Question AMD Rembrandt/Zen 3+ APU Speculation and Discussion

Page 80 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Kaluan

Senior member
Jan 4, 2022
429
895
96
A first tangible design for refreshed or rebranded mobile Rembrandt SKUs:


Likely called Ryzen 7735H/HS, so similar or identical to 6800H/HS. Also seems to come with a Radeon RX 7000S dGPU. Dunno if Navi32 or 33 based.
 

scineram

Senior member
Nov 1, 2020
339
268
106
Wow, 7000S already? Probably launches at CES. S series is low power for thin and light, so probably N33. Was there any N22 is 6000S? I think N23 only.
 
  • Like
Reactions: Kaluan

ahimsa42

Senior member
Jul 16, 2016
225
194
116
Wow, 7000S already? Probably launches at CES. S series is low power for thin and light, so probably N33. Was there any N22 is 6000S? I think N23 only.
launch? if rembrandt is any indication, these probably won't be available until mid 2024 at best. maybe by then we will have some good 6800U options too.
 

LightningZ71

Golden Member
Mar 10, 2017
1,533
1,766
136
I haven't seen many good comparisons on rembrandt between native 1440, 1440 with XeSS, 1440 with FSR, and 1080p minimum. It would be i terrsting to see how capable it is with the upscaling techs and how Phoenix behaves by comparison in the future.
 
Last edited:

tamz_msc

Diamond Member
Jan 5, 2017
3,366
3,299
136
Just picked up a Thinkpad P16s with the 6850u. Let me know if you have any questions.
Well you can do the honors by testing the performance delta on battery and plugged in. Set Best Performance in Power Mode in Settings -> System -> Power & battery. Keep everything else on default. Run Geekbench 5 and Speedometer 2.0. Report the Geekbench ST score and Speedometer result when on battery and when plugged in.
 

tamz_msc

Diamond Member
Jan 5, 2017
3,366
3,299
136
Geekbench 5 Plugged In:
Single-Core:1521
Multi-Core: 8811

Geekbench 5 Battery:
Single-Core:1361
Multi-Core: 6078
That's a 11% loss in ST while on battery. My 11370H (which is basically configured like a 1165G7) loses only 4% on battery. So this is an area where AMD has to continue improving.

Try the same thing with Speedometer 2.0, with a Chromium based browser. That will align with a more real-world scenario.
 

lightmanek

Senior member
Feb 19, 2017
360
683
136
That's a 11% loss in ST while on battery. My 11370H (which is basically configured like a 1165G7) loses only 4% on battery. So this is an area where AMD has to continue improving.

Try the same thing with Speedometer 2.0, with a Chromium based browser. That will align with a more real-world scenario.
You can't determine that without knowing how long this task will run for on a same capacity battery as this might simply be an OEM decision to target long battery life versus burst performance.
On the other hand, Intel's 10nm process is somewhat better suited to higher clocks, therefore not loosing efficiency as quickly going up in clocks (before hitting the wall).
 

tamz_msc

Diamond Member
Jan 5, 2017
3,366
3,299
136
You can't determine that without knowing how long this task will run for on a same capacity battery as this might simply be an OEM decision to target long battery life versus burst performance.
On the other hand, Intel's 10nm process is somewhat better suited to higher clocks, therefore not loosing efficiency as quickly going up in clocks (before hitting the wall).
That's why I said select Best Performance setting. That overrides any OEM specific optimisations and ensures that Turbo/Boost is attained quicker.
 

Abwx

Diamond Member
Apr 2, 2011
9,868
2,313
136
That's a 11% loss in ST while on battery. My 11370H (which is basically configured like a 1165G7) loses only 4% on battery. So this is an area where AMD has to continue improving.
It s the other way around, it s Intel that has to improve something...

In single core power is to be limited as well as it s well over 15W in your laptop and this considerably reduce the battery life, it was once proved in this forum that battery life is optimal when CPU power is equal to rest of the system power, so certainly lower than 15W for the CPU.
 
  • Like
Reactions: igor_kavinski

lightmanek

Senior member
Feb 19, 2017
360
683
136
That's why I said select Best Performance setting. That overrides any OEM specific optimisations and ensures that Turbo/Boost is attained quicker.
I think BIOS still is in charge of power limits regardless of power settings in Windows. That's the reason we have such a wide spread of performances between different laptop models using technically the same CPU.

