9-Way AMD Ryzen 7000 Series Benchmarks On Linux 6.5 / Ubuntu 23.10

[phoronix]

Phoronix: 9-Way AMD Ryzen 7000 Series Benchmarks On Linux 6.5 / Ubuntu 23.10

For those currently weighing between the (currently) nine different AMD Ryzen 7000 series processors for Linux use, here are fresh benchmarks of the Zen 4 desktop CPU line-up while testing with Ubuntu 23.10 and the Linux 6.5 kernel...

https://www.phoronix.com/news/AMD-Ryzen-7000-Linux-6.5

[skeevy420]

7800X3D vs 7700X are interesting since they take out all the CCD and mixed core factors. There's not a whole lot of performance gained over the X3D for a noticeable increase in power usage. However, it also shows that that IPC is still King in a lot of places. I don't regret getting the X3D. All things considered, at this level of performance I prefer the power savings and the cache benefits when they rear their heads.

Those results did make me wonder something:

Michael Have you ran the 7700X locked to 5.05GHz? Or any other X3D/Non-X3D at the X3D's frequency? I thought it be interesting to see how they perform, their power usage to performance ratio, when limited in frequency like the X3D.

I can't get openbenchmarking.org to load to look at anything else

[elbar]

Making Heat powerlines and return it back into Power when motherboard with CPU onboarded on it could have also connection of power to earth on the otherside of motherboard or better idea to leverage such heat space and maybe sometime return it back to hungry CPU... Heat is also radiation and therefore also electricity...

[Mark625]

Making Heat powerlines and return it back into Power when motherboard with CPU onboarded on it could have also connection of power to earth on the otherside of motherboard or better idea to leverage such heat space and maybe sometime return it back to hungry CPU... Heat is also radiation and therefore also electricity...

Are you trying to suggest turning the computer's waste heat back into electricity and using that to power the computer? If so, how? What method would be efficient enough and affordable enough to make any such effort cost effective?

Waste-energy recapture is not worthwhile even at industrial scale in most cases. In a home or office environment it would be even less so.

[uid313]

Wow, compare Ryzen 9 7900 against Ryzen 9 7950X, the latter is horribly energy inefficient!

The 7900 gets 78.95% of the performance at less than half the energy!

[skeevy420]

Wow, compare Ryzen 9 7900 against Ryzen 9 7950X, the latter is horribly energy inefficient!

The 7900 gets 78.95% of the performance at less than half the energy!

And that the X3Ds are where it's at in regards to power and performance. The 7950X has 1/3 or so more power usage for minuscule amounts of performance gain compared to the 7950X3D.

What's flippin crazy is the 7900X is cheaper than the 7900 at both Amazon and Newegg right now. That makes me wonder how it performs in a 65W mode.

[JanW]

Yes, the power consumption numbers are rather surprising (not new, mind you). It seems that the -X3D parts sip power. Are the power consumption numbers from an independent source (power meter at the wall minus some baseline platform load) or is this the motherboard reporting? If the latter, how does it compare to wall power, do the differences check out? Just wondering if there was some of the package power for the X3D that's not accounted for, since the differences are so stark... Apologies if this has been discussed, I do remember that results were similar in the first benchmarks when the 7950X3D came out.

[ddriver]

Amd wins!

3d parts run lower voltages and peak clocks for various reasons, so they are nearly on the perfect power / performance curve.

[loganj]

that 7600x sure draws too much power for its performance

[peterdk]

I would prefer a few more results in the article. The openbenchmarking is not readable at mobile form factor.