AMD Ryzen 7 9700X

[Vincero]

It isn't. At least not in gaming. Other sites like German ComputerBase have tested this and upped the power limit from 88W (default 65W TDP) to 142W (same PPT as the 7700X). The 9700X achieved +1% better results on average (+3% best case), i.e. at least for games it is completely pointless to run this CPU at a higher power target.

That's disappointing in the sense that it would seem that there isn't much more headroom - either there is some other inefficiency or it looks like 4nm vs 5nm is really a minor bump / offers no extra headroom.

 

[evernessince]

I have been reading a few reviews across the internet and I think alot of people are missing what makes these chips interesting. In its stock state these things sip power especially compared to the Intel equivalent. Its not going to beat the 7800X3D in gaming but we knew that, its not crammed with cores for heavy multi-tasking, and its not even the top end part. Realistically we need to see what the top end does (Ryzen 9) to see how those hold when the limits are removed.

I am curious how the Ryzen 9's will do.

Look at the efficiency charts in the review, it's not better than the 7700 in terms of efficiency which means any power limited 7000 series CPU (which you can do by changing a single setting the BIOS) is going to be comparable.

 

[RandallFlagg]

Why would anyone buy 9700x when he can get 7800X3D for 350€/$?

Pricing makes no sense. It should cost 300 bucks max imho.

B/C a lot of people prefer a rig that is good at everything, not just a one trick pony for gaming - a performance boost which itself is irrelevant unless you have a 4090.
The 7800X3D is a very poor performer outside of games, and requires a high end GPU to show a notable benefit in games.

 

[noel_fs]

8 cores for 360$ thats a nice joke

 

[regs]

They also could add 5 or 4 Zen 5c cores at the top of logic chip, by moving logic chip little closer to the north bridge chip. There seem to be enough space.

 

[RedelZaVedno]

B/C a lot of people prefer a rig that is good at everything, not just a one trick pony for gaming - a performance boost which itself is irrelevant unless you have a 4090.
The 7800X3D is a very poor performer outside of games, and requires a high end GPU to show a notable benefit in games.

Lots of cache really makes a huge difference when it comes to microstutters. I went from 5600X to 5800X3D and now 7800X3D. All I can say as a VR flight simmer these were all huge jumps. Stable 0.1%/1% lows are all that really matters when it comes to VR and 7800X3D delivers big time on this front. Even on my mediocre 4070TI Super GPU. I went from stuttering 15-30fps 0.1% lows on 5700X to 45-55fps (5800X3D) and now above 60 fps 0.1% lows on 7800X3D in MS FS2020 (Varjo Aero native res). This means going from hardly playable sim to being quite enjoyable. It looks like 9700X would be a backstep in this regard.

 

[RogueSix]

Its not going to beat the 7800X3D in gaming but we knew that...

Actually, we didn't. Yes, AMD first claimed that 7800X3D would maintain a marginal lead but later on, at a tech event in July, AMD stated that 9700X would be 2% faster than 7800X3D.

AMD outright lied, especially if you look at the benchmarks on other sites. TPU is by far the only website that shows 9700X on par with the 7000 X3D CPUs. Just about every single other site shows that 9700X is approximately 15% to 20% behind the X3Ds in gaming. If anyone can find another site with even remotely the same results as TPU then please post it here.

That aside, AMD outright lied in their Zen 5 reveal in general. If you look at the slides from back then and now at the real world performance: Nope. Not sure what happened between now and then but "good guy AMD" have lost a ton of credibility right there.

It makes me wonder if technical issues might be a factor here because, according to ComputerBase, there was quite the commotion going on behind the scenes prior to launch. AMD had to release five new BIOS revisions within a single week to address certain memory compatibility problems. CB are saying that a July 31st launch, as originally planned, would have ended up in a "disaster" and that this was the most "bumpy" launch they have witnessed from AMD in a long time.
The frequent new BIOS revisions and the feedback from testers around the world are supposed to have mostly fixed the issues by now but CB is expecting more frequent patching to come during Zen 5 early days.

 

[persondb]

Come on 15-16% IPC >>>> 3-4% from way back when zen+ launched.

The main reason this is looking bleak is that zen 3 - zen4 had IPC+high clock speeds!

It`s more like 5% to 8% Performance increase or so. The 15=16% are biased from the SIMD and Vector heavy stuff that is not very common for normal users.

 

[jpvalverde85]

The good sides are cheaper and less power (than 7700X), everything else unfortunately is meh, the 7800X3D still beats it everywhere. For such a notable frontend change a lot of workloads don't get a significant IPC change, and the double-wide FP crunching facility per core barely shows up and never with twice the speed, maybe there is some bottleneck inside the cores.

 

[AMDK11]

You'll wrong. Just some works on frontend.

