***OFFICIAL*** Ryzen 5000 / Zen 3 Launch Thread REVIEWS BEGIN PAGE 39

[coercitiv]

Im sure you all know that 10600K was way faster vs 3700X in gaming that what 5600X is against 10700 today.

I'm not interested in this 10700 vs 5600X debate, but I would like to remind you something: on any GPU bellow the 2080Ti the difference between 10600K and 3700X was under 3%, rapidly converging towards zero.

Spoiler: 5.5% @1080p

Spoiler: 3.5% @ 1440p

 

[Turtle.Man]

So you can save around 35 quids and than spend around 150 more for high end ram, and 50 more for bigger PSU to have similar performance

How is this competition? This is the definition of utter lack of it.

But the o/c'd Intel chip will also heat your house in winter: I'll bet the Ryzen can't do that!

 

[Hitman928]

I really dont have any numbers for this so I cannot reply, but i dont believe the time difference is that high that will have a meaningful impact in gaming. But im open to any links that might be taking a look in to this.

I just saw an article about 2 weeks ago comparing game load times and the pcie4 drives showed faster load times. The difference was not large between gen 3 and gen 4 nvme drives and obviously the biggest change was from a platter drive to any SSD, but some people value shaving even a few seconds off of load times. I'll try to see if I can find the article again.

 

[Kenmitch]

But the o/c'd Intel chip will also heat your house in winter: I'll bet the Ryzen can't do that!

Pair it with a power hungry Ampere and it'll get the job done just fine! Nice try! /s

 

[amrnuke]

I didnt say it does more, I said its a better choice because it has 2C 4T more and might mature better due to consoles that have 8C 16T.

Let's take an extreme MT test like CB20 MT, the 5600X (6/12) is within 3% of a 9900K (8/16) and beats the 2700X (8/16) by nearly 10%. Clearly, even one of the most MT-dependent benchmarks is responsive also to single-threaded performance. To assume the worst case scenario with games, that they'll be able to fully utilize 8/16, one should also consider that they will be single-thread responsive as well, and the overall single-thread crown is worth bearing in mind when making purchase decisions.

 

[AtenRa]

I'm not interested in this 10700 vs 5600X debate, but I would like to remind you something: on any GPU bellow the 2080Ti the difference between 10600K and 3700X was under 3%, rapidly converging towards zero.

Spoiler: 5.5% @1080p

View attachment 33563

Spoiler: 3.5% @ 1440p

View attachment 33564

True, same applies for the 10700 vs 5600X. The higher the resolution the performance difference is getting meaningless.

 

[Dave3000]

It does nothing for gaming. Not even with a RTX3090

The only reason why I'm hesitant to buy a 10700k now and waiting for 5800x's or 5900x's to be in stock again is because of PCIe 4.0 and I intend to buy a PCIe 4.0 NVme drive in the near future. Also I intend to buy either a 6800XT, 6900XT, or a RTX 3080. XBox One Series X and PS5 have PCIe 4.0 and a PCIe 4.0 NVme drive. However, if PCIe 4.0 is only going to make only a few percent difference for PC at best, despite those two consoles have PCIe 4.0, then I might just go for the 10700k and a Z490 motherboard as I have other parts (power supply, case, and CPU cooler) that I already purchased 3 weeks ago waiting to be installed in my upcoming rig and would like to know if they are working and not broken.

 

[Thunder 57]

Please, if anyone needs help is people who think AMD still needs help from fanboys. 3600X was bad value, 5600X is bad value. AMD is milking early adopters and faithful cause they can, but let's stop with the spin that somehow makes 5600X value champion and lets call it overpriced 6C with many cheaper options from both companies.

redacted If it's a bad value, don't buy it. Get a cheaper six core. No one is calling the 5600X a "value champion". That is a straw man.



Profanity in tech is not allowed, nor is the use of the word fanboy.


esquared
Anandtech Forum Director

 

[Rigg]

This notion of a 10700 system being a viable alternative is a joke. Does anyone seriously think that a 10700, cheap B460, 2666 RAM, stock cooler combo would compete with a 5600x, B550, 3600 CL16 RAM stock cooler combo? The intel can't compete in gaming without faster RAM. Not to mention it would get trashed in MC loads without a cooling solution and mobo that can handle disabling the power limits. Intel can't compete without Z490.

 

[Hitman928]

The only reason why I'm hesitant to buy a 10700k now and waiting for 5800x's or 5900x's to be in stock again is because of PCIe 4.0 and I intend to buy a PCIe 4.0 NVme drive in the near future. Also I intend to buy either a 6800XT, 6900XT, or a RTX 3080. XBox One Series X and PS5 have PCIe 4.0 and a PCIe 4.0 NVme drive. However, if PCIe 4.0 is only going to make only a few percent difference for PC at best, despite those two consoles have PCIe 4.0, then I might just go for the 10700k and a Z490 motherboard as I have other parts (power supply, case, and CPU cooler) that I already purchased 3 weeks ago waiting to be installed in my upcoming rig and would like to know if they are working and not broken.

