First Game Test With the Ryzen 7 5800X3D Appears as Promised

[SL2]

Cherry picked game or not, what about the different GPU's? If the 3D had a 3090 Ti as well, those number should have been even higher?

I don't believe this for a second, because if true, the 3D would cost more. Neither AMD nor Intel gives something extra for nothing like that..

..well, unless there are very few CPU's available, like so many have suggested, then price can be low like this.

 

[PerfectWave]

"Well, it gets an average FPS number of 231, which is a bit odd, since it's not an average of around say 230 FPS, which would've made sense based on how the Intel numbers were reported." WHATTTTTT??????????

 

[Warigator]

I wrote that the 12900KS is a bad joke. 5800X3D is 15% faster in games and $280 cheaper! And that's in April, because in August or September they will release Ryzen 7800X, which will be in games 15-20% faster than 5800X3D. Raptor Lake needs to be 40% faster in games than Alder Lake, which is doable in my opinion, but very difficult. Over are times of stagnation and same architectures for years.

 

[Pumper]

Why wouldn't you use the in-game benchmark, which provides easily comparable results?

The answer is in the question.

 

[DeathtoGnomes]

clickbaiting at its finest. "oh lookie here, performance tests!'

 

[Punkenjoy]

The in game benchmark does not have enough CPU load to be a relevant CPU tests. It still depend on the game, but if you take section of the game and you run it for long enough, the randomness should be flatten and you would get a better idea of the real load.

As for this test, they tested in 720p low, but with max shadows and ultra details. Theses settings have a good impact on CPU load.

But 720p test are not as synthetic as something like 3dmark, but they are not that useful as the mode you will end up using. I find it always best to get the information. The 1% low are probably a more important metrics than the average frame rate for CPU. You don't need a 5800x3d to play game at 4k, but in some instance, you can still see huge comfort gain over an old CPU while still not having huge gain on average FPS.

CPU are used to run the OS, to load game and decompress assets (for now at least). Having a faster CPU really improve the comfort and can reduce the stutter someone might occasionally get. (or like other have pointed out, speed up games where the simulations is the bottleneck)

And it's not at all about numbers of cores. If thoses numbers are true, then this cpu with 8 somehow slower cores is kicking at to a 8P+8E/24thread cpu and a 16cores/32 thread cpu. while being clocked lower.

 

[mb194dc]

Until you play a properly CPU limited game like Stellaris, CK3, Cities Skylines, Civ 6, Old Wolrd, Factorio etc etc.

Then you can be getting 2,000 fps at 4k Ultra but slow turn times or slow tic rates make them a pain to play on thr largest maps which totally changes the game play experience.

True re strategy games where CPU is not just on rendering.

I doubt the 3d cache will help with that, probably regular 5800x maxed crushes the 3d version in those games.

I'd like to see proper tests with 1% low on max OC 4790k and later generations in typical gamer workloads. Especially at 4k.

We won't get that though as the results probably upset AMD and Intel too much!

 

[ThrashZone]

clickbaiting at its finest. "oh lookie here, performance tests!'

Hi,
That was the geekbench test

 

[TheLostSwede]

"Well, it gets an average FPS number of 231, which is a bit odd, since it's not an average of around say 230 FPS, which would've made sense based on how the Intel numbers were reported." WHATTTTTT??????????

Sorry, poorly worded and fixed.

 

[ThrashZone]

I don't get the 720 thing though.

 

[Punkenjoy]

True re strategy games where CPU is not just on rendering.

I doubt the 3d cache will help with that, probably regular 5800x maxed crushes the 3d version in those games.

I'd like to see proper tests with 1% low on max OC 4790k and later generations in typical gamer workloads. Especially at 4k.

We won't get that though as the results probably upset AMD and Intel too much!

One of the reason these games get slow is the dataset that become too large to be handled quickly in the CPU. In stelaris, you can cleary see that after some point, the performance just crash. It's quite possible that a larger cache will push a little bit further where the performance would tank but it remain to be seens, i really hope they will test thoses kinds of games to see if major improvements are coming with larger cache.

 

[TheLostSwede]

clickbaiting at its finest. "oh lookie here, performance tests!'

I specifically wrote test, not tests.