In every area of the core, Zen5 has been redesigned. including a unified schedule for 6 ALU ports. Rebuilt and expanded FPU block. The frontend has dual prediction and prefetching with 2x 4 (8-wide) instruction decoding instead of 4-wide as in Zen 4. These are just first examples, without going into further details.

Dual decoder and predictor can be a problem because it behaves like a Zen4 width (wide decoder 4) in the worst case.

I wonder how a single 8-Wide decoder will behave in LionCove with ArrowLake-S.

 

[R0H1T]

The 15=16% are biased from the SIMD and Vector heavy stuff that is not very common for normal users.

It's 13-14% integer & 25-26% on FP based on Spec2017 ~

And then you have 387 tests from phoronix! Come on enough with the exaggerations

 

[AnotherReader]

The good sides are cheaper and less power (than 7700X), everything else unfortunately is meh, the 7800X3D still beats it everywhere. For such a notable frontend change a lot of workloads don't get a significant IPC change, and the double-wide FP crunching facility per core barely shows up and never with twice the speed, maybe there is some bottleneck inside the cores.

The doubled throughput for AVX-512 has been confirmed by actual benchmarks. This is what the developer of y-cruncher has to say about the increase in AVX-512 throughput:

or y-cruncher (which I'm the developer of):

• The regular Pi benchmarks and computations gain almost nothing (1-3%) on Zen5 due to being bottlenecked by memory bandwidth.
• If you run single-threaded, you remove the memory bottleneck and get a ~50% IPC improvement on Zen5 thanks to AVX512. (less than 2x due to Amdahl's Law)
• y-cruncher's BBP test (now a benchmark) shows 98% IPC improvement due to the pure AVX512 without any memory access.

In fact, the AVX512 improvement on Zen5 created a memory bottleneck so large that it became the primary reason why I promoted the BBP mini-program from a tool for verifying Pi records to a formal benchmark. The regular benchmarks wouldn't do Zen5 (and future processors) any justice. At least until someone can figure out how to get DDR5-20000 on AM5...

 

[Super XP]

Performance seems pretty piss poor for a 'next gen' chip

ZEN 6 is supposed to be that "Generational Chip". Not ZEN 5. But nevertheless people can upgrade from ZEN 1-2-3 with ZEN 5. Those on ZEN 4 I would probably wait till ZEN 6 for that next gen performance increase.

8 cores for 360$ thats a nice joke

At least it won't die on you like the Intel chips. What even more maddening is how Intel is still in denial mode lol

 

[Frick]

Those were ~2 years ago, when you could get one below $200. I grabbed mine for $179 on sale.

Now they're almost back up to launch MSRP, and with ryzen 5000 production lo longer the priority, I wouldnt count on it getting cheaper again.

If you see it or the 5700x3d anywhere near $200, grab it immediately.

They've never been that cheap here (they did dip to around €330) and anyway they released april 2022 so I don't know what you're on about.

Also by affordable I mean like €150 or so.

 

[Zach_01]

At least in blender has the efficiency top spot.



Here the 7700X at default and at 86W limit against 9700X
AMD really gone for efficiency and lower temps here leaving performance get away.



I know many will think that while any IPC gain has been lost in the lower power target of the CPU, this can be "fixed" with just an extended power limit.
I would say think again. Probably (most likely) AMD has binned these CPUs to perform on this specific power target.
If so, then there are little to be gained by extending power limit.

Dont know what can be done here through manual advanced PBO settings and CurveOptimizer, but if anything cannot be claimed as the default settings.
Most users dont go near those settings.

They've never been that cheap here (they did dip to around €330) and anyway they released april 2022 so I don't know what you're on about.

Same here. 5800X3D never gone under 270€.
Now the are back at 320~330€

Last 2 years of pricing

 

[Dr. Dro]

It does make the difference
From 9:23

Interesting. But it's a very nominal improvement over stock considered that in Roman's testing, it hit around 170 watts at the PBO Max configuration running R23. However, even in this situation, looks like it still loses to Raptor Lake's P-cores as long as AVX-512 is not involved. In Roman's testing he got 22026 multi and 2057 points single, well, just for you, BenchMate'd so you have insurance I'm not BSing you. This is the stock clock configuration with Intel's latest power spec, with my cooling on automatic (quiet). I can probably extract more out of it if I do some careful configuration.

I believe the 65 W spec (88 W PPT) is the sweet spot and a home run for AMD here. The power consumption and heat generation in this setting for the performance extracted are simply amazing, and while the processor might not be the absolute fastest, that efficiency is simply awesome. I just have to recognize that.

 

[R0H1T]

Same here. 5800X3D never gone under 270€.
Now the are back at 320~330€

There was a dip to around ~190 or so, why don't try some price tracker sites like camelcamel or whatever works over there?

 

[Craptacular]

And yet they are still loosing to the 7700 and 7800X3D in power efficiency. It's only better than 7000 series chips that are pushed way out of their sweet spot.

