AMD Ryzen 7 9700X

[THU31]

You're talking stagnation? Sitting on dual core ULV for 6-8 years or quad core desktop chips for 8-10 years that is called stagnation or underwhelming! Lots of exaggeration here with very little facts to back them up

Intel stagnated for years and people hated it. Is that an excuse for AMD doing the same now?

They pushed core counts with the release of Zen after a decade of quad cores from Intel. We got a jump to 8 cores, and then quickly to 16 cores. Intel responded well with their hybrid architecture and are miles ahead in terms of multi-threaded performance in the lower segments.
For some reason AMD are not willing to increase core counts again. 8 cores should be entry level at this point. Ryzen 5 and 7 have the same core counts as they did back in 2017.
If core counts are the same, while IPC and efficiency increase are marginal, it's the definition of stagnation.

 

[R0H1T]

Intel stagnated for years and people hated it. Is that an excuse for AMD doing the same now?

Intel did that deliberately ~ that is a major difference! They had 10c chips on HEDT by the time Zen launched & it cost 4-5x their desktop flagship! They also went from 4c to 10c in 3 years post Zen release. Find me an AMD chip with 16c on 1 chiplet, except probably the zen5c for Turin?

AMD's gone from 8c-16c in 4-5 years, just a wait a bit more for your 32c chips. Like I said many times before this as well you will need more bandwidth to keep feeding the cores otherwise they are useless!

 

[Dyno]

Would you build a 7000-series non-X3D CPU now that these are out? No.
Would you replace a 5800X3D with a 9700X? No.
Would you build a 7800X3D system now that 9700X3D is just around the corner? No.

This staggered launch is a typical "dumb AMD marketing" mistake again. They need to launch new generations by putting their best foot forward. That's the 9800X3D and the 9950X. These two CPUs are great compared to their respective 7000-series equivalents, but those equivalent 7700X and 7600X are some of the weakest and least popular CPUs in AMD's current lineup. The 7800X3D and 7500F are their two bestsellers by a country mile.

Isn't it suppose to be the 9800X3D and not the 9700X3D that is coming soon?

 

[W1zzard]

@W1zzard, wouldnt a more accurate comparison be to benchmark the 9600/9700 set to 105W or even 95w?

Check the PBO Max numbers. Stock is stock, and the way most people will use the processor, only AMD can change the default spec

 

[Chrispy_]

Isn't it suppose to be the 9800X3D and not the 9700X3D that is coming soon?

Typo. I got it right in the sentence after that.

 

[Avro Arrow]

I'm really confused about the Zen5 launch because different tech press institutions have opinions that vary wildly. Techspot/Hardware Unboxed and Gamers Nexus say it's (in not so polite terms) disappointing while TechPowerUp and Tom's Hardware love it. I've honestly never seen such a massive difference of opinion across these four sources.

One thing that caught me off guard is the fact that the R7-9700X is barely faster (5%) in gaming than the R5-7600 (or my own R7-5800X3D). I know it's not specialised for gaming like my 5800X3D but neither is the R5-7600. The performance difference between Zen3 and Zen4 is far greater than the difference between Zen4 and Zen5. This makes me wonder if the performance uplift from Zen3 to Zen4 was truly a result of improved architecture or if it was artificially inflated by the speed difference of DDR4 and DDR5.

I think it's possible that AMD is making their mainline processors more optimised for productivity because they know that gamers will always flock to the X3D parts. The only way we'll know for sure is when they release the R7-9800X3D.

 

[AnotherReader]

I'm really confused about the Zen5 launch because different tech press institutions have opinions that vary wildly. Techspot/Hardware Unboxed and Gamers Nexus say it's (in not so polite terms) disappointing while TechPowerUp and Tom's Hardware love it. I've honestly never seen such a massive difference of opinion across these four sources.

One thing that caught me off guard is the fact that the R7-9700X is barely faster (5%) in gaming than the R5-7600 (or my own R7-5800X3D). I know it's not specialised for gaming like my 5800X3D but neither is the R5-7600. The performance difference between Zen3 and Zen4 is far greater than the difference between Zen4 and Zen5. This makes me wonder if the performance uplift from Zen3 to Zen4 was truly a result of improved architecture or if it was artificially inflated by the speed difference of DDR4 and DDR5.

I think it's possible that AMD is making their mainline processors more optimised for productivity because they know that gamers will always flock to the X3D parts. The only way we'll know for sure is when they release the R7-9800X3D.

Zen 4 clocks significantly higher than Zen 3 while Zen 5 clocks lower than Zen 4 in multithreaded workloads. That reduces the impact of the real microarchitectural improvements.

