Review AMD Ryzen 5 9600X and Ryzen 7 9700X Review: Zen 5 brings stellar gaming performance

[bit_user]

Im saying at any wattage, whether it's at the 170 PBO max or at the stock 88w, any intel chip beats it. By a lot.

What you're proposing is that someone would buy an i7-1x700K and run it at the same Watts as a 9700X. This is unrealistic.

Man I honestly feel like you are trolling me.

No, just trying to hold you accountable.

You've not been introducing new information into the discussion, for a while. At this point, you're just trying to control the narrative. You've made your points numerous times. It's very clear to everyone who's been following the thread what you think. The only point of continuing to make these claims is to try and drown out others.

It's literally one of the most inefficient CPUs for the price in MT performance.

See, you just keep cooking up these whoppers, so matter how many times and how conclusively they're proven to be false. The only way you can make them true is by tweaking one CPU to run at non-stock settings that either make the AMD CPU run less efficiently or make the Intel CPU run more efficiently. If this were an honest contest of optimal efficiency, why wouldn't you tweak both CPUs to run respectively at their most efficient settings?

It's like you just sit around trying to dream up scenarios that make Intel look more efficient and then try to come up with some kind of justification for it. This level of fakery puts to shame spongiemaster's complaints about 720p gaming benchmarks!

Nobody goes about buying a CPU by saying "I need exactly a certain number of TFLOPS; therefore I must boost or throttle my CPU to that performance level". The kind of informed CPUs buyers we care about are weighing all of a CPUs pros and cons against their particular needs and making the choice that's the right fit for them. Rather than run any CPU at iso-performance to some other CPU, they're either going to run it at stock, motherboard defaults, or whatever custom settings best suit their needs. And motherboards don't default PBO to "on".

 

[TheHerald]

What you're proposing is that someone would buy an i7-1x700K and run it at the same Watts as a 9700X. This is unrealistic.

Why is it unrealistic? If you are after efficiency why wouldn't you do just that? What's unrealistic is saying someone cares about MT efficiency and ends up buying a 9700x. That's what's unrealistic.

See, you just keep cooking up these whoppers, so matter how many times and how conclusively they're proven to be false. The only way you can make them true is by tweaking one CPU to run at non-stock settings that either make the AMD CPU run less efficiently or make the Intel CPU run more efficiently. If this were an honest contest of optimal efficiency, why wouldn't you tweak both CPUs to run respectively at their most efficient settings?

By setting up the CPUs at the same power you do exactly that, you tweak both CPUs to run at their efficient settings. Neither is running at "their most efficient" setting at 88w so your point is moot.

It's like you just sit around trying to dream up scenarios that make Intel look more efficient and then try to come up with some kind of justification for it. This level of fakery puts to shame spongiemaster's complaints about 720p gaming benchmarks!

Nobody goes about buying a CPU by saying "I need exactly a certain number of TFLOPS; therefore I must boost or throttle my CPU to that performance level". The kind of informed CPUs buyers we care about are weighing all of a CPUs pros and cons against their particular needs and making the choice that's the right fit for them. Rather than run any CPU at iso-performance to some other CPU, they're either going to run it at stock, motherboard defaults, or whatever custom settings best suit their needs. And motherboards don't default PBO to "on".

Oh I agree. That's exactly why I think anyone that actually cares about MT efficiency wouldn't be going near any of those midrange AMD chips.

Let me make it simple for you. A friend of yours asks you what CPU to buy - at 300-350$ - he cares about MT efficiency. The answer to that question has to be anything but the 9700x. Get a 7900, or a 7900x and power limit it, or a 13700k / 14700k and power limit. That's the only answer that makes sense. Suggesting the 9700x does not. Now if he doesn't want or doesn't know what a bios is and where to find that, then sure, your suggestion can be the 9700x, but at the end of the day you are suggesting a subpar product for the convenience factor which is what im trying to get across.

Just remove intel from the conversation cause obviously it triggers you, don't you think the 7950x is a better suggestion than the 5950x for someone that cares about efficiency? Does it make any sense to buy the 5950x instead of just spending 5 seconds in the bios to power limit the 7950x? Do you actually consider the 5950x more efficient than the 7950x? Cause that's just crazy to me.

