Review Ryzen 7 9700X Reviews

[H T C]

The 9700X with PBO is comparable to the 7700X. The 9700X sans-PBO can be compared to the 7700. They're just comparing/equating the TDP values for each to measure the efficiency. Nothing out of sort there.

No it's not: it's comparable to the 7700X PBO.

Most, if not all, that turned PBO on with the 9700x did so with an unlimited power setting, which is not the same as a 7700x.

Exactly.

 

[jdubs03]

Most, if not all, that turned PBO on with the 9700x did so with an unlimited power setting, which is not the same as a 7700x.

Maybe it isn’t “exactly”, but in their charts they show the power consumption and while it is more than the 7700X it’s not massively higher.
And say you benchmark the 7700X with PBO. That is going to narrow the difference in performance (and power).

Performance:

AMD Ryzen 7 7700X Review - The Best Zen 4 for Gaming

With the Ryzen 7 7700X, AMD is introducing their fastest Zen 4 processor for gamers. In our review we found out that gaming on the 7700X runs better than the 7900X and 7950X, thanks to the single CCD design of the 7700X. Just like on other high-end AM5 CPUs, temperatures are a problem though.

www.techpowerup.com

AMD Ryzen 7 9700X Review - The Magic of Zen 5

The AMD Ryzen 7 9700X boasts a robust eight-core design and exceptional single-threaded performance with great power efficiency. In our review we saw good gains with PBO overclocking and TDP increases. Unfortunately, the $360 price tag is steep, and there are several good alternatives available.

www.techpowerup.com

Power consumption:

AMD Ryzen 7 7700X Review - The Best Zen 4 for Gaming

With the Ryzen 7 7700X, AMD is introducing their fastest Zen 4 processor for gamers. In our review we found out that gaming on the 7700X runs better than the 7900X and 7950X, thanks to the single CCD design of the 7700X. Just like on other high-end AM5 CPUs, temperatures are a problem though.

www.techpowerup.com

AMD Ryzen 7 9700X Review - The Magic of Zen 5

The AMD Ryzen 7 9700X boasts a robust eight-core design and exceptional single-threaded performance with great power efficiency. In our review we saw good gains with PBO overclocking and TDP increases. Unfortunately, the $360 price tag is steep, and there are several good alternatives available.

www.techpowerup.com

 

[Hitman928]

Maybe it isn’t “exactly”, but in their charts they show the power consumption and while it is more than the 7700X it’s not massively higher.
And say you benchmark the 7700X with PBO. That is going to narrow the difference in performance (and power).

Performance:

AMD Ryzen 7 7700X Review - The Best Zen 4 for Gaming

With the Ryzen 7 7700X, AMD is introducing their fastest Zen 4 processor for gamers. In our review we found out that gaming on the 7700X runs better than the 7900X and 7950X, thanks to the single CCD design of the 7700X. Just like on other high-end AM5 CPUs, temperatures are a problem though.

www.techpowerup.com

AMD Ryzen 7 9700X Review - The Magic of Zen 5

The AMD Ryzen 7 9700X boasts a robust eight-core design and exceptional single-threaded performance with great power efficiency. In our review we saw good gains with PBO overclocking and TDP increases. Unfortunately, the $360 price tag is steep, and there are several good alternatives available.

www.techpowerup.com

Power consumption:

AMD Ryzen 7 7700X Review - The Best Zen 4 for Gaming

With the Ryzen 7 7700X, AMD is introducing their fastest Zen 4 processor for gamers. In our review we found out that gaming on the 7700X runs better than the 7900X and 7950X, thanks to the single CCD design of the 7700X. Just like on other high-end AM5 CPUs, temperatures are a problem though.

www.techpowerup.com

AMD Ryzen 7 9700X Review - The Magic of Zen 5

The AMD Ryzen 7 9700X boasts a robust eight-core design and exceptional single-threaded performance with great power efficiency. In our review we saw good gains with PBO overclocking and TDP increases. Unfortunately, the $360 price tag is steep, and there are several good alternatives available.

www.techpowerup.com

PBO unlimited lets the 9700x use up to ~170 - 180 W for little benefit, while the 7700x is capped at 142 W. That’s 20 - 25% more power while being deep into diminishing returns. If you enabled unlimited power on the 7700x and tried to push that much wattage, you would close the power gap with almost no performance benefit, it will favor the 9700x for sure. Both should be tested at stock and within sane power limits if you want to compare efficiency, not pushing one or both to the limit where the power increases per clock increase explodes.

 

[Hitman928]

Thanks to @Geddagod for linking it in the other thread, here is what the power curve looks like between the 9700x and 7700x. At the 9700x's stock of 88 W, the efficiency difference isn't that large. At the 7700X's stock of 142 W, the efficiency improvement grows to ~15%. Testing the 9700x at ~170 W while keeping the 7700x at stock and making an efficiency comparison makes no sense and makes the 9700x look worse than it is.

