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

[amrnuke]

What's bad faith about 3600 vs 5600X, i.e. $200 vs $300, i.e. $100 price increase?
What's bad faith about 3700X vs 5800X, i.e. $330 vs $450, i.e. $120 price increase?

Again, there are no 5600 or 5700X CPUs at the moment and these are direct comparisons.

Twisting logic, using facts selectively and applying double standards have become AMD fans favourite game.

Yes, but there were 3600X and 3600XT and 3800X and 3800XT. You are choosing to compare to the lower-binned parts because it helps you feel better about your worldview. It's a wrong worldview.

An analogous error would be to compare the 10600K to the 9400. But you wouldn't make that mistake, would you?

 

[Veradun]

Gaming kings even on Linux! Fascinating stuff, to squeeze so much performance from the same process node is pure engineering masterpiece.

Forcing to retest every GPU on the new king platform is the real masterpiece here, on every account. Lisa Su is the reincarnation of Sun Tzu.

 

[yeshua]

Kinda just the reality. It's not like the prices are that out of line with what we've become accustomed to paying. Definitive performance crown + covid hardware demand in-elasticity, what do you expect?

Why is it absolutely not OK for Intel and NVIDIA to raise prices when they introduce new better products and suddenly it's just fine for AMD? And today I've already heard next to a dozen quite far-fetched arguments why it's perfectly OK. Kinda makes you think that people are extremely biased against Intel and NVIDIA despite AMD starting to copy their worst behavior. How many years ago did Intel last abuse their market power? Ten? Should I remind you about the FX-57 CPU by AMD which was exorbitantly priced (over $1K in 2005) because Intel kept pushing their brain-dead power hungry underperforming Pentium Prescott CPUs? COVID demand, right. Should we mention aliens or Solar eclipses as well? Actually, COVID right, except most people prefer to use laptops when working remotely and everyone who needed to work remotely has started doing so months ago and they don't really need to upgrade their desktop computer. I guess the demand for the Ryzen 5000 series comes mostly from tech-enthusiasts or dedicated gamers who want to extract each extra FPS while playing at 1080p because they have the money for that.

The prices are absolutely out of line if you don't live in a first-world country. Even Ryzen 3600 at $200 is far too expensive for the majority of people on Earth and now we are talking about at the very least $300.

 

[UsandThem]

Should we mention aliens or Solar eclipses as well?

Really? In a tech discussion thread you're using those two things as analogy?

Actually, COVID right, except most people prefer to use laptops when working remotely and everyone who needed to work remotely has started doing so months ago and they don't really need to upgrade their desktop computer.

Have you been to everyone's house to see what device they use for work? You never stopped by my house, because along with my wife and kids, we all prefer using a desktop for actual work.

I guess the demand for the Ryzen 5000 series comes mostly from tech-enthusiasts or dedicated gamers who want to extract each extra FPS while playing at 1080p because they have the money for that.

Uh, yeah. This is an hardware enthusiast site. I doubt anyone who isn't an "enthusiast" gets excited over CPU releases.

 

[yeshua]

Notice how little people have complained about Intel/Nvidia pricing? "Gotta pay to play" etc. hardly a peep. Now AMD slightly increases prices and BLARGHHGGHGGGH how could AMD do that?!?!?!

My take: I will happily pay more for AMD in the short term and I hope AMD kicks Intel around for a few more years and builds up a gargantuan war chest of cash. If people don't agree or like the above, that's great then go out and buy Intel.

Like really? In almost every thread and every review about RTX 3000 cards at least a few times it's mentioned how overpriced RTX 2000 cards were and how NVIDIA played everyone and that the RTX 3000 generation only looks good because of the RTX 2000 GPUs before.

 

[AnandThenMan]

How many years ago did Intel last abuse their market power? Ten?

When for years AMD outright bribes OEMs to only use their products to the point where Intel nearly goes under, then you can start to make an apples to apples comparison.

 

[coercitiv]

Even Ryzen 3600 at $200 is far too expensive for the majority of people on Earth and now we are talking about at the very least $300.

Get some sleep mate, this is getting out of hand. The 3600 is not the CPU the vast majority of people on Earth require.

 

[inf64]

Back to technical stuff, a beautiful comment from AT article:

+19% IPC

The key metric offered by AMD was a +19% IPC uplift from Zen 2 to Zen 3, or rather a +19% uplift from Ryzen 5 3800XT to Ryzen 5 5800X when both CPUs are at 4.0 GHz and using DDR4-3600 memory.

In fact, using our industry benchmarks, for single threaded performance, we observed a +19% increase in CPU performance per clock. We have to offer kudos to AMD here, this is the second or third time they've quoted IPC figures which we've matched.

In multithreaded SPECrate, the absolute gain was only around 10% or so, given that faster cores also require more bandwidth to main memory, which hasn’t been provided in this generation. This means that there are some bottlenecks to which a higher IPC won’t help if more cores require the same resources.

