Poll: Ryzen 7950X3D, 7900X3D, 7800X3D

[rare]

So, in all honesty, going from a 7900 to 7950X3D do you think one would notice much of a difference?

Depends on many factors, what GPU and what resolution?

 

[Iron Man]

So, in all honesty, going from a 7900 to 7950X3D do you think one would notice much of a difference?

I don’t think it would make that much of a difference. Going up a GPU class would make more of a difference.

 

[01001101]

Depends on many factors, what GPU and what resolution?

7900XTX @ 3440x1440

 

[rare]

7900XTX @ 3440x1440

At that resolution I don’t think you will see a big uplift. Minimums will improve no doubt. If you look at reviews of the 7900 at 4k even with high end GPU‘s there’s not much in it performance wise.

 

[01001101]

At that resolution I don’t think you will see a big uplift. Minimums will improve no doubt. If you look at reviews of the 7900 at 4k even with high end GPU‘s there’s not much in it performance wise.

Fair enough. That's good to know. Thank you for your answer and advice.

 

[uscool]

Z690 and Z790 are 99.9% identical. Main difference is official support (QVL) for £500+ 8000Mhz DDR5 kits, which not many would buy anyway.

13900k running flawlessly on a Z690 hero here

Obviously AM5 will last longer, though it's a much more expensive platform. DDR5 + AM5 vs DDR4 and £85 board, no comparison from cost point of view.

this ? MSI PRO H610M-B DDR4 not sure I would wanna pair anything with this

 

[rare]

Pickup a cheap 7600 from OCUK in the meanwhile, sensible thing to do is the test your parts to ensure none are DOA.

Very tempted by this considering you can buy one for about £230.

Heck you can even get a 7700x for £312 if you look about.

Might skip x3d altogether at this rate and save a few quid considering I will be playing at 4k with a 4090.

 

[timorous]

Can't wait for all the useful 1080p 700 fps benchmarks

You do know how to use benchmarks don't you.

If you want 1440p 120fps then you need a CPU that can achieve 120 fps in your desired titles / engines. If in the 'useful 1080p' tests your CPU can only manage 60fps then no matter what GPU you pair it with you are maxing out at 60fps and you are short of your goal.

So get the CPU that does 120+ fps in the titles you want, find a GPU that does 120fps in the titles you want at your desired resolution and then pair the two together.

To be honest it never occurred to me that people did not understand this because I thought it was blindingly obvious.

 

[timorous]

So, in all honesty, going from a 7900 to 7950X3D do you think one would notice much of a difference?

Entirely depends on the games you play. AAA fare and GPU matters more. Sim racing / flight sim and you will see differences unless you have a really weak GPU. If you want to improve Stellaris tic rates or Cities tic rates or the tic rate of other games then you should see pretty big uplifts. 5800X3D was 60% faster in Stellaris tic rate than the 5900X and you see similar gains in Factorio. Football Manager is another candidate where X3D might be disproportionately faster due to the cache but nobody tests it so we don't actually know.

 

[timorous]

At that resolution I don’t think you will see a big uplift. Minimums will improve no doubt. If you look at reviews of the 7900 at 4k even with high end GPU‘s there’s not much in it performance wise.

At that resolution the games played matters more. This high res = GPU bound always and CPU doesn't matter nonsense needs to stop because it is BS.

Tell Star Citizen players that CPU does not matter, even at high res, or Stellaris / other Grand Strategy games or Factorio and those style of games. Same for MMOs where the limiting factor in large PVP or end game raids is CPU. What about Path of Exile which is heavily CPU bound in late game maps. So many games do not fit the standard AAA mantra of 'only GPU matters'.

Also Stellaris, HOI4, EU4, CK3, Factorio, Civ 6, Football Manager, PAth of Exile are all games in the steam top 100 and they rarely get benched in reviews where as we frequently see Watch Dogs: Legion or Borderlands 3 or Tomb Raider or Far Cry 6 etc, games that are far less popular.

TBH CPU benchmarking for games needs a bit of a shakeup and if someone had the time and a bit of starting budget to come in with a channel that did a combination of the current popular AAA games with games that are representative of popular engines like UE4/UE5 and then 4/5 games where the key metric is not FPS but tic-rate then I think they would be really successful.

 

[SilverCider]

You do know how to use benchmarks don't you.

