AMD Clarifies Ryzen 7000 "Zen 4" TDP and Power Limits: 170W TDP, 230W PPT

[btarunr]

The mention of "170 W" in one of the slides of AMD's Computex 2022 reveal of the upcoming Ryzen 7000 "Zen 4" desktop processors, caused quite some confusion as to what that figure meant. AMD issued a structured clarification on the matter, laying to rest the terminology associated with it. Apparently, there will be certain SKUs of Socket AM5 processors with TDP of 170 W. This would be the same classical definition of TDP that AMD has been consistently using. The package-power tracking (PPT), a figure that translates as power limit for the socket, is 230 W.

This does not necessarily mean that there will be a Ryzen 7000-series SKU with 170 W TDP. AMD plans to give AM5 a similar life-cycle to AM4, which is now spanning five generations of Ryzen processors, and the 170 W TDP and 230 W PPT figures only denote design goals for the socket. AMD, in a statement, explained why it needed to make AM5 capable of delivering much higher power than AM4 could—to enable higher CPU core-counts in the future, more on-package hardware, and for new capabilities like power-hungry instruction-sets (think AVX-512). AMD has been calculating PPT as 1.35 times TDP, since the very first generation of Ryzen chips. For a 105 W TDP processor, this means 140 W PPT, and the same formula continues with Ryzen 7000 series (230 W is 1.35x 170 W).



The AMD statement follows.

"AMD would like to issue a correction to the socket power and TDP limits of the upcoming AMD Socket AM5. AMD Socket AM5 supports up to a 170 W TDP with a PPT up to 230 W. TDP*1.35 is the standard calculation for TDP v. PPT for AMD sockets in the "Zen" era, and the new 170 W TDP group is no exception (170*1.35=229.5).

This new TDP group will enable considerably more compute performance for high core count CPUs in heavy compute workloads, which will sit alongside the 65 W and 105 W TDP groups that Ryzen is known for today. AMD takes great pride in providing the enthusiast community with transparent and forthright product capabilities, and we want to take this opportunity to apologize for our error and any subsequent confusion we may have caused on this topic."

View at TechPowerUp Main Site

 

[thesmokingman]

Damn this could be a huge win if they stuff greater than 16 cores into a desktop chip.

 

[ModEl4]

"AMD would like to issue a correction to the socket power and TDP limits of the upcoming AMD Socket AM5"

a correction to Robert Hallock, in the PCWorld interview he clearly mentioned 170W PPT, lol he didn't even have clear in his mind what the TDP/PPT figures were?

 

[Daven]

Base: 65, 105, 125W
Turbo: 170W
Socket: 230W (overclocking and future)

 

[Deleted member 202104]

I've said it before, and I'll say it again:

 

[Pumper]

Wonder how many of the watts are dedicated for the iGPU. Other than that, the higher TDP and reliance of higher clock speeds makes me think that Zen4 will be closer to factory overclocked Zen3 than something brand new.

 

[ncrs]

I've said it before, and I'll say it again:

View attachment 248916

It's not really about the raw power usage alone, but more about power/perf:

Intel has improved in this department with Alder Lake, but previous generations were awful by comparison.

 

[john_]

They can move over 170W in future AM5 mobos if Intel manages to REALLY push them with something more than a combination of Performance + Marketing cores.

 

[Deleted member 202104]

It's not really about the raw power usage alone, but more about power/perf:

Intel has improved in this department with Alder Lake, but previous generations were awful by comparison.

I don't disagree with any of that.

At this point we don't know anything about Zen 4's performance - it could have the highest perf/watt ever, or it could be AMD's Rocket Lake. People (in this thread even) are already saying higher power consumption is no big deal.

 

[oxrufiioxo]

I don't disagree with any of that.

At this point we don't know anything about Zen 4's performance - it could have the highest perf/watt ever, or it could be AMD's Rocket Lake. People (in this thread even) are already saying higher power consumption is no big deal.

Yeah, I'm not sure how this is going to go down Ryzen 5000 already has a high thermal density this looks to be even worse. I'm waiting to see if amd are pushing these out of the efficiency window for the sake of performance.

 

[usiname]

The performance is 45% higher than 12900k. Even if the consumption is same, this is win. Undervolt will be possible for sure with almost no loss of performance like the previous Zens, but lets wait for the reviews, 45% is not guaranteed and we don't know much about the real consumption under load.

 

[phill]

I'll hang on for the reviews but I'm really looking forward to the reviews!!

 

[HD64G]

Point is that the 170TDP/230PPT limits aren't of any upcoming in the near future CPU but a platform limit to be utilised in next gens of CPUs on AM5 if needed. So, not a big deal for anyone atm to bother with. Only board manufacturers will have trouble and make expensive VRMs to withstand that power figure.

 

[Jism]

The performance is 45% higher than 12900k. Even if the consumption is same, this is win. Undervolt will be possible for sure with almost no loss of performance like the previous Zens, but lets wait for the reviews, 45% is not guaranteed and we don't know much about the real consumption under load.

Its just a standard offering a higher max TDP/PPT and such. AM4 was limited to if i'm correct at 144W.

They can still release perfectly fine 65w, 95w, 105w or even 140w TDP based CPU's.

This will be interesting tho. See if a 50$ board handles a high end CPU.