For real-world tests, across our whole suite, we saw an average +24% uplift. For explicitly multithreaded tests, we saw ranges from even performance up to +35%, while for explicitly single threaded tests, this ranged from even performance up to +57%. This comes down to execution/compute bound tests getting bigger speedups over memory bound workloads.

 

[yeshua]

When for years AMD outright bribes OEMs to only use their products to the point where Intel nearly goes under, then you can start to make an apples to apples comparison.

Remind me what competive CPUs did AMD have between AMD 64 and Ryzen. Remind me what made AMD almost sink itself by purchasing ATI which I still find to be a very questionable move because they have never truly realized the goals behind the merger (they wanted to create a Fusion APU which just never happened - we have something close to it in Xbox/SPS but that could have been done without this merger). You cannot blame everything on one company when the other company through mismanagement or dubious cash spendings is comming suicide (Intel was in the same boat with Brian Krzanich). I'm not defending Intel. I'm trying to defend healthy competition.

 

[Ajay]

I have zero faith in RocketLake: 14nm, huge dies, insane power consumption. The earliest Intel can redeem themselves seems to be their new 7nm node CPUs at the end of 2022 (they seem to be unable to fix their 10nm node power consumption) but yeah, I'd love to have healthy competition in the x86 market which is now run by a single company while the other one keeps ripping the rewards of their ~14 years near monopoly (2004-2018).

It's looking more and more like the 7nm desktop CPUs are going to 2023. Even if RKL matches Zen3 - it looks like Intel will be out of bullets after that (ADL just doesn't look like it will be faster, given that it's on 10nm).

 

[yeshua]

Back to technical stuff, a beautiful comment from AT article:

Yeah, and I'm especially curious how Zen 3 has become so much faster at compressing using WinRAR - a performance increase is insane, over 65%. That is not reflected in this review, but here are the results of the integrated benchmark:

 

[amrnuke]

Why is it absolutely not OK for Intel and NVIDIA to raise prices when they introduce new better products and suddenly it's just fine for AMD? And today I've already heard next to a dozen quite far-fetched arguments why it's perfectly OK. Kinda makes you think that people are extremely biased against Intel and NVIDIA despite AMD starting to copy their worst behavior. How many years ago did Intel last abuse their market power? Ten? Should I remind you about the FX-57 CPU by AMD which was exorbitantly priced (over $1K in 2005) because Intel kept pushing their brain-dead power hungry underperforming Pentium Prescott CPUs? COVID demand, right. Should we mention aliens or Solar eclipses as well? Actually, COVID right, except most people prefer to use laptops when working remotely and everyone who needed to work remotely has started doing so months ago and they don't really need to upgrade their desktop computer. I guess the demand for the Ryzen 5000 series comes mostly from tech-enthusiasts or dedicated gamers who want to extract each extra FPS while playing at 1080p because they have the money for that.

The prices are absolutely out of line if you don't live in a first-world country. Even Ryzen 3600 at $200 is far too expensive for the majority of people on Earth and now we are talking about at the very least $300.

I'd love to hear your opinions on the exorbitant cost of a Toyota Corolla that the majority of people on Earth can't afford. AMD will be releasing a broad line-up of chips. Lest you forget, they released a $100 Ryzen 3100 just 6 months ago. Meanwhile, Intel still charge $100 for some of their 2-core Celerons.

You are taking things way out of context because you picked a "team" instead of appreciating the advancements being made.

 

[rainy]

Back to technical stuff, a beautiful comment from AT article:

This is another (after performance) pleasant surprise: even as AMD fan, I'm always quite sceptical about marketing numbers/slides and looks like it wasn't exaggeration on their side.

 

[inf64]

Yeah, and I'm especially curious how Zen 3 has become so much faster at compressing using WinRAR - a performance increase is insane, over 65%. That is not reflected in this review, but here are the results of the integrated benchmark:

This is (most likely) the effect of +50% more loads and stores per cycle in Zen3 vs Zen 2 :

 

[itsmydamnation]

i only just got up (australia) just started reading anad review

FPU looks like it has more ports but not wider ( more store ports etc)?

So anyone noticed that Zen3 now look like a big K7/K8
also seemed like the OOOE window didnt grow much.

will need to keep reading but if all that is true , then this is massive for AMD , they have very real room to grow come Zen5 etc

 

[inf64]

i only just got up (australia) just started reading anad review

FPU looks like it has more ports but not wider ( more store ports etc)?

