• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.
  • The forums have been upgraded with support for dark mode. By default it will follow the setting on your system/browser. You may override it by scrolling to the end of the page and clicking the gears icon.

Share your CPUZ Benchmarks!

https://valid.x86.fr/n9bhnv
CPU-z 24-7.png


I haven't ran this in a while, but did Saturday as I was just tweaking my UV a little on my 24/7 OC and ran a bunch of testing so I had a recent run at this. I've been playing with it since I updated my BIOS last time (which updated the micro-code, which boosted voltages...). I have 1-4 core loads run 6.0GHz, then 5-8 are 5.8, 5.7, 5.6, 5.5GHz in order, e-cores at 4.6GHz. When I first set the system up (old micro-code) I had done some testing at 5.6Ghz and 5.7GHz all-core and the gaming performance just wasn't worth the increased heat and power at all, so I settled on boosting e-cores and boosting a few cores higher as that gave me the most gains in games. One of my tests back then, pushing all-core a little harder: https://valid.x86.fr/ju16u8 (Multi-thread: 17883, Single-thread: 972).
 
1695770838424.png

CPU-Z.png
CPU-Z Multi Threads.png
 
Last edited:
Here are mine, wrapping up my new build

Dual AMD Epyc 7773x 128/256
1TB Ram
Dual 4090 FE RTX
64TB raid array
1600W psu

Silent build, compact, small form factor, all air cooled.
amd idle temps 27C
amd load temps 44C

2x 4090 FE temps idle 29C, Load 58C
 

Attachments

  • cpuz res 0.jpg
    cpuz res 0.jpg
    101.4 KB · Views: 246
  • cpuz res 1.jpg
    cpuz res 1.jpg
    101.6 KB · Views: 224
  • cpuz res.jpg
    cpuz res.jpg
    70.5 KB · Views: 274
https://valid.x86.fr/n9bhnv

I haven't ran this in a while, but did Saturday as I was just tweaking my UV a little on my 24/7 OC and ran a bunch of testing so I had a recent run at this. I've been playing with it since I updated my BIOS last time (which updated the micro-code, which boosted voltages...). I have 1-4 core loads run 6.0GHz, then 5-8 are 5.8, 5.7, 5.6, 5.5GHz in order, e-cores at 4.6GHz. When I first set the system up (old micro-code) I had done some testing at 5.6Ghz and 5.7GHz all-core and the gaming performance just wasn't worth the increased heat and power at all, so I settled on boosting e-cores and boosting a few cores higher as that gave me the most gains in games. One of my tests back then, pushing all-core a little harder: https://valid.x86.fr/ju16u8 (Multi-thread: 17883, Single-thread: 972).

@Bobaganoosh
Nice.. Your post gave me the itch to tweak things. What voltage and mode are you running? I'm at 1.305v, Override/Fixed, LLC6 on an MSI Z690 Edge WiFi DDR4.

I'm stuck at 59x1, 56x5, 55x8, with e-cores all on 44x. Ring is Auto. I can't get more than half of the e-cores on 45x or even p-cores set to 56x6 before I start throwing the occasional WHEA errors with HWInfo. Running per the above seems to be OK and error free after an hour of looping multi-core and single core in CB23 and simply letting the computer run all day. I'll try 2x59 or 1x60 at some point in the near future once I'm confident the above is stable.


capture.jpg
 
Updated results with Buildzoid Timings for Ryzen 7000 series:

First time cracking 800pts for Single-Thread :)
CPU-Z Buildzoid timings new.png
 
@Bobaganoosh
Nice.. Your post gave me the itch to tweak things. What voltage and mode are you running? I'm at 1.305v, Override/Fixed, LLC6 on an MSI Z690 Edge WiFi DDR4.

I'm stuck at 59x1, 56x5, 55x8, with e-cores all on 44x. Ring is Auto. I can't get more than half of the e-cores on 45x or even p-cores set to 56x6 before I start throwing the occasional WHEA errors with HWInfo. Running per the above seems to be OK and error free after an hour of looping multi-core and single core in CB23 and simply letting the computer run all day. I'll try 2x59 or 1x60 at some point in the near future once I'm confident the above is stable.


View attachment 316137
Override will help you keep power low for all core, but limit your peak per-core OC. I have an evga z690 Classified running in adaptive voltage with -40mV on the 5400 and 5700 V/F points (the rest at 0). It ends up keeping the voltage around 1.295-1.3 at 5.5GHz all-core load, but lets individual cores go up to default (~1.395ish I think) when they're boosting to 6GHz. I have VDroop set to - 75% (minimum droop), FLL override to mode 1 and PWM frequency at 600. Some of that is based on guides Luumi did for these EVGA boards. I've been thinking about doing some comparisons between my current settings and dropping e-cores to 4.5 with an all-p-core 5.6GHz. Not sure what the power/heat/performance trade there will be.
 
Updated results with Buildzoid Timings for Ryzen 7000 series:

First time cracking 800pts for Single-Thread :)
View attachment 316151

Well done! How can one achieve more than 800 single-threaded (ST) performance points with this CPU, whats the secret?
 
New bios for the W790 SAGE now lets tile Xeons go above 5.2GHz without BCLK adjustments, running 5.4GHz 2 core, 5.3GHz on 2 and 5.2GHz on the other 12

Screenshot 2023-10-18 213037.png
 
  • Like
Reactions: N/A
I keep hoping it will get faster.. :D

Good to see the old girl still has a little pull left in her :cool:

cpuz.png
 
My 5700G at 45w Boost OFF on DeskMini with 2x8gb BallistiX 2666 (XMP 3200 not enabled) :
! Everyday usage settings !
1699257062470.png



5700G at 65w Boost ON on DeskMini 2x8gb BallistiX XMP 3200 :
! Set only for BENCH !
1699257321246.png


2.08 x64 (updated) :

 
Last edited:
Yeah but this is the case with many benchmarks that don't take into account "cache" which is MUCH faster than system ram.

If you want to measure "latency" it appears PYprime is really good at this. MUCH better than Aida64 for example.

With the SAME 2X16GB DDR5 6200MT/s CL30 same timings / 2200FCLK my 7900X hit 57.2 ns is Aida64 while my 7800X3D is at 62ns.
In PYprime (2B) my 7900X hit 9.5 seconds vs. 8.0 seconds for my 7800X3D.

Haven't tested many games but in Destiny 2 I've seen a 30-40FPS improvement over my 7900X with 7800X3D. If you just looked at common "Benchmarks" one would assume the 7900X would be faster..
"Cache is King". I think the next consoles should definitely utilize 3D V Cache due to power / efficiency. Just imagine if they had CPU's ~ 5800X3D..
 
Last edited:
it appears PYprime is really good at this. MUCH better than Aida64 for example
I haven’t run it for awhile but I have 10.5 on the bot, might have to play around with it some more :)
 
First try, new rig presently in the build process, default cpu settings, ddr5-7200 CL34 XMP on 14700k... oh and no errors 2 passes memtest86.... bring the hate X3D gamerz :rockout:

https://valid.x86.fr/nvvhq5
1700275796378.png
1700275803849.png

1700279081366.png
 
Last edited:
CPU-Z Validator.png
 
Last edited:
Back
Top