Question AMD Rembrandt/Zen 3+ APU Speculation and Discussion

[izaic3]

Alright, so we've had some leaks so far. I don't know if any of it's been confirmed yet, as it's pretty early, but here is what I've surmised so far (massive grain of salt of course):

• Zen 3 "enhanced"
• DDR5 and LPDDR5 support
• PCIE4
• 6nm
• CVML
• Navi 2/RDNA/RDNA 2
• 12 CU
• Socket FP7/AM5
• Possible USB4
• 2021(unlikely) -2022 (more likely) release date

If if turns out to have RDNA 2 and 12 CU, I could see iGPU performance potentially almost doubling over Cezanne.

If I've made any mistakes or gotten anything wrong, please let me know. I'd also love to hear more knowledgeable people weigh in on their expectations.

 

[ahimsa42]

If the GPU market doesn't even out anytime soon, AMD could make a killing selling these as desktop APUs. 1650's are going for ~$350 right now. When the Zen 4 APUs eventually come out, we may be to the point where 60 FPS 1080 gaming can be done without needing a discrete GPU.

from what i have read, zen 5 is also going be a tremendous improvement over zen 4 apu's with RDN3 & hybrid 3nm cores. my plan is to buy RMB this year, skip phoenix & wait for strix point in 2024/2025. it's definitely a very exciting time for apu enthusiasts to say the least.

 

[Abwx]

If the GPU market doesn't even out anytime soon, AMD could make a killing selling these as desktop APUs. 1650's are going for ~$350 right now. When the Zen 4 APUs eventually come out, we may be to the point where 60 FPS 1080 gaming can be done without needing a discrete GPU.

At 5nm they can pack 1024 SPs at a smaller size than the current 768 ones, and they can get roughly 10% frequency uplift at same TDP, not counting eventually faster RAM.

Expect Zen 4 next gen APUs to perform 50% better, wich should yield much higher framerate than 60 FPS.

 

[Insert_Nickname]

from what i have read, zen 5 is also going be a tremendous improvement over zen 4 apu's with RDN3 & hybrid 3nm cores. my plan is to buy RMB this year, skip phoenix & wait for strix point in 2024/2025. it's definitely a very exciting time for apu enthusiasts to say the least.

Bought a 5700G*, so I'm good for at least until second gen AM5. We'll see how things work out by then.

*Pretty much same CPU part as Rembrandt, but compatible with AM4.

 

[Mopetar]

At 5nm they can pack 1024 SPs at a smaller size than the current 768 ones, and they can get roughly 10% frequency uplift at same TDP, not counting eventually faster RAM.

Expect Zen 4 next gen APUs to perform 50% better, wich should yield much higher framerate than 60 FPS.

To some degree the ultimate limiter will be memory bandwidth so we may not see AMD increase the CU count much beyond what they have now.

The counter to that would be having something akin to the SLC that Apple has in their newest chips where any part of the hardware can make use of it. That would allow the L3 to effectively act the same as the infinity cache in RDNA2 GPUs and help alleviate a lot of that bandwidth bottleneck. If that can be further enhanced with v-cache it would make for a very powerful APU.

 

[ahimsa42]

Bought a 5700G*, so I'm good for at least until second gen AM5. We'll see how things work out by then.

*Pretty much same CPU part as Rembrandt, but compatible with AM4.

a 5700g may be similar to RMB cpu wise but even in the 6800U, the RDNA2 igpu will crush the cezanne vega igpu.

 

[Thibsie]

I guess nobody knows how availability of RMB will be for OEMs ?
I need a lappy but if RMB are there in about 2or 3 months in Thinkpads, getting Cezanne now would be a shame IMO.

 

[ahimsa42]

I guess nobody knows how availability of RMB will be for OEMs ?
I need a lappy but if RMB are there in about 2or 3 months in Thinkpads, getting Cezanne now would be a shame IMO.

no one does know for certain but i think if at all possible it would be wise to wait a few months to see how things shake out with RMB. you can always buy a 5000 series laptop if there is major RMB scarcity when it is released but if you buy now & 6000 series laptops are plentiful & reasonabliy priced you may have big regrets about not waiting to snag one instead.

 

[flatwhite]

I guess nobody knows how availability of RMB will be for OEMs ?
I need a lappy but if RMB are there in about 2or 3 months in Thinkpads, getting Cezanne now would be a shame IMO.