 

[HD64G]

Since AMD's marketing slides show 5800X3D winning by 10% over 12900K in this game, it is logical for the gap to be bigger at 720P. Only thing is, how many games take advantage of that huge L3C. For those that do so, it is almost sure for AMD to retake the gaming-CPU crown back.

 

[TheinsanegamerN]

-Why are you running games at 720p in 2022???
-Because at higher resolution there is barely any difference between CPUs!
-Oh... OK then...

Why do we even test GPUs at 4k???? We should only bench at *arbitrary settings* because that's what REAL people use!

- straw man

They test at 720p because this is supposed to show the sheer difference between cpus with no other bottleneck present, that means reducing gpu load to the point the gpu is not a bottleneck. Just as running a 6600 at 4k ultra is not realistic either.

 

[napata]

Until you play a properly CPU limited game like Stellaris, CK3, Cities Skylines, Civ 6, Old Wolrd, Factorio etc etc.

Then you can be getting 2,000 fps at 4k Ultra but slow turn times or slow tic rates make them a pain to play on thr largest maps which totally changes the game play experience.

Strategy & RTS games on average are not bottlenecked by cache or the memory subsystem but the rather pure computational power of a CPU. In that sense it's possible a 5800x would outperform a 5800x3D in these games. It depends on the game and probably even the scene though because I think SC2 for example does gain from more L3 cache.

 

[TheinsanegamerN]

Strategy & RTS games on average are not bottlenecked by cache or the memory subsystem but the rather pure computational power of a CPU. In that sense it's possible a 5800x would outperform a 5800x3D in these games. It depends on the game and probably even the scene though because I think SC2 for example does gain from more L3 cache.

Sc2 benefits greatly from zen 3's avx improvements, as does sins of a solar empire. I don't know if the cache makes that much of a difference, unless there's a 5800x vs 5900x comparison I missed

 

[SL2]

clickbaiting at its finest. "oh lookie here, performance tests!'

Well they're not getting my clique muhahahaha.

 

[GaryPoisonOak]

I have a 3600x on a x570 board, so I’m very interested in what the 5800x3d will end up delivering. I see myself either getting it or the 5800x this year. Running native pc games is only one piece of the pie, though. I’ll be curious to see what if any kind of uplift it has in single core performance. It could be a situation where the 5800x is better at emulation for instance, due to the higher clock speeds.

 

[defaultluser]

I'm curious to see how much improved emulation is on these things (ps3 and 360 are still both well-shy of 60 fps, even on 5800X)

 

[SL2]

I have a 3600x on a x570 board, so I’m very interested in what the 5800x3d will end up delivering. I see myself either getting it or the 5800x this year. Running native pc games is only one piece of the pie, though. I’ll be curious to see what if any kind of uplift it has in single core performance. It could be a situation where the 5800x is better at emulation for instance, due to the higher clock speeds.

The 3D chip will most likely not make much difference in games in a computer that's limited by the GPU.

Emulation on the other hand is a different topic, and it's possible that we'll improvements there.

 

[DeathtoGnomes]

I specifically wrote test, not tests.

preemptive strike.

 

[noel_fs]

720p is how you can best compare cpu performance lol

 

[kapone32]

I don't think AMD would be doing the Youtube rounds if they didn't have a compelling product (based on the last 4 years of launches). Being that most Games are produced for AMD hardware for at least 3 years (Consoles) and they may actually have a gotcha for Intel with this chip.

 

[btk2k2]

Strategy & RTS games on average are not bottlenecked by cache or the memory subsystem but the rather pure computational power of a CPU. In that sense it's possible a 5800x would outperform a 5800x3D in these games. It depends on the game and probably even the scene though because I think SC2 for example does gain from more L3 cache.

Would be wonderful for there to be useful testing of this though rather than speculation. Also even when RTS or strategy games do get tested it tends to be more FPS tests rather than tic rate tests. You sometimes see factorio tests and Civ6 turn time tests but that is about it which is pretty meagre for such an oft played genre.

 

[Space Lynx]

Won't be surprised if it's faster, AMD created it on our purpose for this.
But I don't think it would make any difference for gaming at reasonable resolutions such as 1440p or 4k, at Fullhd most people won't be using a 3090 or a 6900XT hence are GPU limited anyways.

even if its not fake its a stupid test... no one games at 720p... this is just dumb. and does not deserve my attention