AMD Ryzen 9 7950X

[W1zzard]

in the beginning on the page or at the end;

Added the following to the power measurement page:

All power measurements on this page are based on a physical measurement of the voltage, current and power flowing through the 12-pin CPU power connector(s), which makes them "CPU only", not "full system." We're not using the software sensors inside the processor, as these can be quite inaccurate and will vary between manufacturers. All measurements are collected and processed at a rate of 30 data points per second, on a separate machine, so the power measurement does not affect the tested system in any way. Our new data processing pipeline allows us to link recorded data precisely with benchmark runs, so we can easily create the charts below.

 

[oxrufiioxo]

Did gamers buy the 5950X?

I did

I'm both impressed and underwhelmed with these results although other sites seem to have gotten better gaming results although they also used a 3090ti.

Just like with any cpu it is likely very game and even maybe scene dependent though.

Either way can't wait for 3DV$ versions.

 

[ERazer]

like many 5950x user 7950x didn't stuck me to upgrade, cost to newer power platform is not compelling and AMD did mentioned they still have plans for probably higher 3dx cpu for am4

 

[InVasMani]

Value for dollar isnt' quite there with 7950x then again it rarely is for chips at that end of the spectrum.

 

[outpt]

The winner of this yawn fest is: both sides.
Hope there’s more to the gpu’s than this.

 

[champsilva]

Low clock vs High clock , really?

Well, AMD choose to use high clocks to fight intel. So is their fault, right?

 

[Richards]

@W1zzard great review... we need Spider-Man for future benchmarks...

Intel 10nm competing well in efficiency with tsmc 5nm.. meteor lake on intel 4 will destroy every cpu

 

[THU31]

I wonder what kind of clocks this could do at 65 W and 95 W for AVC/HEVC encoding utilizing all 32 threads. Undervolting these CPUs is supposed to be amazing. I hope people will test it.

 

[Pumper]

BTW, the leaked 7950X review shows great performance in 65 ECO mode. Any plans to test that? Would be nice to see what an optimezed setup is capable of and what kind of temps and power draw is possible.

 

[W1zzard]

we need Spider-Man for future benchmarks...

It was released too late to make it into this test bench

 

[AnotherReader]

@W1zzard great review... we need Spider-Man for future benchmarks...

Intel 10nm competing well in efficiency with tsmc 5nm.. meteor lake on intel 4 will destroy every cpu

Intel 10 nm isn't really competing in efficiency with TSMC's N5. Take a look at this slide from Anandtech when they limited the 7950X to 65 W, i.e. 88 W total power draw:

A full power 12900k is unable to even come close to a hobbled 7950X. this is going to be a bloodbath in servers which is the actual focus of Zen 4.

 

[defaultluser]

@W1zzard great review... we need Spider-Man for future benchmarks...

Intel 10nm competing well in efficiency with tsmc 5nm.. meteor lake on intel 4 will destroy every cpu

By throwing in a shitload of e-cores. And still somehow managing to increse power consumption by nearly 100 more watts?

 

[RandallFlagg]

Such irony...

Zen 4 was supposed to be the killer in games, and it isn't.

But it was supposed to lose in productivity apps, which it didn't, at least not to alder lake.

The new arch also seems to lose to Zen 3 in some of the areas alder lake was losing in, like compression.

Seems like a very mixed bag, and not what AMD said it was at all with all of their hyping about Zen 4 game prowess.

It's not bad, it's just weird.

 

[ARF]

The integrated graphics performance is so low. Lower than the junk brand intel:

Core i9-12900K is 5% faster at 1080p:

The IPC improvement over Zen 3 is a measly 3%!

 

[AnotherReader]

Great review W1zzard! Is adding results for ECO mode for each of the three Zen 4 samples on the cards? Are there any AVS-512 enabled applications that you could use to test the uplift due to this addition to Zen 4? Phoronix tested some Linux applications and saw some noticeable performance uplifts.

Alder Lake still competes in single threaded tasks and Raptor Lake is likely to be the single threaded champion, but will have a tough time in multi-threaded applications. All in all, the situation seems like this:

• Intel's focus on big cores and single threaded performance is probably the best approach for desktop users
• AMD's focus on relatively small cores and excellent power consumption at saner clock speeds will be a boon for both servers and laptops
• Anandtech reports that AMD's target was 6 Ghz clocks for Zen 4. Zen 5 will probably push the clock speeds even higher than Zen 4

The integrated graphics performance is so low. Lower than the junk brand intel:

Core i9-12900K is 5% faster at 1080p:

View attachment 263074


The IPC improvement over Zen 3 is a measly 3%!