There's also the SAM feature that you gain access to with a RDNA2 GPU, Zen3 CPU, and pcie4. I keep forgetting about that one. It might not be a huge boost but an extra 5 - 10% for free if your components work out isn't a bad thing.

 

[amrnuke]

True, same applies for the 10700 vs 5600X. The higher the resolution the performance difference is getting meaningless.

Exactly.
10700K ($370) + Z490 mobo ($130+) + 3600CL16 RAM ($75) = $575
5600X ($300) + B550 mobo ($110) + 3600CL16 RAM ($75) = $485

That $90 is more than the difference between a Radeon 6800 and 6800XT, and that difference will matter much more than the difference between a 10700K and 5600X.

 

[richierich1212]

Exactly.
10700K ($370) + Z490 mobo ($130+) + 3600CL16 RAM ($75) = $575
5600X ($300) + B550 mobo ($110) + 3600CL16 RAM ($75) = $485

That $90 is more than the difference between a Radeon 6800 and 6800XT, and that difference will matter much more than the difference between a 10700K and 5600X.

Not only that but the 10000-series Intel CPUs suck down much more power when overclocked. That's a huge factor in the long term in places where electricity isn't cheap.

 

[ondma]

Maybe. And then they'll require a different chipset.

All the preliminary information I have seen so far has said you will not need a new motherboard for Rocket Lake. You are simply incorrect or making stuff up.

 

[AtenRa]

Exactly.
10700K ($370) + Z490 mobo ($130+) + 3600CL16 RAM ($75) = $575
5600X ($300) + B550 mobo ($110) + 3600CL16 RAM ($75) = $485

That $90 is more than the difference between a Radeon 6800 and 6800XT, and that difference will matter much more than the difference between a 10700K and 5600X.

I was talking about 10700 non K

10700 ($348) + B460 mobo ($110+) + 3600CL16 RAM ($75) = $533
5600X ($300) + B550 mobo ($110) + 3600CL16 RAM ($75) = $485

Total difference 48$

With cards like the 6800/6800XT the vast majority of gamers will use them at 1440p to 4K. At those resolutions the CPU performance difference is negligible between 10700 and 5600X.
In fact if you are in a low budget but still aiming at 1440p/4K , it will be much better to go for the cheaper Core i5 10400 and invest the price difference to a better GPU.

 

[shady28]

Let's take an extreme MT test like CB20 MT, the 5600X (6/12) is within 3% of a 9900K (8/16) and beats the 2700X (8/16) by nearly 10%. Clearly, even one of the most MT-dependent benchmarks is responsive also to single-threaded performance. To assume the worst case scenario with games, that they'll be able to fully utilize 8/16, one should also consider that they will be single-thread responsive as well, and the overall single-thread crown is worth bearing in mind when making purchase decisions.

3% of a 9900K, or a 10700K, on multi? Not hardly, is that some kind of bizarre napkin math you're doing?

These charts show big mult-thread gains of 10700K / 9900K vs 5600X - at stock.

If you want to run everything stock, these charts are a good representation.

However the only reason to buy a 10700K and not a 10700 is to overclock it, and the only reason to build your own in general is to tweak it.

I those cases both can perform far higher than these numbers - like 10-25% higher - using good air cooling or a relatively inexpensive AIO water cooler.

But on to the stock comparison -

From Guru3d :

 

[Rigg]

I was talking about 10700 non K

10700 ($348) + B460 mobo ($110+) + 3600CL16 RAM ($75) = $533
5600X ($300) + B550 mobo ($110) + 3600CL16 RAM ($75) = $485

Total difference 48$

With cards like the 6800/6800XT the vast majority of gamers will use them at 1440p to 4K. At those resolutions the CPU performance difference is negligible between 10700 and 5600X.
In fact if you are in a low budget but still aiming at 1440p/4K , it will be much better to go for the cheaper Core i5 10400 and invest the price difference to a better GPU.

Dude.....You can't run above 2666 RAM on b460. I pointed this out in my previous post. Your argument is completely flawed. It's z490 or bust.

EDIT: I just remembered it's 2933 for the i7 and i9 and 2666 for the lower tier stuff. The point remains.

 

[Hitman928]

3% of a 9900K, or a 10700K, on multi? Not hardly, is that some kind of bizarre napkin math you're doing?

These charts show big mult-thread gains of 10700K / 9900K vs 5600X - at stock.

If you want to run everything stock, these charts are a good representation.

However the only reason to buy a 10700K and not a 10700 is to overclock it, and the only reason to build your own in general is to tweak it.

I those cases both can perform far higher than these numbers - like 10-25% higher - using good air cooling or a relatively inexpensive AIO water cooler.

But on to the stock comparison -

From Guru3d :

View attachment 33578

View attachment 33579


View attachment 33584

A 10700 is a power limited and lower clocked version of the 10700k, without even taking overclocking into account. It doesn't perform the same as a 10700k at stock and will be significantly slower in heavy multi threaded situations.