 

[zlobby]

What if mobo manufacturers design mobos that can handle, say 300W package power? Will AMD allow boost to this new target?

 

[Jism]

No they wont, but it does offer headroom in terms of overclocking.

FX line for example was rated for up to 25A on the 12V line or something, and with higher end boards you where able to override that with a additional 40% extra (35a max).

That only comes into play when your chasing records and using LN2. Knowing how accurate XFR is manual OC's won't yield that much anyway.

 

[DeathtoGnomes]

I gonna guess that the AMDs iGPU contributes to the TDP cap, not a huge amount. I hope this comes out in reviews.

 

[tabascosauz]

I gonna guess that the AMDs iGPU contributes to the TDP cap, not a huge amount. I hope this comes out in reviews.

I doubt it. That's almost 90W increase in PPT over what we have now. Vega 8 in 5700G when balls to the wall draws roughly 35-45W and doesn't really go past that on ambient. The "iGPU" in mainstream Zen 4 is much smaller and slower. New iGPU easily fits into 142W if they make the new 6nm IOD and their Fabric links a little more efficient.

230W isn't unattainable for 2CCD Zen 3. It's just not very easy to cool and excludes pretty much everything outside of water, custom water, and full blast dual towers.

Robert said a lot about the exciting "5.5GHz" core clock, but as we all know on Zen 3 the exciting ST clock =! actual sustained ST clock =! MT clocks. We'll see. Zen 3 did offer big MT clock uplift with lower temps and lower volts over Zen 2, however, so it's not impossible even if N5 is dense.

 

[ModEl4]

The most probable scenario is that we are going to have at launch higher than 105W TDP offerings (at least one)
It kinda seems he confirmed that we are going to have 65W/105W/170W options, at least that's how my interpretation of the below:
"The confusion stems from me misspeaking. I had misread some tech docs and got my wires crossed. Sorry to the community for that. :
Ryzen sockets are CPU_TDP*1.35 = PPT (maximum socket power).
So 65W TDP = 88W PPT (no change from AM4), 105W TDP = 142W PPT (no change), and 170W TDP = 230W PPT (new option)"
IF we are going to have something between 105-170W TDP, i don't have high expectations to be somewhere in the middle (135W TDP) since based on his comment:
"The Computex processor was a 16-core prototype sample not yet fused to specific power/TDP values, but it was operating in a range below the new 170W TDP group we've developed. It's a conservative figure."
He didn't say somewhere above 105W or a lot below 170W, or he could use the expression somewhere in the middle, combine that they already confirmed that they used Blender because it's free in order the community to be able to replicate the result and since the performance must be at least at Computex's level, the most probable scenario is that we are going to have a 16c 170W TDP option at launch (especially if there is an additional 105W option 16core part, because they would need to differentiate them)
They could do multiple TDP SKUs like in the past, but it may bring confusion to the market, but again 170W should be one of the two TDP options.

 

[freeagent]

I don’t see the big deal, Zen 3 can already put these ppt numbers out..

 

[Minus Infinity]

Point is that the 170TDP/230PPT limits aren't of any upcoming in the near future CPU but a platform limit to be utilised in next gens of CPUs on AM5 if needed. So, not a big deal for anyone atm to bother with. Only board manufacturers will have trouble and make expensive VRMs to withstand that power figure.

Zen 5 is going BIG.little and it appears the little cores will be Zen 4C cores. I could easily imagine a Zen 5 8950X with 16 Zen 5 cores and 8/16 4C cores requiring the higher TDP/PPT

 

[AlwaysHope]

Enthusiasts into OC & next gen high end GPU will be into shopping for new PSU, think 1Kw minimum!
But its all about the return on investment isn't it? so no big deal...

 

[DeathtoGnomes]

I doubt it. That's almost 90W increase in PPT over what we have now. Vega 8 in 5700G when balls to the wall draws roughly 35-45W and doesn't really go past that on ambient. The "iGPU" in mainstream Zen 4 is much smaller and slower. New iGPU easily fits into 142W if they make the new 6nm IOD and their Fabric links a little more efficient.

230W isn't unattainable for 2CCD Zen 3. It's just not very easy to cool and excludes pretty much everything outside of water, custom water, and full blast dual towers.

Robert said a lot about the exciting "5.5GHz" core clock, but as we all know on Zen 3 the exciting ST clock =! actual sustained ST clock =! MT clocks. We'll see. Zen 3 did offer big MT clock uplift with lower temps and lower volts over Zen 2, however, so it's not impossible even if N5 is dense.

So 35-45w less without iGPU, thats assuming you can disable it with a dGPU. Or do I have it wrong?

 

[tabascosauz]

So 35-45w less without iGPU, thats assuming you can disable it with a dGPU. Or do I have it wrong?

Did you read what I wrote? That's Vega 8 maxed out at 1.2V. Raphael isn't an APU. The AM5 chiplet "iGPU" isn't even in the same ballpark of performance and power draw... that's not where the PPT is coming from.

 

[oxrufiioxo]

Did you read what I wrote? That's Vega 8 maxed out at 1.2V. Raphael isn't an APU. The AM5 chiplet "iGPU" isn't even in the same ballpark of performance and power draw... that's not where the PPT is coming from.

Yeah from the sound of it it's meant for diagnosing the pc or at most basic web browsing. I doubt even including the IO chipset it consumes more than 15w.