Where do you get the 3% IPC improvement? Anandtech ran SPEC and saw sizable improvements in many sub tests. In the SpecInt test which is most relevant to common desktop workloads, the average uplift is 23% over the 5950X.

Test	7950X	5950X	Uplift	12900k	Increase over 12900K
perlbench	10.1​	7.93​	27%​	9.7​	4.1%​
gcc	11.8​	10.01​	18%​	11.7​	0.9%​
mcf	9.6​	7.69​	25%​	7.7​	24.7%​
omentpp	7​	5.71​	23%​	6.1​	14.8%​
xalancbmk	8.3​	6.95​	19%​	7.7​	7.8%​
x264	14.9​	12.08​	23%​	13.7​	8.8%​
531.deepsjeng_r	7.3​	6​	22%​	6.4​	14.1%​
541.leela_r	7.5​	6.28​	19%​	6.1​	23.0%​
548.exchange2_r	15.4​	11.78​	31%​	12.2​	26.2%​
557.xz_r	6.4​	5.28​	21%​	4.8​	33.3%​
GEOMEAN	9.39​	7.65​	23%​	8.15​	15.3%​

 

[natero62]

i cant understand why test include undervolting and more like that options?

its not for normal user way,jusr cpu for mobo,turn PC on and run test.

anyway,i cant se uundervolting intel cpus...

but, dissapointment amd zen 5nm cpu

power eat i say is higher than intel 10nm cpu,so much handicap 5nm zen4 get.

still loose gaming..and we i took 12900k not 13900k

i say good value,not editor choice.

again, not taken took that different 5nm cpu and 10nm cpu, its massive different..
i hope intel raptor was 5nm cpu also, we seen truth.
well, few years,but i think 7nm meteor lake show direct..

still, good test if we forget thouse tricks aka undervolting and others.

 

[ARF]

Where do you get the 3% IPC improvement? Anandtech ran SPEC and saw sizable improvements in many sub tests.

AMD Ryzen 9 7950X review - Conclusion (guru3d.com)

 

[Tomgang]

Meh, I'm a little disappointed. Yes 7950X is fast. But compared my 5950X. 7950X have som red flags. Power consumption, heat, gaming performance is not so great as I hoped for.

5950X is easy to air cool, is not that much slower in games and is more power efficient. All in all I stick to my 5950X. It is still more than fast enough for my use. I'm glad I actually just upgraded last year to Zen 3. I had regreted waiting another year for Zen 4.

Looks like I'm better off just upgrading gpu this year.

 

[AnotherReader]

View attachment 263076
AMD Ryzen 9 7950X review - Conclusion (guru3d.com)

They use one application, Cinebench, to determine IPC. I don't think that's a valid way to determine IPC. Moreover, on the IPC test page, they write:

Coming from the Ryzen 9 5950X (Vermeer), we see an increase of 13% in IPC over last gen, exactly what AMD claimed it to be.

 

[LifeOnMars]

Weird times.

 

[ARF]

My conclusion is that this launch is a mess.

The Ryzen 9 7950X is simply a rebadged Threadripper which for some weird reason AMD were not brave enough to launch.

Only the 6-core and 8-core SKUs deserve to be Ryzens but that would put AMD far far BEHIND the Raptor Lake launch tomorrow.

 

[THU31]

Intel 10 nm isn't really competing in efficiency with TSMC's N5. Take a look at this slide from Anandtech when they limited the 7950X to 65 W, i.e. 88 W total power draw:

View attachment 263070

A full power 12900k is unable to even come close to a hobbled 7950X. this is going to be a bloodbath in servers which is the actual focus of Zen 4.

This puts joy in my eyes.

It means it probably stays over 4 GHz in that 65 W test. And that is just a power limit, without manual undervolting.

 

[dayne878]

I'm a bit disappointed by this. It's comparable to the 12900k by the looks of things and that suggests 13900k is going to blow it out of the water.

It IS good for competition, however. It keeps Intel on their toes. It just might not be good for our wallets, since Intel could theoretically price the top tier 13th generation CPUs at higher than the top tier AMD CPUs.

I'm not planning on upgrading for a year or two (have a 10700k from my 2020 build), but if I were this fall/winter I would (even with Intel CPUs still not officially released) be buying Intel still.

 

[BorisDG]

Same here... still waiting for my next Intel HEDT.

 

[efikkan]

hmm , gaming performance

The gaming performance is excellent, but it's not outperforming Alder Lake. But as I've been saying, when most current games are not bottlenecked by Alder Lake, we shouldn't expect something to perform significantly better either until more demanding games arrives. The same will hold true with Raptor Lake too, there will probably be a few games improving, but overall it's only going to be slightly faster in gaming.