Either they have made very little progress in terms of efficiency or none at all.

The same could be said about the processor as a whole, not much worth mentioning. As other's have stated this seems more like a Zen+ than a new architecture.

For AMD's sake this processor better be laying the groundwork for a large next gen uplift because in every other metric it's an abject failure.

That is one way of looking at it, the other way is to look at the 7700x power consumption. Where it is 25 watts less, which is a 29% reduction in power consumption. The 9700x replaces the 7700x and as we can see the only big improvement was the power consumption. Same thing with the 9600x, it is replacing the 7600x, that is 21% reduction in power consumption. If AMD does in fact ever release a 9700 non-x and 9600 non-x then it will be interesting to see the power consumption. Same is true for the 9800x3d.

 

[jpvalverde85]

That is one way of looking at it, the other way is to look at the 7700x power consumption. Where it is 25 watts less, which is a 29% reduction in power consumption. The 9700x replaces the 7700x and as we can see the only big improvement was the power consumption. Same thing with the 9600x, it is replacing the 7600x, that is 21% reduction in power consumption. If AMD does in fact ever release a 9700 non-x and 9600 non-x then it will be interesting to see the power consumption. Same is true for the 9800x3d.

View attachment 357869

Yeah, kinda they could have dialed a bit down the clocks and called them vanilla 9700 and 9600 bundled with an elcheapo cooler and release the X ones with 105w TDP to sprint on lowly threaded tasks.

 

[Macro Device]

Performant is not a word in english, do not use it!

Purism is what kills the language, not the other way around. I'd rather address nonsensical writers who use your/you're interchangeably and put "of" instead of "have" after their "would" and "shoulds." These are a real insult to English. Borrowing words is fine, this is how you know the language is alive.

Regarding the 9700X: not that I had any expectations but still... Heavily reminiscent of Kaby Lake. 7700K only was marginally better than 6700K and essentially was a 6700K on 'roids. AMD are now virtually the Intel from 2011–17.

 

[BoggledBeagle]

I checked Igors 9600X review and the first reader comment in the first sentence says "schwer enttäuscht". I find this very funny.

If any linguists now all over the world want to study THE LANGUAGE OF DISAPPOINTMENT, they just need to check comments under all the Zen 5 reviews.

 

[phanbuey]

honestly it just shows there’s a bottleneck somewhere else in the system.

it’s being held back by memory/or some other component in the design - since zen 5 is basically improved zen 4. I’m wondering if they were able to pump the infinity fabric higher and give it faster ddr5 sweet spot if the story would have been different.

Or maybe if the new am5 chipsets or new agesa come out and there is performance to be gained.

This should have been X3D only like they originally planned - the non variants seem like waste of sand.

 

[mkppo]

I think there are two things that stand out:

Single core performance is pretty much what was expected - around 15%
Multicore performance is less, because each core now consumes a bit more power and they're hilariously back end and IO limited. I know these tests aren't done yet, but 9xxx will scale much better with memory overclocking/tuning.

7700 and 7700x doesn't scale much, but 9700x scales almost 6% in applications on average with PBO. When power normalized, that puts 9700x at a 10% increase in application performance and held back by memory more than the 7700x.

People who are saying this is Intel back in the days isn't really true because the double digit IPC gains are there. It's only in MT where it's held back by power and memory. Crank it up, and you're left with a 10% MT and 15% ST gain. Not great, but not terrible either

 

[evernessince]

That is one way of looking at it, the other way is to look at the 7700x power consumption. Where it is 25 watts less, which is a 29% reduction in power consumption. The 9700x replaces the 7700x and as we can see the only big improvement was the power consumption. Same thing with the 9600x, it is replacing the 7600x, that is 21% reduction in power consumption. If AMD does in fact ever release a 9700 non-x and 9600 non-x then it will be interesting to see the power consumption. Same is true for the 9800x3d.

View attachment 357869

The way I'm looking at it is from an architectural standpoint. Comparing the best perf per watt a given architecture can produce gives you a true sense of it's efficiency. The 7700X isn't indicative of the 7000 series's efficiency, it's essentially a 7700 pushed outside it's sweetspot for very minimal performance gain. It's essentially dead on with the 9700X OC results in that chart.

Saying the 9700X reduces power consumption by 29% is only correct is the very literal sense but completely misses the point that the architecture itself does not appear to be any more efficient. When you compare both architectures within their ideal efficiency range there appears to be zero to potentially negative efficiency improvement. Maybe the 9000 series can be tuned even further for better efficiency but there is not much of a performance gain over the 7000 series for that to happen. Overall it does not paint a good picture.

 

[BoggledBeagle]

Or maybe if the new am5 chipsets or new agesa come out and there is performance to be gained.

Those are the HIDDEN POWERS I already wrote about!