 

[Garrus]

It's all fine until you look at the new higher prices. The Ryzen 7600 was $250 CAD in Canada when I built a system for my roommate last month, the 9600X is $400 today. No.

I think if reviews are confusing and not clear, that in itself means the CPU is not good. The 7800X3D is completely obviously good when you use it. $500 CAD for the 7800X3D versus $400 for the Ryzen 9600? Obvious.

Too bad. I hope the 9800X3D delivers or the entire Zen5 launch will go down as AMD's biggest mistake in the last 10 years. We already saw that the new Laptop AI CPUs are terrible for gaming. Latency makes them slower than last year's CPUs in laptops.

Every Zen5 product is a no buy so far. That is not what you want to see.

 

[Chrispy_]

I think it's possible that AMD is making their mainline processors more optimised for productivity because they know that gamers will always flock to the X3D parts. The only way we'll know for sure is when they release the R7-9800X3D.

I think a lot of high-volume OEM partners are wanting more efficient CPUs that can be driven by boards with cheaper VRMs, cheaper PSUs, and cooled with cheaper coolers.

AM5 has taken a couple of years of valid criticism for having motherboards that are too expensive, as well. Making processors that run to their full potential in cheaper (A-series) motherboards is a big deal for AMD in both the consumer and OEM markets; I would guess that people who care about PCIe Gen5 or overclocking are a relatively small minority, possibly even an insignificant minority when it comes to AMD's bottom line.

 

[mkppo]

The more I look into the data, the more it seems like Zen 5 design choices will benefit servers more than consumers. Traditional SSE workloads don't benefit much and INT uplift is masked by slightly higher power consumption and in turn lower clocks at the same power, coupled with larger memory bottleneck than what was present in Zen 4.

Zen 5 looks like an interesting CPU to play around with memory. Maybe I will snag a 9950x with 870 when prices fall a bit.

Don't forget that raw numbers for the core itself are pretty staggering - SPEC 1T has a 13% INT uplift and 26% FP uplift. around 6-7% of that FP can be attributed to AVX512 but even then, the uplift is massive for single threaded workloads. But the cores seem to have trouble offloading it to memory, so there's an obvious bottleneck at the back end and IO.

Memory scaling and tuning tests with Zen 5 should be very interesting.

 

[W1zzard]

while TechPowerUp and Tom's Hardware love it

Wait.. what? Love? Did we read the same conclusion?

 

[Zach_01]

The performance difference between Zen3 and Zen4 is far greater than the difference between Zen4 and Zen5. This makes me wonder if the performance uplift from Zen3 to Zen4 was truly a result of improved architecture or if it was artificially inflated by the speed difference of DDR4 and DDR5.

I think it's possible that AMD is making their mainline processors more optimised for productivity because they know that gamers will always flock to the X3D parts. The only way we'll know for sure is when they release the R7-9800X3D.

Its the jump of power consumption from 0.6~0.8+GHz over Zen3. Of course there was some IPC gains too but a lot of performance was from clocks and power increase.
Now AMD for these 2 low-mid CPUs has decreased power. On the other high-end ones power stays the same as Zen4.

---------------------------------------------

There are hints that Zen5 is hungry for memory bandwidth
6000 vs 6400MT/s
At least some of them can do FCLK 2133-2200Mhz for 6400 DDR5



Honestly AMD does not know how to promote its products.
They should've released their entire line at once. Wait a month or even 2 if necessary and do it and I mean X3Ds too.
That way you have a product for every taste out there. Good efficiency, raw productivity performance and gaming performance.
In my opinion they gone too conservative with 65W TDP (88W PPT). At least 95W TDP (125W PPT) would be better and still there is room for PBO and all of its tweaks in there for those who want to experience the 170~180W.

Also maybe they could've slightly improve I/O Die to be able to play FCLK at 2200~2300MHz 6600-7000MT/s if not more. Of course this increases power further.
Unless this speed can be improved by better board tracing (see 800series) and maybe new AGESA(s)
Will see...

 

[azrael]

It's all fine until you look at the new higher prices. The Ryzen 7600 was $250 CAD in Canada when I built a system for my roommate last month, the 9600X is $400 today. No.

I think if reviews are confusing and not clear, that in itself means the CPU is not good. The 7800X3D is completely obviously good when you use it. $500 CAD for the 7800X3D versus $400 for the Ryzen 9600? Obvious.

Too bad. I hope the 9800X3D delivers or the entire Zen5 launch will go down as AMD's biggest mistake in the last 10 years. We already saw that the new Laptop AI CPUs are terrible for gaming. Latency makes them slower than last year's CPUs in laptops.

