3900x with higher frequency yet slower benchmark results - Overclock.net - An Overclocking Community

Forum Jump: 

3900x with higher frequency yet slower benchmark results

 
Thread Tools
post #1 of 3 (permalink) Old 10-24-2019, 12:15 PM - Thread Starter
New to Overclock.net
 
Join Date: Oct 2019
Posts: 11
Rep: 0
3900x with higher frequency yet slower benchmark results

Hi all,

As I mentioned in a another thread I currently have two 3900x's and one will be going back to amazon. The first 3900x has a idle of 33-43 during idle (fluctuates) on a NH-D15 and can get a fclk of 1900. The problem with this processor was that under a exteme load it would dip under 3.8ghz which according to AMD techsupport isnt normal so I decided instead of having a new chip RMA'ed through AMD i would just get amazon to send me a new one.

Well this new chip in initial results seemed better other than it refuses to work with a fclk of 1900 but will work with 1866. The temps are way lower with it operating at 27-37ish (not sure about the high end) on idle. And when under load the cores don't dip below 3.8 and overall seems to have more cores hitting higher clock rates compaired to the first.

With that said I bench-marked both chips with a fclk of 1866 to be fair to both and the first chip seems to beat the second one in every benchmark I have done. Probably not by alot but it still does. In games the performance difference is negligible but I also dont have a power house video card at the moment and I am working on a spare TV for now so low resolution.

In geek bench I am seeing a delta of about 200-300 points or so in multi core preformance. Sorry I dont have more concrete numbers right now but I didnt bring my results with me to work.

Anyone have any idea what could cause a seemingly higher clocking and cooler chip from beating the hotter lower clocking one? All the voltages other then SOC and vggd are on auto. My motherboard is a x570 Master with the F7c bios.

Just trying to rule out something I have done in error on my side. Other wise I guess I send back the second chip.

Thanks
BMDJag is offline  
Sponsored Links
Advertisement
 
post #2 of 3 (permalink) Old 10-24-2019, 12:40 PM
GPU Enthusiast
 
rv8000's Avatar
 
Join Date: Sep 2012
Posts: 2,506
Rep: 199 (Unique: 146)
Without average clocks from the benchmark runs (resetting hwinfo for the duration of the benchmark) for both CPU's you don't have a super accurate way to compare the previous CPU with the new one. The newer CPU may boost higher with the updated 1.0.0.3 ABBA based bios but it's average MC turbo may be overall lower than the previous CPU for dozens of reasons

As long as all the other bios settings remained the same between testing, comparing overall average boost clock under multithreaded loads would probably give you the best information. Now for a list of things that could be causing the score difference:

- chipset driver differences
- bios change
- mismatched mem clk and fclk
- different memory timings
- windows power plan settings
- silicon quality requiring higher voltage for X clock speeds causing more clock dips due to hitting boost limits
- various driver changes and or background programs

When all is said and done, a 200-300 points difference in Geekbench is a ~1-2% performance difference, seems a bit extreme to return the CPU over unless you are competitively benching. In that case you wouldn't even be running the cpu multi at auto so the stock performance with vsoc and vddg changed is moot.

Steins Gate
(16 items)
CPU
3700X
Motherboard
ASUS CH8 WiFi
GPU
EVGA 2080S FTW3
RAM
2x8GB DDR4 3200 c14 @ 3800 C16 1T
Hard Drive
960 Evo
Hard Drive
850 Evo
Power Supply
EVGA 750w G2
Cooling
EVGA CLC 360
Case
Fractal Design S2
Monitor
Acer ed323qur
Keyboard
CM MK730
Mouse
G703
Mousepad
Logitech Powerplay
Audio
Polk tsi 100
Audio
HD518
Audio
Denon PMA-60
▲ hide details ▲
rv8000 is offline  
post #3 of 3 (permalink) Old 10-24-2019, 12:50 PM - Thread Starter
New to Overclock.net
 
Join Date: Oct 2019
Posts: 11
Rep: 0
Quote: Originally Posted by rv8000 View Post
Without average clocks from the benchmark runs (resetting hwinfo for the duration of the benchmark) for both CPU's you don't have a super accurate way to compare the previous CPU with the new one. The newer CPU may boost higher with the updated 1.0.0.3 ABBA based bios but it's average MC turbo may be overall lower than the previous CPU for dozens of reasons

As long as all the other bios settings remained the same between testing, comparing overall average boost clock under multithreaded loads would probably give you the best information. Now for a list of things that could be causing the score difference:

- chipset driver differences
- bios change
- mismatched mem clk and fclk
- different memory timings
- windows power plan settings
- silicon quality requiring higher voltage for X clock speeds causing more clock dips due to hitting boost limits
- various driver changes and or background programs

When all is said and done, a 200-300 points difference in Geekbench is a ~1-2% performance difference, seems a bit extreme to return the CPU over unless you are competitively benching. In that case you wouldn't even be running the cpu multi at auto so the stock performance with vsoc and vddg changed is moot.
Hi,

I will use hwinfo and see if I can get a better idea of the average operating frequency. But when watching in ryzen master it seems to sustain its higher clocks. But maybe its dipping between frequency updates in Ryzen Master?

As for the other differences:

The system/software differences are not a factor as both chips are being swapped out and placed into the same system. They are the variables. I make sure I reset the bios between swaps and then load up a profile that is common to both of them when swapped. So the Ram frequencies and bus frequencies are all the same.

As for silicon quality, am I incorrect in assuming the chip that operates at a higher temp would be the one that requires more voltage? Also is there any other voltages I should keep a eye on while its operating other then the core voltage?

I am pretty sure I shut down a lot of the back ground processes but I will double check tonight.

Are there any recommended benchmarks I should run. I am running a few games (FarCry 5, Shadow of the Tomb Raider), Geek Bench 4/5, and using Prime 95 as a stress test.

Thanks again for the feedback.
BMDJag 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