 

[uzzi38]

AVX2 VAES support just landed into Geekbench, seems to have a pretty big effect on scoring.

5800X GB5.3 Windows running 4.85GHz single core (this is stock behaviour - it's the maximum boost allowed by Precision Boost 2)

ST score: 1717.

Generic - Geekbench

Benchmark results for a Generic with an AMD Ryzen 7 5800X processor.

browser.geekbench.com

AES scoring is in the same region as TGL now, so it's a noticable bump to scores of around 5%. Might even be enough to push Linux Zen 3 scores into the 1800pts range.

 

[AtenRa]

Dude.....You can't run above 2666 RAM on b460. I pointed this out in my previous post. Your argument is completely flawed. It's z490 or bust.

EDIT: I just remembered it's 2933 for the i7 and i9 and 2666 for the lower tier stuff. The point remains.

A 10700 is a power limited and lower clocked version of the 10700k, without even taking overclocking into account. It doesn't perform the same as a 10700k at stock and will be significantly slower in heavy multi threaded situations.

10700K all core turbo = 4.7GHz
10700 all core turbo = 4.6GHz

Gaming doesnt stress the CPU that much, the performance difference between 10700K and non K model is 5% at best.
Also, memory speed is not that much of a performance uplift with Intel CPUs as is with Zen 1 and 2

The video bellow perfectly illustrate the extremely small performance difference between the Core i7 10700 + 2933MHz memory vs Core i7 10700K + 3600MHz memory.



i7 10700 (B460 + 2933) vs i7 10700k (Z490 + 3600) with RTX2080Ti

 

[lobz]

Why would i complain when they released obviously better choices in 10400F @155 eur and 10700F @320 eur ? Takes true AMD fan to slot in 5600X @ 340 eur and also call it a value when 3600 also exists in budget choices.
But i remember complaining and hoping Intel unlocks memory speed for lower rank chipsets. That did not happen and i hope Intel's bottom line will burn for that.

Have completelt lost your mind? What sick game do you wanna play with us? It was YOU who started to talk about the 5600X as it had to be a good choice for the budget gamer.

So... Let me help you: the 5600X offers good value with its performance AT ITS OWN PRICE POINT.

Geez this has been such a freaking dense topic... Richie Rich could learn a lot frok you guys.

 

[lobz]

Nah my next box will probably be an M1.

That explains a lot.

 

[Timorous]

True, same applies for the 10700 vs 5600X. The higher the resolution the performance difference is getting meaningless.

Not in Civ 6 or Stellaris or a bunch of other games where they become very CPU bound even at 4K. It is a shame that these games are not tested in reviews as much as improving the performance of late game stellaris or reducing AI turn time in Civ 6 are meaningful improvements.

As far as Civ 6 is concerned the 5600X is faster than the 10900K let alone the 10700 and since this is turn time resolution will have no impact on the performance improvemet.

 

[lobz]

Please, if anyone needs help is people who think AMD still needs help from fanboys. 3600X was bad value, 5600X is bad value. AMD is milking early adopters and faithful cause they can, but let's stop with the spin that somehow makes 5600X value champion and lets call it overpriced 6C with many cheaper options from both companies.

It's the best gaming CPU for about $300 with no obvious contenders within +-$50 right now. So tell me, how the freaking molding fudge is it bad value?

 

[Veradun]

I didnt say it does more, I said its a better choice because it has 2C 4T more and might mature better due to consoles that have 8C 16T.

Ok, let's translate:

10700 is in your opinion a better choice, compared to 5600X, because with only 50$ bucks more it will perform worse and lacks PCIe4. Cool.

I strongly advise to ignore AtenRa advices on hardware.

 

[krumme]

As it is today people that buy amd does so because perf/cost ratio, not because brand. Intel is still a tad overpriced. Might perfectly change with all the amd hype atm.
I am not running 2933 ram today when 3600c16 is nearly same price. Lets be realistic neither will Altenra nor anyone here. Its z490 for gaming or ryzen3 line. And ryzen is just faster for gaming, so its not that difficult to choose.
When gaming you run into different bottlenecks. If we look back 3 years Battlefield 1 killed those 4 core cpu in mp64 battles due to lacking throughput, but eg PUBG killed the ryzen 1 line due to memory latency.
We are a much better place today.
I think the 3600 on a cheap b550 and 3600c16 is a fine balance here for budget options if you intend to play those games that demand throughput. And it can be upgraded. If funds alow 5600x. And it goes for every workload. Intel is much cheaper than it was and better value, but ryzen 3 seems to have no weak sides. Its a safe buy.
If Intel lowers prices even more and stop the idiotic motherboard ram speed limitation, well we just buy that. But they are still worse value as it stands. High price on 5600x doesnt change that. I understand people want cheaper 5600x its expensive for 6 core. The sad reality is just the alternatives (10700 with 2933 ram/10600k 490 3600c16) is just slower.