So anyone noticed that Zen3 now look like a big K7/K8
also seemed like the OOOE window didnt grow much.

will need to keep reading but if all that is true , then this is massive for AMD , they have very real room to grow come Zen5 etc

I agree and I personally think Zen5 has the same ~40% IPC goal versus Zen3 as AMD has set in the past ( Zen1 Vs EX and Zen3 vs Zen1). Zen4 should be a very solid ~13-15% IPC jump, just about enough to stay above the competition. Zen5 might have some neat tricks as well, such as more threads (>2) per core

 

[GoodRevrnd]

Remind me what competive CPUs did AMD have between AMD 64 and Ryzen. Remind me what made AMD almost sink itself by purchasing ATI which I still find to be a very questionable move because they have never truly realized the goals behind the merger (they wanted to create a Fusion APU which just never happened - we have something close to it in Xbox/SPS but that could have been done without this merger). You cannot blame everything on one company when the other company through mismanagement or dubious cash spendings is comming suicide (Intel was in the same boat with Brian Krzanich). I'm not defending Intel. I'm trying to defend healthy competition.

A) Intel was pulling that crap through Athlon 64 days.
B) I get the sense your only definition of healthy competion is an outright price war. $200 3600s isn't good enough, when you could get them for $160 for a time. And you don't seem to believe the demand pricing in-elasticity despite how quick hardware components are selling out, the lack of availability in all sorts of hardware all year, and the price creep in Ryzen 3000 SKUs since May.

It's not like I want to pay more but some people need to face certain realities. To some degree it seems more that demand for performance is driving prices up rather than competition driving price down. The 7700K was ~$350 inflation adjusted and the last CPU not under AMD threat. 8700K went up a bit, and Ryzen 1's had some premium to them for multithreaded performance and we've more or less stayed there. Raising the relative performance ceiling for the high-end is contributing.

 

[amrnuke]

Yeah, and I'm especially curious how Zen 3 has become so much faster at compressing using WinRAR - a performance increase is insane, over 65%. That is not reflected in this review, but here are the results of the integrated benchmark:

As someone else mentioned part of it is just increased load/store speed.

But from another compression/decompression-specific point of view, IIRC WinRAR is somewhat cache-dependent and someone single-threaded even though technically it is multithreaded. As you can see in various comparisons it scales more tightly with SPECint2006 than with SPECint2017 Rate-N

When comparing 10900K to 3950X and 5950X the key reasons Zen3 pull ahead are multiple, but I'd suspect a large core access to L3$ (Zen2 core = 16MB L3$ access, Comet Lake 10900K core = 20MB L3$ access, Zen3 core = 32MB L3$ access) and the load/store changes, faster recycling after branch mispredictions, and so on are all at play.

 

[CP5670]

I read the Anand review, quite impressive uplift across the board. It reminds me of the Core 2 or Athlon 64 launches back in the day. I'm especially surprised by how much faster it is in Civ 6. In FPS-type games there is much less difference though. If I was buying today I would definitely get a 5800X over a 10700K, but my existing 10700K will be fine for at least a few years.

 

[inf64]

I also wanted to add that it looks like Load/Store was one of the major bottlenecks in Zen2 design as we can see with Zen3 - like a cork has been pulled out from a fine bottle of wine , goodness flows in abundance
This bodes well for Zen4+ designs, on 5nm they can widen the core and expand all the key structures and maybe uncork another bottleneck in Zen3. DDR5 should be a major uplift as Zen3 looks to be severely memory BW bound in some multi threaded workloads.

 

[Adonisds]

Did someone take die shots?

What is the transistor count?

Did transistor density increase?

 

[Rigg]

This BIOS feature is a game changer for overclocking AMD. Especially x SKU's with their high SC boost. I really hope this becomes a standard bios feature. I'm hoping my crosshair 8 will get this feature in a BIOS update. This eliminates all of the risks and compromises of all core and CCX overclocking. I've been waiting for them to do something like this for a long time.

 

[KompuKare]

I don't see Flight Sim 2020. X-Plane is also heavy on the CPU and the results in the Phoronix review look outstanding:

AMD Ryzen 9 5900X/5950X Linux Gaming Performance Review - Phoronix

www.phoronix.com

Oops, you are right.
I though I had seen MS Flight there, but must have been mistaken. Due to the way they do their site for mobile, on the phone it wasn't obvious to see when replying.

 

[Veradun]

Remind me what competive CPUs did AMD have between AMD 64 and Ryzen. Remind me what made AMD almost sink itself by purchasing ATI which I still find to be a very questionable move because they have never truly realized the goals behind the merger (they wanted to create a Fusion APU which just never happened - we have something close to it in Xbox/SPS but that could have been done without this merger). You cannot blame everything on one company when the other company through mismanagement or dubious cash spendings is comming suicide (Intel was in the same boat with Brian Krzanich). I'm not defending Intel. I'm trying to defend healthy competition.

Nope.

We would have had nvidia and intel consoles otherwise. Instead it is a monopoly.

 

[tamz_msc]

Very impressive performance in StarCraft II.

Source(Russian):