 

[bit_user]

By setting up the CPUs at the same power you do exactly that,

It's a false experiment. People buying a i7-1x700K don't set its TDP to 88W. You picked an artificial configuration for that CPU, just because it happens to match the stock configuration of the 9700X. Why not pick a more efficient configuration for the 9700X? Why should it have to remain at stock?

You set these artificial goal posts that don't match what people do, in the real world. That's why almost no professional reviewers test CPUs this way. Sure, you can find a couple cases where it's been done, but if it were truly as important as you seem to think, then it would be standard practice.

That's exactly why I think anyone that actually cares about MT efficiency wouldn't be going near any of those midrange AMD chips.

That's a false dichotomy. You're presuming they want MT efficiency at max performance. Depending on what they're doing, they might be fine with stock performance. If it were me, I'd be running it at or near stock.

It's a weird presumption, because if someone cares about MT efficiency at max performance, then pretty much any CPU they get would be cranked up well beyond the range where it runs efficiently.

Let me make it simple for you. A friend of yours asks you what CPU to buy - at 300-350$ - he cares about MT efficiency.

That's an unrealistic example - it's too simplistic. As I already explained, "MT efficiency" is underspecified. There should be a performance requirement/desire on some class of workloads. Also, whether or not the CPU is used for gaming or other performance-intensive, lightly-threaded tasks and how important performance on those is to the user.

 

[TheHerald]

It's a false experiment. People buying a i7-1x700K don't set its TDP to 88W.

Agree, cause most people don't care about efficiency that much.

The point isn't that someone will buy a 7900x or a 13700k and se them to 88w. The point is that no one will buy the 9700x for it's MT performance or efficiency, cause it has neither. Testing all CPUs at ISO power isn't because "thats how people will use them". It's to show that they shouldn't buy the 9700x for efficiency since the cpus it competes against are better at that.

That's an unrealistic example - it's too simplistic. As I already explained, "MT efficiency" is underspecified. There should be a performance requirement/desire on some class of workloads. Also, whether or not the CPU is used for gaming or other performance-intensive, lightly-threaded tasks and how important performance on those is to the user.

Sure. I want a CPU that can do around 20k -25k - 30k in CBR23 at the least amount of wattage possible. Budget is around 350$. Is the 9700x a consideration in either of these 3 scenarios? No, cause it needs a truckload of wattage compared to any other cpu at it's price range to his those numbers, or it just doesn't hit them at all.

Man if you are suggesting that the 9700x is good in either flat out MT performance or MT efficiency then in my opinion you are just flat out wrong and there is no point talkng about it any further. Go ahead, have the last word, this is boring.

 

[bit_user]

Sure. I want a CPU that can do around 20k -25k - 30k in CBR23 at the least amount of wattage possible. Budget is around 350$.

Now you're overconstraining it. You're fishing for an exact answer. That's obvious and makes the question neither very interesting nor very useful.

A more worthwhile exercise would be to pose the question like this: Under what circumstances would you recommend an i7-13700K? Given: CPU budget of $360, but no prior motherboard or RAM on hand. To that, I'd answer:

Someone doing lots of rendering or who compiles large software packages somewhat frequently should get the i7-13700K, provided they don't mind its ST performance could be around 7.4% lower.
​

For someone concerned about gaming, but not MT performance, the R7 7800X3D is a tempting option. But, for those who really value single-threaded performance over and above gaming (i.e. maybe they game at higher resolutions or don't have the fastest dGPU), then that might be where the R7 9700X fits best. Its CB24 ST performance is 18.4% better than the 7800X3D!

 

[YSCCC]

Now you're overconstraining it. You're fishing for an exact answer. That's obvious and makes the question neither very interesting nor very useful.

A more worthwhile exercise would be to pose the question like this: Under what circumstances would you recommend an i7-13700K? Given: CPU budget of $360, but no prior motherboard or RAM on hand. To that, I'd answer:

​

Someone doing lots of rendering or who compiles large software packages somewhat frequently should get the i7-13700K, provided they don't mind its ST performance could be around 7.4% lower.​

​

For someone concerned about gaming, but not MT performance, the R7 7800X3D is a tempting option. But, for those who really value single-threaded performance over and above gaming (i.e. maybe they game at higher resolutions or don't have the fastest dGPU), then that might be where the R7 9700X fits best. Its CB24 ST performance is 18.4% better than the 7800X3D!

seriously, just let him speak to himself and nobody will care the crazy narrative...

ppl compare the products stock, is because that is what guaranteed spec in the price point that one will have, making crazy arguement on soon to discontinue products on value or trying to nerf the very expensive TOTL ones just to compete in efficiceny, or suggesting a flawed design needing tons of user tweak just to survive is ok is as genius as it can get.