Every Zen5 product is a no buy so far. That is not what you want to see.

I don't know why people get so hung up on the price. They're not higher than when the equivalent 7000 series CPUs launched, but lower. Just because the last gen now has dropped sufficiently in price doesn't mean the new gen should be given away.

I also believe Zen5 is quite good. What's bad is (once again) AMD's marketing. TBH they should fire the whole team. A much better way to market the new series (or at least the first two CPUs) would've been to plainly state that they're tuned for efficiency and if you would rather sacrifice that for performance then just enable PBO. Or perhaps add that functionality to Ryzen Master or some other tool. What's needed is the opposite of the "Eco Mode".

 

[ShrekOC]

When it comes to CPUs, we can't really judge anything based on launch prices. Let's see how prices settle in over the next 90 days. That said, the only AMD CPUs that interest me are the 5600, 5700X3D, 7500F and 7700 because those are available from AliExpress for about 1/3 less than the US retail price (except the 7500F which isn't sold in the US).

 

[ARF]

I also believe Zen5 is quite good.

Not that it is particularly so, but:

What's bad is (once again) AMD's marketing. TBH they should fire the whole team. A much better way to market the new series (or at least the first two CPUs) would've been to plainly state that they're tuned for efficiency

This I agree with. They should have invented a new naming convention for these. Something like Ryzen 5 9500 Eco Edition 65 or similar, in order to aggressively and proactively emphasize on that particular product quality.
Because, as marketed like they currently are, neither the average joe that doesn't read reviews, nor the retail stores consultants will ever know about it..

 

[R0H1T]

I'm really confused about the Zen5 launch because different tech press institutions have opinions that vary wildly. Techspot/Hardware Unboxed and Gamers Nexus say it's (in not so polite terms) disappointing while TechPowerUp and Tom's Hardware love it. I've honestly never seen such a massive difference of opinion across these four sources.

One thing that caught me off guard is the fact that the R7-9700X is barely faster (5%) in gaming than the R5-7600 (or my own R7-5800X3D). I know it's not specialised for gaming like my 5800X3D but neither is the R5-7600. The performance difference between Zen3 and Zen4 is far greater than the difference between Zen4 and Zen5. This makes me wonder if the performance uplift from Zen3 to Zen4 was truly a result of improved architecture or if it was artificially inflated by the speed difference of DDR4 and DDR5.

I think it's possible that AMD is making their mainline processors more optimised for productivity because they know that gamers will always flock to the X3D parts. The only way we'll know for sure is when they release the R7-9800X3D.

Think of it this way, just based on the launch it's a "relatively" disappointing product. But it doesn't exist in a vacuum. Take any number of CPU's from $100~1000 and AMD will have at least one VFM king in each $100-200 bracket. So you can say this exists where it is because AM4 CPU's are still selling. This should be easy enough for most forum dwellers to understand & as always buy it on sale! Except sub $100 I doubt there's anything worth buying from Intel for now especially if AMD CPU's are cheap at your place.

 

[Hecate91]

Not that it is particularly so, but:



This I agree with. They should have invented a new naming convention for these. Something like Ryzen 5 9500 Eco Edition 65 or similar, in order to aggressively and proactively emphasize on that particular product quality.
Because, as marketed like they currently are, neither the average joe that doesn't read reviews, nor the retail stores consultants will ever know about it..

IMO Zen 5 is good, not sure why people expected it to beat the X3D in everything, though I like that AMD went with efficiency over just cranking up the power, I think its nice to have a power efficient cpu which is still decent for gaming.
And I think the 9600X and 9700X would've been better without the X sku naming, save the X for a higher clocked version, or just make the X3D the high performance CPU.
Also the the average joe isn't building their own system, they're just going to buy whatever gaming system from an OEM.

 

[sil3ntearth]

Man, I love TPU, I don't see how this is a recommended product. This launch is crazy similar to the 11th gen core release, minus the better temps and power consumption. That performance difference is embarrassing.

IMO Zen 5 is good, not sure why people expected it to beat the X3D in everything, though I like that AMD went with efficiency over just cranking up the power, I think its nice to have a power efficient cpu which is still decent for gaming.
And I think the 9600X and 9700X would've been better without the X sku naming, save the X for a higher clocked version, or just make the X3D the high performance CPU.
Also the the average joe isn't building their own system, they're just going to buy whatever gaming system from an OEM.

I wasn't expecting it to beat any of the x3d stuff, but I was expecting a much better generational improvement. I think the fact that AMD themselves held back this launch is all you need to know.

 

[Neo_Morpheus]

After reading the Phoronix and Wendell Linux review, it makes me think that once again Windows is holding AMD back.