My guess is that Legion series & Asus G14/G15 may deliver first batch of orders in late Feb/early March in SEA region

 

[ahimsa42]

AMD, Intel and NVIDIA processors and graphics cards will become more expensive in 2022 - all because of rising manufacturing costs

AMD, Intel and NVIDIA's processors, graphics cards and specialized ASIC chips will rise in price this year, DigiTimes predicts. All of these...

www.ryzencpu.com

more insane inflation & not a good sign for RMB buyers if 6800/6900 laptops are going to cost 10-20% more than cezzane. for example, if accurate it would mean a product which would have cost $800 would go for $880-$960 instead!

 

[eek2121]

https://twitter.com/x/status/1483837240212168704

Apparently, Rembrandt takes quite a performance hit at 15W. However, I think people are still misunderstanding why Rembrandt was tested at 28W to begin with. Apple started this trend. Their current laptop chips are 30W. AMD and Intel are just changing things up to match. Cores are getting too big to run at peak performance when capped to 15W. Capping to 15W means sacrificing single core performance. Battery and cooling tech has advanced far enough that running this chips in thin/light laptops should rarely be an issue.

 

[moinmoin]

Wonder if that's in reference to a U or an H variant. A couple of days ago I found out that my 4500U laptop allows me to to freely change cTDP limits at any time (as a CLI user ryzenadj is nice and simple), and testing with Speedometer beyond 11W there was no more score improvement to be had, a main thread maxing out at 4GHz as advertised it seems. Significantly more heavy MT and especially GPU usage are very likely a different matter for cTDP of course.

 

[leoneazzurro]

https://twitter.com/x/status/1483837240212168704

Apparently, Rembrandt takes quite a performance hit at 15W. However, I think people are still misunderstanding why Rembrandt was tested at 28W to begin with. Apple started this trend. Their current laptop chips are 30W. AMD and Intel are just changing things up to match. Cores are getting too big to run at peak performance when capped to 15W. Capping to 15W means sacrificing single core performance. Battery and cooling tech has advanced far enough that running this chips in thin/light laptops should rarely be an issue.

He has no Rembrandt to test, he tested Cezanne and he found that "there is a substantial hit" at 15W vs 25W but he did not specify how much, the only testing he spoke about was Rainbow Six which was 66FPS instead of 74FPS. Which means a 12% more on Cezanne going from 15W to 25W. Which also means that instead of 2x we would get a 1,78x comparing a 28W Rembrandt to a 25W Cezanne which is not a +100'% increase but still damn good at practically same power level and probably better bettery life. Techepifany lately is all AMD bashing, often without much sense.

 

[Thibsie]

Techepifany lately is all AMD bashing, often without much sense.

Agreed

 

[IntelUser2000]

How is that bashing? He's just pointing out AMD is also resorting to misleading presentations. So pointing out the facts are called bashing?

He may have become salty(just guessing I don't know) but it's not bashing. Remember if you attack someone to change their views it often has the opposite effect.

He says even after that the gain will be 1.5-1.7x, which is really damn good, and they could have shown that instead.

Should always clamour for apples to apples comparisons, regardless of the size of the company, or your bias.

 

[DrMrLordX]

How is that bashing? He's just pointing out AMD is also resorting to misleading presentations. So pointing out the facts are called bashing?

His analysis is also misleading. Instead of running current hardware (Rembrandt) at 28W (when we all wanted to see it running @ 15W), he's running older hardware (Cezanne) @ 15W. Which is really no better.

 

[IntelUser2000]

His analysis is also misleading. Instead of running current hardware (Rembrandt) at 28W (when we all wanted to see it running @ 15W), he's running older hardware (Cezanne) @ 15W. Which is really no better.

There is no Remembrandt laptops to buy. How is that a misleading comparison? Cezanne is the latest available, and the predecessor of Rembrandt. AMD claims 2x the performance compared to 5800U which is Cezanne are they not? He's saying if you give 28W to Cezanne and faster LPDDR4X-4266 then the gap will go from 2x+ to 1.5-1.7x.

 

[tamz_msc]

There is no Remembrandt laptops to buy. How is that a misleading comparison? Cezanne is the latest available, and the predecessor of Rembrandt. AMD claims 2x the performance compared to 5800U which is Cezanne are they not? He's saying if you give 28W to Cezanne and faster LPDDR4X-4266 then the gap will go from 2x+ to 1.5-1.7x.

I have a feeling that RMB iGPU performance claims are going to be similar to Polaris perf/W claims.

Apparently, Rembrandt takes quite a performance hit at 15W. However, I think people are still misunderstanding why Rembrandt was tested at 28W to begin with. Apple started this trend. Their current laptop chips are 30W. AMD and Intel are just changing things up to match.

Wait, what? I'm pretty sure TGL-U launched before M1.