Reviews being reviews is to test the market segmentation of the vendor with the price and/or performance they are targetting. The Zen 5 lower end don't seem to be eye openingly good and not really bad, market reaction will tell if it's really a flop or a success.

And in reality we know that let the strange behaving troll to die off is the best way to not keep any discussion contaminated, so time to claim down and enjoy life, I am gaming with some snacks to see how the discussion goes along

 

[raminpro]

why youtuber like gamernexus , hardwareunbox , jay dont test this powerfull cpu with PBO (turbo) and test cpu in base 4.40ghz ????

in PBO 9700x beat 14600k-14700k-14900k in many games like cyberpank

why they test in minimum ???
this show their knoledge about CPU is 0.00

check techpowerup again , they change their review 1 hour ago (enable pbo)

 

[jeremyj_83]

You can do that now. However, I expect that as the 9700X's price settles, we'll see it slide below the i7-13700K's price. Once we're talking about matching it against an i5, your assertion will no longer hold. That's obvious, given how close the MT performance of the 9700X is to the i5, at its stock power.

That is what matters. My comment #219 gets into that. Basing a comparison of products on arbitrary name i7 vs Ryzen 7 is meaningless, especially since Intel changed their core counts with the hybrid design and pricing doesn't match. When compared on total threads available, the Ryzen of the same thread count usually is faster. Not surprisingly the Intel and AMD chips of equal thread counts are also about equal on price. Due to this we know the Zen 5 chips will come down in price soon. The launch price of the R5 & R7 is too high compared to the target competition. Right now the early adapters are being targeted so they will pay more. In a couple months pricing will reflect the actual target of the i5s. One other thing is the overall performance for the R7 (not just cherry picking applications that load all threads) is equal to very slightly higher than the 14600k despite the total threads disadvantage.

 

[-Fran-]

But what about the effectiveness of the memory controller in Zen 5 compared to Zen 4?
Folks who purchase a 9700X will more than likely be enthusiastic gamers (who don't already have a 7800X3D) or/& are over clockers looking to get every single bit of performance out of their AM5 platform.
If they follow guides such as what SkatterBencher published for his adventure into 9700X, they will find new found areas of performance not seen in 8c/16t chip with relatively low power consumption compared to Intel. Although he uses ASUS ROG Crosshair X670E Hero in his venture, much if not 99% of the settings will be applicable in good quality mid to high end AM5 boards anyway.
Paying for the "X" in an AMD chip assures more than likely the end user will have better quality silicon to indulge in OC affairs.

The I/O die in the Zen5 package is the same exact one as in the Zen4 package (remember, chiplets), so it is to be understood the IMC (which lives in the I/O die) is not going to be radically different in terms of capabilities. This being said, it's still early days for Zen5 and AGESA to further tweak the microcode in hopes to extract a bit more out of it. Not that it'll change much, but maybe they can squeeze a few more percentage points. Also, keep in mind AMD's own recommendation is to use 6000MT/s as that hits the sweet spot for IF and memory clocks most of the time 2:3:3 ratio.

As for the "X". Binning has always been a thing, but luck of the draw goes both ways: you can still get a very good OC'er in a non-X CPU from AMD. Specially when the process is as mature as 5/4nm is now.

And as for the CPU itself. Much like when Zen4 launched, I have the same feeling as back then on this one: AMD doesn't have a performance issue, but a price issue. That is resolved with some time in the market. The truth of the matter here is: anyone looking to have a decent gaming experience WILL BE ABSOLUTELY OK with a R5 7600, i5 12600K and up, from both camps. So you just need to get what is cheaper and splurge on the GPU if you care about games first and foremost.

Regards.

 

[-Fran-]

I need to spend some more time digging into the meat of that article, because all of the benchmarks I looked at (CB ST, CB MT, MP3 (ST), Gaming, Apps) are in line with what I would've expected. So, I'm not quite sure what they're on about.

The needle moves, but it's still very little; close to a nothing-burger. I think the interesting finding there is how SMT affect AMD very little overall and the 7800X3D doesn't even care about it, which is funny.

