• Now's your chance win big! Join our community and get entered to win a RTX 2060 GPU, plus more! Join here.

    Pi Cast Episode 3 streams live on Tuesday, August 4th at 2:30 pm ET (7:30 PM BST). Watch live right here!

    Catch Scharon on the Tom's Hardware Show live on Thursday, August 6th at 2:00 pm ET (7:00 PM BST). Click here!

Info Misinforming people about the Ryzen 3950x

lastguytom

Prominent
Nov 23, 2018
5
1
515
0
Why does your website misinform people about the Ryzen 3950x?
I ran the same tests on my Ryzen 3950x in Ashes of Singularity using the same setting you used and I got scores of 65.5 my setting was 1.3 volt 4.4ghz all core 16core/ SMT turn off.
I did the same test with 16 core, SMT on 4.3 GHz/ 1.3 volts to all cores got a score of 64.7 only .8 slower
The AMD web site has a post by my friend Micheal Nager on how to set up your Ryzen 3000 series processor
PLEASE correct your ARTICLE OR TAKE IT DOWN, IT'S NOT TRUE
 

DSzymborski

Champion
Moderator
Why does your website misinform people about the Ryzen 3950x?
I ran the same tests on my Ryzen 3950x in Ashes of Singularity using the same setting you used and I got scores of 65.5 my setting was 1.3 volt 4.4ghz all core 16core/ SMT turn off.
OK then, let's see evidence of your exact configuration and the actual data from your test. Screenshots and/or video, please, not your "word." I can say I got Windows 10 running on my Game Boy; without evidence there's no reason to take it seriously.
 

Rogue Leader

It's a trap!
Moderator
Why does your website misinform people about the Ryzen 3950x?
I ran the same tests on my Ryzen 3950x in Ashes of Singularity using the same setting you used and I got scores of 65.5 my setting was 1.3 volt 4.4ghz all core 16core/ SMT turn off.
I did the same test with 16 core, SMT on 4.3 GHz/ 1.3 volts to all cores got a score of 64.7 only .8 slower
The AMD web site has a post by my friend Micheal Nager on how to set up your Ryzen 3000 series processor
PLEASE correct your ARTICLE OR TAKE IT DOWN, IT'S NOT TRUE
Except the article IS true. You manually overclocked your 3950X. The test was done with stock settings and PBO on. The article was also done in November of 2019, between then and now there have been many BIOS improvements, new chipset drivers, optimizations and more.

Also 15fps difference we would need to see some significant telemetry proof to believe that difference.

Please feel free to come back with proof of your claims before lobbing accusations on the editorial team.
 

Phaaze88

Illustrious
Ambassador
Yep. And then if you get the ram speed too high, you cut the speed of the Infinity Fabric in half... :sweatsmile:
If the user undervolts too much, these cpus actually lose performance; I believe it's similar to X299's Phantom Throttling.
The user won't see the loss until performance benchmarks are run.

*sigh
What a mess Ryzen 3000 OC'ing has turned into... I'm frankly tired of those and the high idle temp threads.
 
Last edited:

wyliec2

Honorable
Apr 4, 2014
81
2
10,645
4
And yet again, why you don't manually overclock Ryzen 3000.
The cpu can tune itself better than you can.
I’ve made several runs at overclocking my 3950x (ASUS Crosshair VIII Formula, 32 GB g.skill Z Neo 3600 RAM, Kraken X62 AIO) and keep coming back to this same conclusion.

Through various settings I’ve gotten 9500+ score on Cinebench R20 but hitting mid-80 C temps drawing 220 watts. Aida64 stress testing is also mid-80’s. I might consider using these settings if it produced a meaningful difference in my heavy work which is processing Blu Rays and encoding to MKV. Running this task I also see fairly high temps.

In largely default settings (everything on Auto except with XMP enabled), I get 70-72 C temps with Cinebench and Aida 64. The Cinebench score drops to 9000 however the actual Blu Ray processing is slightly FASTER with the lower settings and the temps under 70C.

I’ve concluded that for my real-world work (YMMV), the CPU/MOBO combo I have does a better job managing voltages, power parameters and clock speed than my trying to fine tune the variable boost settings or applying an all-core overclock.

This isn’t my first rodeo having overclocked an I7-4770k (de-lidded) and I7-5960x in the past. Priority has always been stability and longevity.
 

Phaaze88

Illustrious
Ambassador
I’ve made several runs at overclocking my 3950x (ASUS Crosshair VIII Formula, 32 GB g.skill Z Neo 3600 RAM, Kraken X62 AIO) and keep coming back to this same conclusion.

Through various settings I’ve gotten 9500+ score on Cinebench R20 but hitting mid-80 C temps drawing 220 watts. Aida64 stress testing is also mid-80’s. I might consider using these settings if it produced a meaningful difference in my heavy work which is processing Blu Rays and encoding to MKV. Running this task I also see fairly high temps.

In largely default settings (everything on Auto except with XMP enabled), I get 70-72 C temps with Cinebench and Aida 64. The Cinebench score drops to 9000 however the actual Blu Ray processing is slightly FASTER with the lower settings and the temps under 70C.

I’ve concluded that for my real-world work (YMMV), the CPU/MOBO combo I have does a better job managing voltages, power parameters and clock speed than my trying to fine tune the variable boost settings or applying an all-core overclock.

This isn’t my first rodeo having overclocked an I7-4770k (de-lidded) and I7-5960x in the past. Priority has always been stability and longevity.
You may want to benchmark and compare performance on your board's stock settings and the different PBO levels. There are varying degrees to which the motherboards' violate AMD's guidelines for stock operation.
This is just the usual shenanigans from the motherboard vendors trying to one-up one another during reviews - it's the end user that gets the short stick; you've probably heard about some the complaints about high voltage and thermals? ~Yeah...

Asrock's X570 Taichi is pretty bad; the stock settings outperform most of the PBO profiles...

Asus appears to be following AMD's and Intel's guidelines the closest out of all the vendors, so it may not be that bad in your case.
 

ASK THE COMMUNITY

TRENDING THREADS