" limit. The power usage of the CPU will change with the temperature. It's pretty interesting to see what happens there. In HWINFO you can double-click on an entry in the sensor window, and it will then open a small history graph window for that sensor. If you track power usage like that while running some stress test and the CPU uses some constant voltage, you can see that the power usage goes up and down when you play around with fan speeds and the temperature changes. If you are then the PPT limit, the CPU will try to keep the power usage constant. It will reduce MHz and voltage to do that.

The "TDC" limit is in practice very similar to "PPT", you can use either one of those two to limit things to a certain power usage. But with TDC, the temperature will not influence things. When the CPU gets hotter, the power usage goes up.
deepor is offline  
post #1946 of 1946 (permalink) Old 02-12-2020, 09:02 PM
New to Overclock.net
 
hazium233's Avatar
 
Join Date: Jul 2017
Posts: 99
Rep: 0
Quote: Originally Posted by deepor View Post
In my experience, at default settings, only the "EDC" value will matter. This will be the only limit the CPU will hit. The default EDC value makes it so the voltage will be very low when the CPU is under load. The power usage is always far away from the PPT and TDC limits. This then makes it so the temperature will not matter if it's below 85°C.

Where the temperature matters is if you hit the "PPT" limit. The power usage of the CPU will change with the temperature. It's pretty interesting to see what happens there. In HWINFO you can double-click on an entry in the sensor window, and it will then open a small history graph window for that sensor. If you track power usage like that while running some stress test and the CPU uses some constant voltage, you can see that the power usage goes up and down when you play around with fan speeds and the temperature changes. If you are then the PPT limit, the CPU will try to keep the power usage constant. It will reduce MHz and voltage to do that.

The "TDC" limit is in practice very similar to "PPT", you can use either one of those two to limit things to a certain power usage. But with TDC, the temperature will not influence things. When the CPU gets hotter, the power usage goes up.
I think since I had not accepted the PBO warning in bios, I couldn't view those in hwinfo yet. Either that or it might not work for the X370-F, I guess I will find out when I get around to that.

Highest power usage I have recorded was 136.92W, that was running Realbench under the Prism. I have not done a whole lot of small FFT, but for the time I ran that I only saw 130W or so, the different sizes seemed to have slightly different wattage. Voltage does seem to get pretty low in heavy loads, I saw down to 1.212V in small FFT. These were all just using default 2s polling time, so the real peak or nadir may have potentially been missed.

COL1
(3 items)
CPU
Ryzen 7 2700X
Motherboard
Asus Strix X370-F (4012)
RAM
Ballistix Sport AT 2666 (D9TZV)
▲ hide details ▲
hazium233 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: 2 (1 members and 1 guests)
Valnjes
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