If you want 1440p 120fps then you need a CPU that can achieve 120 fps in your desired titles / engines. If in the 'useful 1080p' tests your CPU can only manage 60fps then no matter what GPU you pair it with you are maxing out at 60fps and you are short of your goal.

So get the CPU that does 120+ fps in the titles you want, find a GPU that does 120fps in the titles you want at your desired resolution and then pair the two together.

To be honest it never occurred to me that people did not understand this because I thought it was blindingly obvious.

Whilst it is blindingly obvious in most regards, 720p/1080p CPU data does not necessarily guarantee you the same throughput at 1440p/4k. RT is a good example here since at higher resolutions it places a much greater demand on the CPU (regardless of whether the GPU has the grunt to cope).

 

[Nutty667]

You do know how to use benchmarks don't you.

If you want 1440p 120fps then you need a CPU that can achieve 120 fps in your desired titles / engines. If in the 'useful 1080p' tests your CPU can only manage 60fps then no matter what GPU you pair it with you are maxing out at 60fps and you are short of your goal.

So get the CPU that does 120+ fps in the titles you want, find a GPU that does 120fps in the titles you want at your desired resolution and then pair the two together.

To be honest it never occurred to me that people did not understand this because I thought it was blindingly obvious.

To me it's even simpler than that. You just want to remove any potential GPU bottleneck to see how well the CPU performance scales.
The only way to do that is to test the game in a low resolution.

If you bench at 4K and the frame-rates are all the same, does that mean CPU performance is all the same? Of course not. You're just hitting the FPS ceiling of the GPU.

 

[Robert896r1]

People hold onto cpu's longer than gpu's so you want to ultimately see when your CPU would an issue. The only way to do that is to pick a flagship gpu of the current generation and drop resolution until the GPU is no longer a bottleneck.

 

[subbytna]

Entirely depends on the games you play. AAA fare and GPU matters more. Sim racing / flight sim and you will see differences unless you have a really weak GPU. If you want to improve Stellaris tic rates or Cities tic rates or the tic rate of other games then you should see pretty big uplifts. 5800X3D was 60% faster in Stellaris tic rate than the 5900X and you see similar gains in Factorio. Football Manager is another candidate where X3D might be disproportionately faster due to the cache but nobody tests it so we don't actually know.

Football Manager 2023 at 700FPS on a 3D chip I'd love someone to run a quick FPS test on it lol

 

[kindai]

this ? MSI PRO H610M-B DDR4 not sure I would wanna pair anything with this

That is the saddest looking motherboard ive ever seen

 

[Iron Man]

That is the saddest looking motherboard ive ever seen

I have had worse

 

[Deleted member 258511]

You do know how to use benchmarks don't you.

If you want 1440p 120fps then you need a CPU that can achieve 120 fps in your desired titles / engines. If in the 'useful 1080p' tests your CPU can only manage 60fps then no matter what GPU you pair it with you are maxing out at 60fps and you are short of your goal.

So get the CPU that does 120+ fps in the titles you want, find a GPU that does 120fps in the titles you want at your desired resolution and then pair the two together.

To be honest it never occurred to me that people did not understand this because I thought it was blindingly obvious.

The only way to test a cpu is to remove the GPU bottleneck hence people testing at 720p. What confuses some people is just that “I play at 4k, who plays at 720p snnnort” Pointless testing a CPU at 4k. It would be like testing a cars bhp on a rolling road that is capped at 200bhp. No one would have a clue what their car is capable of without removing that cap.

 

[rare]

I couldn’t be bothered waiting until April anymore so I ordered a 7600x for £234. At least that way I can make sure my motherboard and 4090 actually work and then look to upgrade to a x3d in the summer.

 

[pigeonguyuk]

I couldn’t be bothered waiting until April anymore so I ordered a 7600x for £234. At least that way I can make sure my motherboard and 4090 actually work and then look to upgrade to a x3d in the summer.

Considering do this now myself. As if you wait to upgrade to the X3D about 3 months after launch when supply vs demand is better the "launch" premiums will be gone and get a better overall deal

 

[rare]

Considering do this now myself. As if you wait to upgrade to the X3D about 3 months after launch when supply vs demand is better the "launch" premiums will be gone and get a better overall deal

That’s what I’m thinking, I will be able to sell the 7600x for a fair price I’m sure and then pick a x3d at a lower price later in the year. I’ve already had my 4090 about two months in a box, I can’t be arsed waiting until April anymore, might be dead by then!