 

[DisEnchantment]

He has no Rembrandt to test, he tested Cezanne and he found that "there is a substantial hit" at 15W vs 25W but he did not specify how much, the only testing he spoke about was Rainbow Six which was 66FPS instead of 74FPS. Which means a 12% more on Cezanne going from 15W to 25W. Which also means that instead of 2x we would get a 1,78x comparing a 28W Rembrandt to a 25W Cezanne which is not a +100'% increase but still damn good at practically same power level and probably better bettery life. Techepifany lately is all AMD bashing, often without much sense.

Probably salty he did not get sampled on the laptops, a bunch of Tubers got the new ASUS laptops. Next thing he will derive Zen4 performance from Zen3.
He is an avid APU fan I guess not getting those new APUs to test irks him quite a bit.
Not worthwhile to debate until the hardware is there.

 

[uzzi38]

I feel like an absolute idiot tbh. The footnotes for the RMB slides actually say "nominal" TDP of 15/28W.

It's the equivalent of saying PL1 is 15/28W but then not disclosing PL2. It won't matter for the GPU performance, will have a small impact on ST CPU performance and will have a noticable impact on MT performance.

How much of one? We'll have to wait for reviews, but this actually puts Rembrandt at a disadvantage in these slides vs Cezanne given how little the two scale with higher power (you can also see the diminishing returns with AMD's own numbers too).

The CZN laptop in question has a short PPT of 25W though when the STAPM power is 15W. Just as a baseline for the "15W" CZN performance numbers.

 

[leoneazzurro]

How is that bashing? He's just pointing out AMD is also resorting to misleading presentations. So pointing out the facts are called bashing?

He may have become salty(just guessing I don't know) but it's not bashing. Remember if you attack someone to change their views it often has the opposite effect.

He says even after that the gain will be 1.5-1.7x, which is really damn good, and they could have shown that instead.

Should always clamour for apples to apples comparisons, regardless of the size of the company, or your bias.

AMD presentation is misleading but truth to be said, "15W nominal" may be misleading as well as these processors have different power levels and in many cases they can use more than 15W. Then, one must consider that AMD wants to say that the 6800U is a direct replacement of 5800U in the same form factor laptops with same battery and thermals, that is, the comparison is to say "we can gain up to 2x graphic performance in the same chassis" more than "up to 2x at the exact same power consumption". .
And, while AMD's presentation is certainly misleading, what Techepifany does is misleading as well, because instead of testing the actual hardware in standard conditions and making a comparison based on actual numbers, he artificially limites the older APU he has to a certain power cap and compares the numbers he gets (without even knowing if he set the exact same graphic options in games, which has an effect as well) with the percentage shown as "up 2x" in a slide. Which is quite a very imprecise method to proceed, at least. And the conclusion he is deriving is on the tone "6000 series is bad and AMD wants to scam you" which he can tell after the tests the actual hardware, not when he is trying to prove it as a point since the beginning.
And if you follow the trend of his last tweets, it's all the same story. AMD deserves criticism when it does stupid things (and the presentation MUST be criticized for being shady, RX6500/6400 must be criticized for being bad products) but the criticism must be based on crystal clear numbers and apples to apples comparisons, and not on his personal bias or angryness .

 

[moinmoin]

Honestly the only chart that matters in mobile is the efficiency one. Just hope it's not too much cherry picked ("web browsing" should be close to general usage).

The best approach would be having benchmark results along with the amount of joules those ate to get there. Just give me the most performance for the least joules!

The interesting take-away is that: Even when operating with a higher 28W max TDP, the 6800U has a longer battery live as the 15W TDP 5800U in a general use case. With all those new power management features, adaptive power management framework and new sleep states it requires less watts over time in the general use case.

View attachment 55479

View attachment 55478

 

[leoneazzurro]

Omg, he published a video comparing performance numbers between a DDR-3200 and a DDR4-4266 (so, quite probably with IF overclock) - which is NOT LPDDR4-4266 by any means - to "prove" his idea.
:facepalm:

 

[Thibsie]

Omg, he published a video comparing performance numbers between a DDR-3200 and a DDR4-4266 (so, quite probably with IF overclock) - which is NOT LPDDR4-4266 by any means - to "prove" his idea.
:facepalm:

Si he's completely idiot or completely dishonest.

 

[mikk]

Omg, he published a video comparing performance numbers between a DDR-3200 and a DDR4-4266 (so, quite probably with IF overclock) - which is NOT LPDDR4-4266 by any means - to "prove" his idea.
:facepalm:

He don't need to prove, it's logical that 4266 is faster than 3200 on a 5800U because of the bandwidth bottleneck. It should definitely improve the fps by at least 10%

 

[DrMrLordX]

There is no Remembrandt laptops to buy.

Then why is he trying to extrapolate performance from an older product when he has nothing to test it against?