AMD Ryzen 7 5800X3D Review: Better gaming CPU than Core i9-12900K for US$120 less

[Redaktion]

The AMD Ryzen 7 5800X3D is the company's first attempt at a mainstream desktop CPU that leverages 3D V-cache technology. 3D V-cache provides AMD with a shot in the arm in gaming enabling it to beat even the Intel Core i9-12900K without clock or TDP increases. However, its current implementation results in several deficiencies that narrow the Ryzen 7 5800X3D's use cases.

https://www.notebookcheck.net/AMD-Ryzen-7-5800X3D-Review-Better-gaming-CPU-than-Core-i9-12900K-for-US-120-less.616720.0.html

[shake]

The 5800x3d is a good gaming processor, but the 3090ti is roughly twice as powerful as the titan rtx.

[Vaidyanathan]

The 5800x3d is a good gaming processor, but the 3090ti is roughly twice as powerful as the titan rtx.

You're right. That was an unavoidable limitation. That's why I've limited the comparison to just 1080p at lower settings where the CPU matters more.

[Anonymousgg]

If Raptor Lake and Zen 4 are closer together than Alder Lake and Zen 3 were, then the 7800X3D/7950X3D will really shine. That's before the rumored improvements to the 3D cache are considered, e.g. +30% instead of +15%.

I just hope we see some L4 cache in the future, and mega APUs.

[Vaidyanathan]

If Raptor Lake and Zen 4 are closer together than Alder Lake and Zen 3 were, then the 7800X3D/7950X3D will really shine. That's before the rumored improvements to the 3D cache are considered, e.g. +30% instead of +15%.

I just hope we see some L4 cache in the future, and mega APUs.

I am actually super stoked for the Zen 4 X3D versions. That can potentially bridge the gap we see with current Zen 4 vs Raptor Lake benchmark leaks, at least in gaming.

[Jamie]

Your comparisons are completely baseless as you've not controlled any variable. It's actually comparing apples to oranges. This type of article and media is a waste of your time as well as the readers

[Beomagi]

More but not solely.
You should retest with the Titan on the 5800 and replace the tests with the 3090. These results are misleading.

[Joke]

This article and your tests are an absolute joke.

Pairing the chip with a 3090ti vs other chips paired with several years old gpu's with a fraction of the power. Complete and utter nonsense.

Anyone who thoroughly read this absolutely believes you intentionally gave the 5800 this handicap because you know full well it would have the pants beat off it in a fair controlled benchmark.

Dishonest, shameful, and downright insulting to tech journalism. You should be ashamed of yourself for this trash.

[NikoB]

I don't see any real evidence that Zen4 is better than the outdated Alder Lake. Everywhere a complete failure in both single-threaded performance and multi-threaded, with an insignificant difference in consumption for the desktop. AMD is no longer a cake. R.I.P. But the poor, of course, may like the idea that they are winning something there ...

It's been a long known fact - even when Intel lost in performance for multi-threaded mode (but never in single-threaded mode) - the cost of buying a motherboard of the same class + a comparable processor has always been in Intel's favor. What changed? Now AMD is a loser in everything except consumption. But is it really important for gamers? =)

[NikoB]

Rumor has it that AMD is desperately trying to jump into the last carriage of the outgoing train, trying to get TSMC into the "2"-"3" nm process as soon as possible. On what, as you all remember, Intel recently burned out, canceling all deals with TSMC until the end of 2023. But is it so important to Intel now if it is the pure king of performance, with a not very significant difference in price and power consumption?

It suddenly turns out that not Intel, but AMD is more and more clearly reminiscent of the "Titanic"...

railway carriage

[notebooknewuser]

What is the point in testing the 5800X3D with a 3090, then every other CPU with a Titan RTX?  Completely skewed the result, making every single one of your gaming 'benchmarks' irrelevant and pointless.

Who honestly thought that was a sensible idea?  I suggest in future if you want to increase website visitors, you stick to common sense testing methods (i.e use the same components apart from the one your benchmarking), you know...to make it actually fair and not biased??