AMD Ryzen 7 9800X3D

[Tinek]

I've not seen Linus' vid on it, but I've seen similar testing and the average framerates don't change much... But the 1% and 0.1% lows are vastly improved; they matter more than some folk seem to think.

Does Linus' vid not support that?

It seems, in his video, they don't see any significant difference in 1% lows in actual gaming resolutions (1440p and higher with RTX 4090) between 5800x3D, 7800x3D and newest 9800x3D. Either of these 3 is just fine and within at most few FPS difference from each other. Edge cases happen where newest is considerably better, but these are just 1080p edge cases. In productivity it's still mostly below 9700x. In other words, it's an expensive fastest gaming CPU that is really not needed for most people that play in 1440p or higher - much better option to buy faster GPU than invest into this CPU, as 9700x is considerably cheaper and good enough in games, with dropped price.

 

[Loxa]

It is an awesome chip the 9800x3d but i like what Linus has done in his review, where he has shown it at 1440p with high graphical settings and there is next to no difference as usual from a 5800x3d to the 9800x3d in gaming.. i think a lot of people think they are going to see a jump at 1440p and 4k and be disappointed.

That's quite interesting

 

[Tinek]

That's quite interesting

It's mostly that current modern CPUs and all the 3D cache ones, are fast enough when new games hammer GPUs mostly (yes, for RT fast CPU still needed but no need to get the fastest one there is). I can also see in tests that overclocking these CPUs hardly changes anything even in 1080p (2-3% for 20% more power use is a big NO to me).

 

[otester]

Is it worth upgrading from a 5900x to 7800x3d?

I also currently have CPB/PBO disabled due to WHEA errors causing random reboots.

 

[Sedgey123]

I've not seen Linus' vid on it, but I've seen similar testing and the average framerates don't change much... But the 1% and 0.1% lows are vastly improved; they matter more than some folk seem to think.

Does Linus' vid not support that?

Nope, although it is a limited amount of games but at 1440p and 4K the 1% lows are near identical

 

[Grim5]

For my fellow 4k native gamers;

Now show the 1% lows, don't try to hide them with your trickology

 

[Grim5]

She's a good overclocker, 5.75ghz all core with 1.27v 90watts while gaming, 60c with the stock IHS, 45c after delid. Overclock improves scores by 5-10% at 1080p

 

[benjii]

I'll definitely be getting one. I just ordered an MSI MAG X870 Tomahawk WIFI, 32GB Corsair Vengeance 6000mhz cl30 and a 2TB Samsung 990 Pro. It'll be a nice upgrade from my 5900X setup.

 

[Hotwired]

Excuse my ignorance, but why test at 1080p with this CPU?
At this price point and with the other highend (ish) hardware that most will have alongside it, what's the point?

It's an extremely abused form of testing where games are the programs used to benchmark cpus.

Thing is, once the graphical load becomes significant then the cpu becomes almost irrelevant, every modern cpu will be twiddling its thumbs waiting for the graphics card to draw the frames and thus all get the same frames per second.

It is benchmarking for the sake of benchmarking and it won't get changed because as above, it has to be low res, low graphics or you can't see the differences.

But you know what, for the people who are actually using 1080p and low graphics, it's a true reflection of the gaming performance they'll get... with a 9800X3D and a 4090...

 

[Boomstick777]

Looks amazing! I'm going to try my best to hold off until the 9900X3D launches..

RIP Intel.

 

[Poneros]

The performance differences are MASSIVE in many cases (20-30%+ in cases where you're already <100 fps so performance is very much still needed) vs even 7800X3D/14900K, and there's genuine tuning room there for OC, + Direct Die is looking very juicy (check derbauer's vid). Lots of cope in here around resolution, as if you'd be rendering with 8.3M pixels (4K) and <60 fps instead of cranking up DLSS and enjoying >120 - LOL, gimme a break.

 

[Monkeyboysteve]

I've not seen Linus' vid on it, but I've seen similar testing and the average framerates don't change much... But the 1% and 0.1% lows are vastly improved; they matter more than some folk seem to think.

Does Linus' vid not support that?

Most important factor to me when gaming at 4k... the occasional stutter can ruin some games imo.

 

[CaptainRAVE]

For my fellow 4k native gamers;

Looks like the 7800x3d will keep me going for a bit then.

 

[xaldub]

Is it worth upgrading from a 5900x to 7800x3d?

I also currently have CPB/PBO disabled due to WHEA errors causing random reboots.

For gaming yes, for productivity not so much.

 

[Rroff]

Now show the 1% lows, don't try to hide them with your trickology

I linked to where they test minimums in post 406: https://www.techpowerup.com/review/amd-ryzen-7-9800x3d/21.html

Though it is a little more varied story than that graph represents as some games show more of a difference but still.

 

[Dicehunter]

Nice to see progress but playing at 3440x1440, the uplift and cash outlay coming from a 7800X3D isn't worth it to me.

 

[Aretak]

Very impressive results, but I still can't really justify upgrading my 5800X3D. Especially when I'm looking to move to a 4K monitor soon and don't really play any new AAA games.

 

[keyser van someone]

The power and temps don't add up on TechPowerup.
Gaming power and gaming temps:
7800X3D 46W 69.2C
9800X3D 65W 60.8C

"Temperature testing on this page uses air-cooling, for consistency and to show comparable results. All performance testing on the other pages is done using an Arctic Cooling Liquid Freezer III liquid cooling solution that keeps temperatures well below all limits, to ensure there is no thermal throttling."

edit: ok LTT review explains, more power draw but better temps. I have some reading/watching to do

 

[Tetras]

The power and temps don't add up on TechPowerup.

What's the problem with their results?

 

[Stu999]

Looks like it’s the absolute boss for gaming. RIP any chance of getting one this week for a lot of people.