Regards.

 

[jeremyj_83]

Also, keep in mind AMD's own recommendation is to use 6000MT/s as that hits the sweet spot for IF and memory clocks most of the time 2:3:3 ratio.

I'll be intered to see if someone does a Zen 5 RAM speed scaling review.

 

[jeremyj_83]

The needle moves, but it's still very little; close to a nothing-burger. I think the interesting finding there is how SMT affect AMD very little overall and the 7800X3D doesn't even care about it, which is funny.

Regards.

Gaming without SMT helped the Zen 5 but most other threaded applications saw a performance decline without SMT. Since the gaming performance issue seems to be scheduler related, I wonder is MS will be able to fix that with a Win Update for Win11. Or we will have to wait for Win12 to have the fix.

 

[-Fran-]

Gaming without SMT helped the Zen 5 but most other threaded applications saw a performance decline without SMT. Since the gaming performance issue seems to be scheduler related, I wonder is MS will be able to fix that with a Win Update for Win11. Or we will have to wait for Win12 to have the fix.

The underlying implication was "in low threaded apps the impact of removing SMT was margnial or insignificant".

Regards.

 

[AezielOfficiel]

I'd like to see you guys try a test with Star Citizen vs the 7800x3D and 7700x and the new 9800x. Star citizen is very math intensive with very heavy physics. It's a very cpu bottlenecked game because of it. I'm wondering if the new parallel floating point pipelines in the 9000 series will really help.

Benchmark on flight simulator work too in cpu dependant type game.

 

[YSCCC]

Benchmark on flight simulator work too in cpu dependant type game.

Likely, I am playing MSFS 2020 for years using the 12700KF and 14900k did improve the bottleneck a bit, that seems work better on 7800X3D according to a friend, but tbf, at current condition where intel have been in a big mess AMD basically is in a no competition mode

 

[bit_user]

The needle moves, but it's still very little; close to a nothing-burger.

Well, I spent a little more time looking at it, and what's intriguing about their data is how it shows opposite effects between Zen 4 vs. Zen 5 CPUs. That does indeed suggest something funky could be going on with the Windows scheduler doubling-up threads while other cores sit idle.

The magnitude of the effect isn't huge. However, it should be multiplicative with the speedups already observed between Ryzen 9000 CPUs and other CPUs. So, it definitely seems like it could account for some of the "missing performance" people have been complaining about.

 

[bit_user]

Gaming without SMT helped the Zen 5

From what I've heard, many games also perform better on recent Intel CPUs with HT disabled. That's why I wasn't surprised to see that effect on Ryzen 9000. However, what's strange is how games on Ryzen 7000 seem to benefit from SMT, overall.

most other threaded applications saw a performance decline without SMT.

Yeah, but if they're heavily-threaded, then windows scheduler doesn't really pose the same kind of liability with SMT-enabled, because the normal state for such apps is to fully-load all cores, anyhow.

Since the gaming performance issue seems to be scheduler related, I wonder is MS will be able to fix that with a Win Update for Win11. Or we will have to wait for Win12 to have the fix.

If I were AMD, I'd be screaming "bloody murder", if MS dragged their feet on a fix until Win 12. However, I'm not sure it's even MS' bug. Maybe it's a bug in one of the drivers AMD provides, wherein the different "cores" are mislabeled.

 

[jeremyj_83]

If I were AMD, I'd be screaming "bloody murder", if MS dragged their feet on a fix until Win 12. However, I'm not sure it's even MS' bug. Maybe it's a bug in one of the drivers AMD provides, wherein the different "cores" are mislabeled.

Either way hopefully the fix is easy and rolled out soon.

 

[-Fran-]

From what I've heard, many games also perform better on recent Intel CPUs with HT disabled. That's why I wasn't surprised to see that effect on Ryzen 9000. However, what's strange is how games on Ryzen 7000 seem to benefit from SMT, overall.


Yeah, but if they're heavily-threaded, then windows scheduler doesn't really pose the same kind of liability with SMT-enabled, because the normal state for such apps is to fully-load all cores, anyhow.


If I were AMD, I'd be screaming "bloody murder", if MS dragged their feet on a fix until Win 12. However, I'm not sure it's even MS' bug. Maybe it's a bug in one of the drivers AMD provides, wherein the different "cores" are mislabeled.