 

[jdubs03]

PBO unlimited lets the 9700x use up to ~170 - 180 W for little benefit, while the 7700x is capped at 142 W. That’s 20 - 25% more power while being deep into diminishing returns. If you enabled unlimited power on the 7700x and tried to push that much wattage, you would close the power gap with almost no performance benefit, it will favor the 9700x for sure. Both should be tested at stock and within sane power limits if you want to compare efficiency, not pushing one or both to the limit where the power increases per clock increase explodes.

I agree about this. And that’s where the 7700X vs 9700X messes up the simple comparison because the differential between TDPs (in the reviews); and why the 7700 is the nearest easy SKU to compare against there.
That chart is great though, offers a good look at the difference.

I guess I’m just going about it the long way, to state that HBUX is being unfairly criticized for the results they’ve presented. I see it as them being pretty even handed showing the how each chip of these two most recent generations have performed. Particularly about gaming, 2 years between releases and barely any performance uplift is a poor showing; for more money to boot.

 

[H T C]

Thanks to @Geddagod for linking it in the other thread, here is what the power curve looks like between the 9700x and 7700x. At the 9700x's stock of 88 W, the efficiency difference isn't that large. At the 7700X's stock of 142 W, the efficiency improvement grows to ~15%. Testing the 9700x at ~170 W while keeping the 7700x at stock and making an efficiency comparison makes no sense and makes the 9700x look worse than it is.

Exactly.

 

[ondma]

The 9700X with PBO is comparable to the 7700X. The 9700X sans-PBO can be compared to the 7700. They're just comparing/equating the TDP values for each to measure the efficiency. Nothing out of sort there.


Might want to revisit that commentary once an actual ARL product is released.
Doesn't seem right to make comparisons of a released product, with one that isn’t and only has preliminary data about it.

Dont want to derail into intel in this thread, but I cant let it pass. After all the leaks and even Lunar Lake data, are there still people that think some magic is going to happen that turns ARL into some sort of home run?

 

[GaiaHunter]

TLDR - games like caches and clocks, no increase in those, no increase in gaming performance.

 

[coercitiv]

TLDR - games like caches and clocks, no increase in those, no increase in gaming performance.

Games like everything as long as it's faster than before, they are the ultimate bottleneck sniffer. The 12600K had lower clocks and same L3 as 10900K, yet had no issue pulling ahead in gaming.

 

[H433x0n]

Dont want to derail into intel in this thread, but I cant let it pass. After all the leaks and even Lunar Lake data, are there still people that think some magic is going to happen that turns ARL into some sort of home run?

The Lunar Lake data doesn’t look bad but I’d say people that expect magic from ARL are in the minority (at least on this board). It’s still worth waiting for the product to launch though before declaring anything DOA.

With the way things are going, the x86 ISA can’t afford a total dud from Intel or AMD going forward.

 

[jdubs03]

Dont want to derail into intel in this thread, but I cant let it pass. After all the leaks and even Lunar Lake data, are there still people that think some magic is going to happen that turns ARL into some sort of home run?

I’m not making any statement as to ARL being a home run. But you’re using a QS against a now released product. All I’m saying is, it’s not a fair comparison until ARL has been released and reviewed. Should be the standard for any product.

ARL could tread water so to speak and I’d say that’s a disappointment. If they lose the ST crown for now, that’s a disappointment. Same with MT, particularly if it’s still power hungry.

But it needs to be confirmed on release, not based on a still in-development sample.

 

[gdansk]

TLDR - games like caches and clocks, no increase in those, no increase in gaming performance.

L1d did increase to no impact in games, apparently. Zen 5 must have some limits that most games hit and I think it's the front end. It doesn't seem to work even in SPEC 1T and they would have used that when simulating/testing the design.

With the way things are going, the x86 ISA can’t afford a total dud from Intel or AMD going forward.

Yes, but I'll wait to see what Skymont's successor does before signing the death certificate.

 

[poke01]

Yes, but I'll wait to see what Skymont's successor does before signing the death certificate.

Same here, Intel did say IPC is going to be their focus instead of clocks. Let’s see.


As Intel tells us: “If you have really high frequency as your model then it takes a lot more time so we have been shifting our time towards the microarchitecture because that’s where you get a lot more power efficiency.

“This is one of our first major steps for Lion Cove which is our CPU here, Skymont E-Core [for Arrow Lake / Lunar Lake], every generation after [Panther Lake onwards] is the same thing, go after IPC, IPC, IPC.”

Panther Lake CPUs are officially alive – and Intel has shared its plan to beat AMD with future CPUs

Panther Lake chips, that’ll follow next-gen Arrow Lake, remain on track for 2025

www.techradar.com

 

[DAPUNISHER]

Let's get this train back on the rails - 8 pack does some overclocking of his own.

