AMD Ryzen 7 5800X3D Gets Full Set of Gaming Benchmarks Ahead of Launch

[ARF]

I think this CPU is not even targeted for complete upgrade. You won't be upgrading from Ryzen 3 1300 to this...
This is more like a beta-test or showcase of a future technology which may or may not be implemented in any future products.

I think this kind of cache capacity upgrade needs at least the new N3 TSMC process node, so it will take smaller part of the die availability.

 

[Taraquin]

12700F + DDR4 + B660 motherboard won't be that much cheaper than the 5800X3D + DDR4 + B550 motherboard. In the UK the Intel CPU is about £100 less but the motherboard is around £50 more so you are looking at £50 more for the X3D system but it will be faster in games (sometimes by a lot) at the cost of productivity performance.

Also did you look at the lows? The 5800X3D even when tied @ 1080p often had better lows by 10% or more.

Remember that ram in gear 1 is stuck at 3400-3600 on locked alder lakes due to locked SA voktage beliw 1v, while 5800X3D at worst can do 3733, while may do 4000+ if lucky. In many cases a tweaked 12600KF would likely outperform 12700F since you can run ram at 4000-4300 in gear 1.

 

[watzupken]

It gets beaten by the 12900k at 1080p... plus the 12900k was running ddr4 not high speed ddr5 where the performance will increase

I think the idea is to show the difference in performance by just changing the CPU, while keeping as many factors constant as possible. I think people are quick to jump to the fact that the Intel platform is handicapped, but isn’t testing of a component meant to keep other factors constant, while focusing only on that component? The truth is whether you will see a jump in performance with DDR5, I think there are a lot of reviews that showed that it may in some cases, but generally not much difference. Also, you need to consider the price difference of DDR4 vs DDR5. Where I live, the latter is generally at least 2x more expensive than a decent pair of DDR4 3600 kit. And I believe quite a substantial number of Alder Lake owners like myself are likely using DDR4 due to poor or almost 0 availability of DDR5 at the end of last year and the start of this year.

In any case, I feel this is a stopgap solution from AMD, which is not worth buying unless AMD brings the price down further. With Zen 4 on the horizon and 3D cache baked into it, I don’t see any reason to get too excited over it. It is faster in games than the existing Zen 3 chips, but only in selected games and applications.

 

[Taraquin]

I can see a case for the 5800X3D even for a new build assuming the top priority is gaming performance.

The X3D is more then the 12700F but B550 is less than B660 for a good board and if you go DDR4 for both the ram is the same. Overall cost difference is maybe 15% more for the X3D build and in some games it will win by more than that so if the buyer plays those games it could be worth while. More of a tossup though than the existing AM4 owner, especially an AM4 owner who is on a 1xxxx/2xxx series part.

I think many sitting on a descent B450+3600 or 3700X can consider a 5800X3D

 

[Arkz]

i dont get it .. whats the point of this gaming benchmarks? whos gonna game with 3080ti with ultra low 720p? show some real benchamarks

That's how you bench a cpu for games while ensuring any bottleneck will be on the cpu and not gpu.

 

[Taraquin]

That's how you bench a cpu for games while ensuring any bottleneck will be on the cpu and not gpu.

Actually using DLSS/FSR uses input res as low as 720p so for many modern games it is relevant

 

[GeorgeJr]

I can see a case for the 5800X3D even for a new build assuming the top priority is gaming performance.

The X3D is more then the 12700F but B550 is less than B660 for a good board and if you go DDR4 for both the ram is the same. Overall cost difference is maybe 15% more for the X3D build and in some games it will win by more than that so if the buyer plays those games it could be worth while. More of a tossup though than the existing AM4 owner, especially an AM4 owner who is on a 1xxxx/2xxx series part.

Sure but imo, investing in a dead platform is kind of a bad practice. I'd prefer to upgrade to a fresh one with the possibility of further upgrading the CPU if the need exists.

 

[btk2k2]

Sure but imo, investing in a dead platform is kind of a bad practice. I'd prefer to upgrade to a fresh one with the possibility of further upgrading the CPU if the need exists.

Unless you are buying a lower end CPU like a 12400F and a good B660 motherboard I don't think RPL will be that much faster for non production workloads. 5-10% at best and a lot of that may be due to better and cheaper ram.

 

[THU31]

Waiting for TPU review, but this looks promising!

I am actually waiting for TPU to start measuring lows and frametimes, which are so much more important than average framerates.

 

[btk2k2]

I am actually waiting for TPU to start measuring lows and frametimes, which are so much more important than average framerates.

I am waiting for anywhere to start measuring Tic Rates and putting out more AI turn time tests for TBS games. Even for RTS games it would be good to see late game tic rates with lots of units on screen. Once the FPS is high enough the tic rate matters so so much more in a huge variety of games.

 

[ratirt]

Sure but imo, investing in a dead platform is kind of a bad practice. I'd prefer to upgrade to a fresh one with the possibility of further upgrading the CPU if the need exists.

It honestly is funny what you say. I'm sure Intel will release new CPUs on the same platform Alder Lake uses. If you ask people around, 90% of them will say that isn never going to happen. Obviously you can buy 12900K or KS later on but if gaming is what you are after, why would you bother if the difference between 12900K and 12700K is minimal? What sort of an upgrade is that anyway? I can bet you that the new Intel CPU will not share the same platform with Alder Lake. So your upgrade consists of same arch more cores and if you are already sitting on a 12900K or KS you are screwed since there will be nothing more for you there so this one is a dead platform before it had even gotten released.