I'd say it's more AMD's fault than MS on this particular instance, if the case is they didn't patch the Windows Kernel correctly (as it seems in Linux it's a completely different CPU!). On the upside, again if it's a Kernel problem, it should be fix-able via a Windows Update or a driver update/patch (like they did back in the day with Athlon X2).

Also, maybe they removed logic that allowed their previous SMT implementation to be better in order to save space or have better throughput? Pseudo-proof of that is the improved cache subsystem (from what I see; look at MS FlySim numbers and other cache-intensive apps). Re-arranging the AVX512 pipeline must have side-effects and this could be one of them? Until Cheese and Chips doesn't do a deep dive, this is the best I got as high level speculation.

Regards.

 

[jeremyj_83]

I'd say it's more AMD's fault than MS on this particular instance, if the case is they didn't patch the Windows Kernel correctly (as it seems in Linux it's a completely different CPU!). On the upside, again if it's a Kernel problem, it should be fix-able via a Windows Update or a driver update/patch (like they did back in the day with Athlon X2).

Also, maybe they removed logic that allowed their previous SMT implementation to be better in order to save space or have better throughput? Pseudo-proof of that is the improved cache subsystem (from what I see; look at MS FlySim numbers and other cache-intensive apps). Re-arranging the AVX512 pipeline must have side-effects and this could be one of them? Until Cheese and Chips doesn't do a deep dive, this is the best I got as high level speculation.

Regards.

Or is it a combination of the double pumped branch predictor and MS scheduler not playing as well together.

 

[bit_user]

Or is it a combination of the double pumped branch predictor and MS scheduler not playing as well together.

LOL. You're joking, I hope?

 

[King_V]

No.

In other words, I don't care what he thinks or wants. He's behaving like people here are his towel boys or something.

I also don't care what you think.

I know that Zen 5 sucks for desktop, because the evidence of it was plentiful even before its release, and it has since not been debunked. It has rather been very much confirmed. I have stated my thoughts on why it ended up being so.

And what you said about evidence and proof may apply in a court of law, but if you look around, I'm sure you will notice that we are not in one. This is an Internet forum, sir.

And on this one, in other words: "I can make any unsubstantiated claim I want."

Don't do that. If you're going to make a claim, prove it. Don't pull this "you find my evidence for me" nonsense.

 

[thestryker]

Yeah, let's imagine AMD did us wrong and then get upset at them for it.
🙄

I was referencing my earlier post where I used that same verbiage is all 🤣

Here are a few reasons why they probably did it:

1. PBO barely affects single-threaded or gaming performance, hence a higher TDP wouldn't be relevant to many of their customers.
   
   Spoiler

   
   Spoiler

   
2. Feedback from OEMs might've suggested the 7700X was too expensive to cool for its value proposition. Or, maybe AMD was just looking to make the 9700X a more economical option for them.
3. Variance in die quality could mean they can't guarantee all CPUs will overclock as well as the review samples have.

I think you really only need to see the data supporting point #1 and consider that more TDP isn't exactly "free" (see point #2), in order to understand why it doesn't ship with a higher TDP. Point #3 was speculative, but since we're speculating...

Oh I definitely don't think it should have been even a 105W part. The available evidence certainly supports that if it had been in the 75-85W range it would have been better for consumers.

 

[thestryker]

But the 9700X is faster! Its stock CB R24 ST is 16.4% faster than 7700. In the MT case, the difference is 9.2%. In gaming, it's 5.1% faster. Is that worth nothing?

Keep in mind if you're looking at Anandtech/Tom's they're using AMD memory specifications rather than the "recommended" from the review guide. TPU uses like for like memory configurations on Zen 4 and 5.

I need to spend some more time digging into the meat of that article, because all of the benchmarks I looked at (CB ST, CB MT, MP3 (ST), Gaming, Apps) are in line with what I would've expected. So, I'm not quite sure what they're on about.

After reading through that article I think the only thing notable is that there's a proportionally bigger uplift for turning it off on Zen 5 than there was with Zen 4 even though the absolute numbers are still low. I felt like this was sensational headline at work even though the information was very interesting.

 

[bit_user]

Keep in mind if you're looking at Anandtech/Tom's they're using AMD memory specifications rather than the "recommended" from the review guide. TPU uses like for like memory configurations on Zen 4 and 5.

I was looking at the TechPowerUp review of the 9700X.