Not the first time mind you, they have been doing this since the purposely delayed Win XP 64.

 

[NC37]

Why even put an iGPU on it? That makes no sense at all.

 

[Rabit]

Why even put an iGPU on it? That makes no sense at all.

I using this tiny iGPU on my PC, in hybrid GPU mode

 

[SL2]

After reading the Phoronix and Wendell Linux review, it makes me think that once again Windows is holding AMD back.

I'd like to see some discussions between reviewers about the various conclusions. This one was interesting, but ultimately it doesn't change the outcome in a meaningful way.

Why even put an iGPU on it? That makes no sense at all.

Maybe you know something AMD and Intel doesn't know lol..

It DOES make sense, I think they've figured that out by now. Every user can't have a GTX 750 or whatever as a backup for whenever a graphics card isn't present or doesn't work.

Everyone doesn't need a graphics card.

 

[RogueSix]

IMO Zen 5 is good, not sure why people expected it to beat the X3D in everything

While AMD might not have stated that Zen 5 would be beating X3D in "everything", they did in fact claim at a tech event in July (i.e. quite recently) that the 9700X would beat the 7800X3D by 2% in gaming.

Most sites, however, show the 9700X trailing the 7800X3D by 15% to 20% in games. That is quite a huge miss on AMD's prediction. What happened between mid July and now? Anyway, those news made it to most tech sites courtesy of Overclock3d so it should be quite obvious why people had higher expectations with regard to gaming performance.

Finally, even if we just go by AMD's earlier (at the Zen 5 reveal) statement then AMD said that 9700X vs. 7800X3D would be a win for the 7800X3D but a close one. Well, as it turned out now, it is not exactly close.

That is why -imo- people rightfully expected more of 9700X in gaming. AMD themselves promised "us" more...

 

[JWNoctis]

Maybe you know something AMD and Intel doesn't know lol..

It DOES make sense, I think they've figured that out by now. Every user can't have a GTX 750 or whatever as a backup for whenever a graphics card isn't present or doesn't work.

Everyone doesn't need a graphics card.

It's also useful for things when you really need every ounce of VRAM you've got and the only other choices are not reduced performance, but either no GUI or no-go by OOM. GUIs and browsers can take up a surprisingly large chunk of that. It's not always useful, but it is good to have the choice.

Mere HTPCs would be too typical.

After reading the Phoronix and Wendell Linux review, it makes me think that once again Windows is holding AMD back.

Not the first time mind you, they have been doing this since the purposely delayed Win XP 64.

Might well be consumer workloads holding AMD back, since most of the architectual enhancements are things getting wider. Consumer workloads take advantage of that much less than enterprise ones, with few notable exceptions.

The charges of Zen 5 being primarily enterprise-optimized and therefore not benefiting the typical consumer since it has full-fat AVX-512 does have a point in this way, albeit a bit too extreme and a bit of sour grapes, when the corresponding parts on the competitor's current client processors are fused off.

 

[Minus Infinity]

They don't gain anything at 105W - even with PBO enabled the gains are minimal. Better to get the same performance with 35% less power than to get 8% better performance at the same power.

Tom's Hardware and KitGuru beg to differ. This is what is confusing. HUB especially had a 9700X with pathetically poor all-core clocks of just over 4.4GHz, much worse than others, so not surprised his results are so bad. Tom's Hardware showed good gains with PBO in both productivity and gaming. Admittedly, he preferred the 9600X a lot more, but that showed impressive PBO gains. Note at Tom's Hardware they showed the stock 9600X beating 14600K in 1080p gaming by 12% on average and got another 8% with PBO. This is nothing like HUB reported. They rated the 9600X even worse than the 9700X. GN got a dud 9600X and couldn't run the tests at all.

Prices though are stupidly high and $280 for 6 cores in 2024 is a joke IMO. I would have also called these 9600 and 9700 for $220 and $290 and released 9600XT and 9700XT with 105W TDP for $250 and $320. People saying how efficient these new chips are when they only match the old 7700/7600 for efficiency.

Can't believe that AMD could only managed to squeeze out a 5% progress to their uarch after 2 years.

For gamers, just grab a 7800X3D now, 9800X3D wouldn't be much faster anyway. For gamers+productivity, there are way better options out there already

Depends on who you read. HUB is super negative and IMO have something wrong going on but basically had a go at others that got better results than him.

Tom's Hardware got much better than 5% for productivity and gaming even stock let alone PBO'd. Stock he got 12 game geomean of 167 for 9700X and 149 for the 7700X, so that's 12% and PBO added 9% further. HUB claims PBO did bugger all maybe 1-2% if you're lucky.