 

[GaiaHunter]

Games like everything as long as it's faster than before, they are the ultimate bottleneck sniffer. The 12600K had lower clocks and same L3 as 10900K, yet had no issue pulling ahead in gaming.

They do not care for stuff that doesn't increase gaming performance...

Up until now, every new zen generation, had at least clocks increase and/or latency reductions.

In this case the stuff added doesn't impact gaming performance, but does add to other workloads.

 

[DAPUNISHER]

9700X is ranked 42 (the answer to everything. Except good sales) on the Amazon U.S. best sellers list. Talk about a failed launch; oof.

 

[Ranulf]

Let's get this train back on the rails - 8 pack does some overclocking of his own.

Eh, more of the same really. More data for stock vs PBO vs custom tuning at least. Similar data others had while comparing the 9700x to the 7700x.

 

[DAPUNISHER]

Eh, more of the same really. More data for stock vs PBO vs custom tuning at least. Similar data others had while comparing the 9700x to the 7700x.

Peer review, reproduceable results...

Most importantly, on topic.

 

[Ranulf]

HUB on SMT on vs off, "isn't a silver bullet":

https://twitter.com/x/status/1823234152558440685

 

[CouncilorIrissa]

HUB on SMT on vs off, "isn't a silver bullet":

https://twitter.com/x/status/1823234152558440685

He's missing the point. The point is not to show the CPU in the best possible light, the point is to analyse how the new front-end behaves in light of Clark's statements. Even if the performance differential is small, it's still worth doing. Comparing it to Zen 2 is false equivalence because Zen 5 is a big departure from previous Zen uarches, it's the same thing in name only.

Very surface-level take from Steve.

 

[Hitman928]

He's missing the point. The point is not to show the CPU in the best possible light, the point is to analyse how the new front-end behaves in light of Clark's statements. Even if the performance differential is small, it's still worth doing. Comparing it to Zen 2 is false equivalence because Zen 5 is a big departure from previous Zen uarches, it's the same thing in name only.

Very surface-level take from Steve.

I like HWUB, but they are not the place to go to for any kind of technical detail or analysis.

 

[Nothingness]

I like HWUB, but they are not the place to go to for any kind of technical detail or analysis.

That goes beyond not being technical, their claim just defies any logical thought.

 

[CouncilorIrissa]

I like HWUB, but they are not the place to go to for any kind of technical detail or analysis.

I do as well, but he's wrong about this particular topic. Asking to benchmark SMT on/off is perfectly reasonable given the modern x86 CPU landscape, where Intel are ditching SMT on client and AMD are doing exactly the opposite, dedicating more resources than ever.

edit: Steve's whole niche in PC benchmarking is providing more data points than other reviewers do, he's completely fine with doing 50 GPU benchmark comparison videos. He's essentially asked to milk this launch a little more by his own viewers and he's dismissing the topic as case closed while having no basis whatsoever.

 

[coercitiv]

Very surface-level take from Steve.

I like HWUB, but they are not the place to go to for any kind of technical detail or analysis.

You should consider their PoV as well, the testing they do is aimed at answering a very particular question for the average consumer: is product X good for gaming and/or productivity? This is where it all begins and ends for them, and even the dialogue they're having with the channel's audience is centered around this simplified way of looking at things. We may not like their answer, but they're probably answering a different question than we think was asked.

So I fully agree with @Hitman928 , they're not the people we need to go for answers on technical details and analysis. Even if they were, they would be drowned in noise by their audience seeking a different chatter.

Asking to benchmark SMT on/off is perfectly reasonable given the modern x86 CPU landscape, where Intel are ditching SMT on client and AMD are doing exactly the opposite, dedicating more resources than ever.

When ARL comes in, this will likely happen. Until then though, HUB is under siege for talking bad about Zen 5. Their stance of the SMT subject and everything Zen 5 related is probably defensive at this point.

 

[CouncilorIrissa]

You should consider their PoV as well, the testing they do is aimed at answering a very particular question for the average consumer: is product X good for gaming and/or productivity? This is where it all begins and ends for them, and even the dialogue they're having with the channel's audience is centered around this simplified way of looking at things. We may not like their answer, but they're probably answering a different question than we think was asked.

I get that, I just disagree with his reasoning to refuse making an SMT on/off video based on false equivalence. He should've simply said "We aren't going to test this because switching off SMT is something 99% of our audience won't consider" and be done with it.

When ARL comes in, this will likely happen. Until then though, HUB is under siege for talking bad about Zen 5. Their stance of the SMT subject and everything Zen 5 related is probably defensive at this point.

Some especially rabid fans might actually call off the dogs if he came up with a short video showcasing a 2.5% increase in gaming performance in 7 titles. But I do get that producing videos also takes time and not enough people would likely watch this it to justify the effort.

Oh well.