 

[Leshy]

Actually using DLSS/FSR uses input res as low as 720p so for many modern games it is relevant

not rly... scaling is there to save some gpu performance for rays .. upscaling is handled by GPU so .. in gaming GPU is the limit.. benchmarking cpu in game is stupid

 

[Chrispy_]

The 1% lows are the important result here.

Nobody really cares if they're getting 200 or 300fps average but in a busy firefight the minute it drops below vsync or whatever tickrate the engine/server runs at you'll notice it and want to drop quality settings.

 

[ZoneDymo]

This was what i was talking about. Claim vs claim with ddr4.
12900KS was not tested with ful potential here. Lets see more of this now, where this can lead us conclude.

ermm plenty of reviews have already pointed out that DDR5 does not give a benefit as is...

i dont get it .. whats the point of this gaming benchmarks? whos gonna game with 3080ti with ultra low 720p? show some real benchamarks

Is this a joke? the smileys make me believe you are being sarcastic but im not sure.

if its not a joke, its about cpu limited benchmarks to show what the cpu contributes, if it was a 4k then the top 7 cpu's out there all preform the same as its GPU limited.

 

[fevgatos]

It honestly is funny what you say. I'm sure Intel will release new CPUs on the same platform Alder Lake uses. If you ask people around, 90% of them will say that isn never going to happen. Obviously you can buy 12900K or KS later on but if gaming is what you are after, why would you bother if the difference between 12900K and 12700K is minimal? What sort of an upgrade is that anyway? I can bet you that the new Intel CPU will not share the same platform with Alder Lake. So your upgrade consists of same arch more cores and if you are already sitting on a 12900K or KS you are screwed since there will be nothing more for you there so this one is a dead platform before it had even gotten released.

Im pretty sure raptorlake will be supported on z690.

 

[Denver]

I didn't expect the difference to reach almost 30% in some cases. Very good for a simple cache upgrade without architectural changes.

I keep wondering what else AMD can stack on top of the CPU... iGPUs ? Ram ? Asics dedicated to AI or RT ? It will be very interesting to find out.

 

[Taraquin]

not rly... scaling is there to save some gpu performance for rays .. upscaling is handled by GPU so .. in gaming GPU is the limit.. benchmarking cpu in game is stupid

Some games have upscaling but not rays (and many use upscaling w/o rays) and if you want to compare cpu perf then low res is smart since you remove gpu from equation

 

[TheLostSwede]

I didn't expect the difference to reach almost 30% in some cases. Very good for a simple cache upgrade without architectural changes.

Keep in mind that the actual CPU is slower in terms of clock speeds than 5800X.

I keep wondering what else AMD can stack on top of the CPU... iGPUs ? Ram ? Asics dedicated to AI or RT ? It will be very interesting to find out.

Apparently they filed some patents about some AI/ML stuff.

AMD's Future CPUs Could Feature a 3D-Stacked ML Accelerator

Genoa-AI? Bergamo-AI? Turin-AI?

www.tomshardware.com

 

[Tsukiyomi91]

what's gonna be even more interesting is pushing both the 5800X3D and 12900KS over their limits and run benchmarks at 1440p. It should be a really interesting one.

 

[Dr. Dro]

It gets beaten by the 12900k at 1080p... plus the 12900k was running ddr4 not high speed ddr5 where the performance will increase

TPU testing is quite clear on DDR5 not being useful at all for games on Alder Lake - i.e. it's a waste of money, high-end DDR5 performing within a percent of high-end DDR4. You can happily grab any SKU - including the new KS and make full use of it using your existing DDR4 memory.

Hint:

Whoa man that 12900KS is a whole frame per second faster than a 5800X, how can you not justify upgrading smh what are you, poor?

That's why this isn't a CPU for someone building a new PC now. It's for AM4 users who want to upgrade. You can't beat the price of just the 5800X3D compared to a CPU, mobo and RAM.

I'm not entirely sure I agree at the SEP AMD is asking for. Americans with access to Micro Center can purchase the i7-12700K + mid-level motherboard with a $450 budget, and this option will be upgradable to Raptor Lake. There's also the Ryzen 9 5900X available in most places in the world for that same money. Whether the extra cache or the extra four cores and eight threads are the better pick, I suppose it will come down to the specific game one may want to play, and whether they will be streaming, recording, or multitasking with the system, as well.

 

[ratirt]

Actuall

Im pretty sure raptorlake will be supported on z690.

You are correct. I double checked and it would seem the Raptor Lake will be backwards compatible. Good for Intel fans.
About time something does not die with a release for Intel customers.

 

[Vya Domus]

Some of these results seem hard to believe, I really doubt the cache makes that much of a difference.

 

[DeathtoGnomes]

Still not a complete review. A 3700x and a 5950x system should have been used too, not just intel. It was already said this CPU seems to be targeted for upgrading older existing ryzen builds, not as a new system build. Dropping this into a 300/400 series boards should make a huge difference to those with cards that were otherwise bottlenecked.

 

[Taraquin]

Some of these results seem hard to believe, I really doubt the cache makes that much of a difference.

Look at 3100 vs 3300X at 4.0GHz fixed. The major difference is that 3300X has access to double cache vs 3100 that must share it between ccx-es.

 

[THU31]

Some of these results seem hard to believe, I really doubt the cache makes that much of a difference.

Cache is the main reason why Alder Lake is so good. And that is also where the main difference in gaming between i5, i7 and i9 comes from. Cache has always been very important for gaming.

This is also why the new low-end AMD CPUs are bad.