I could test my Ryzen laptop, but I'm rocking an old 4700U HP Envy model ...
 

igor_kavinski

Diamond Member
Jul 27, 2020
7,586
4,387
106
That's a 11% loss in ST while on battery. My 11370H (which is basically configured like a 1165G7) loses only 4% on battery. So this is an area where AMD has to continue improving.
They probably already have done the best they can. I think your Intel CPU will burn through the battery lot faster than the Ryzen. It's a trade-off. You may be able to prevent the end of the world with a quicker calculation in a Mission Impossible type of scenario but the Ryzen will work longer and keep communication alive with the human using it in case of an emergency situation.
 
  • Like
Reactions: lightmanek

tamz_msc

Diamond Member
Jan 5, 2017
3,366
3,299
136
They probably already have done the best they can. I think your Intel CPU will burn through the battery lot faster than the Ryzen. It's a trade-off. You may be able to prevent the end of the world with a quicker calculation in a Mission Impossible type of scenario but the Ryzen will work longer and keep communication alive with the human using it in case of an emergency situation.
Well if Ryzen sips battery because it doesn't want to go fast and thereby causing sluggish performance, then that is a bad compromise.
I think BIOS still is in charge of power limits regardless of power settings in Windows. That's the reason we have such a wide spread of performances between different laptop models using technically the same CPU.

I could test my Ryzen laptop, but I'm rocking an old 4700U HP Envy model ...
PL2 or short-duration power limit doesn't affect ST performance that much, unless you reduce it drastically. Like I said, choosing Best Performance overrides BIOS and OEM optimizations governing boost behavior.
It s the other way around, it s Intel that has to improve something...

In single core power is to be limited as well as it s well over 15W in your laptop and this considerably reduce the battery life, it was once proved in this forum that battery life is optimal when CPU power is equal to rest of the system power, so certainly lower than 15W for the CPU.
Tiger Lake does just fine in the battery life department. It matches Cezanne while being bested by Rembrandt, so it ain't too bad for an older node that is primarily frequency optimized unlike the TSMC 7nm and derivative nodes used by AMD in its mobile processors.
 
  • Like
Reactions: scineram

Abwx

Diamond Member
Apr 2, 2011
9,868
2,313
136
Well if Ryzen sips battery because it doesn't want to go fast and thereby causing sluggish performance, then that is a bad compromise.
It would be more efficient than TGL even without ST frequency cutdown since it consume considerably less in the 4GHz range.

As for being sluggish that s as bad faith as possibly can be, if it s sluggish at 1361 GB ST points it will be as much sluggish at 1521 pts, you really think that such small of a difference could be perceptible elsewhere than in your forever anti AMD rants..?..
 

scineram

Senior member
Nov 1, 2020
339
268
106
Geekbench 5 Plugged In:
Single-Core:1521
Multi-Core: 8811

Geekbench 5 Battery:
Single-Core:1361
Multi-Core: 6078
M2 Air gets 1890/8733. LMFAO absolute rekt! And half the cores are efficiency, no SMT.
This is not to diss your laptop. I would get that over any Mac. Just pissing on Apple silicon disrespecters.
 
Last edited:

gdansk

Golden Member
Feb 8, 2011
1,292
1,275
136
M2 Air gets 1890/8733. LMFAO absolute rekt! And half the cores are efficiency, no SMT.
This is not to diss your laptop. I would get that over any Mac. Just pissing on Apple silicon disrespecters.
M2 may be pretty good but ironic pro-gamer posts are still bad.
 
  • Like
Reactions: scineram

tamz_msc

Diamond Member
Jan 5, 2017
3,366
3,299
136
It would be more efficient than TGL even without ST frequency cutdown since it consume considerably less in the 4GHz range.

As for being sluggish that s as bad faith as possibly can be, if it s sluggish at 1361 GB ST points it will be as much sluggish at 1521 pts, you really think that such small of a difference could be perceptible elsewhere than in your forever anti AMD rants..?..
It is impossible to give proper replies to you because you shift goalposts with every comment. When did efficiency become the topic under consideration?

It is a fact that all Ryzen mobile APUs deliver perciptably lower performance, with the decrease being worse as you go back each generation, when on battery as compared to plugged in. We have the data proving so. Just wait for something real-world like Speedometer to be run on battery to see just how sluggish it can be.
 

moinmoin

Diamond Member
Jun 1, 2017
4,174
6,250
136
What has that got to do with performance delta on battery and plugged in?
There is no performance delta. What you are talking about is the delayed boosting behavior of mobile Zen cores which is completely negligible on longer running tasks.
 

Abwx

Diamond Member
Apr 2, 2011
9,868
2,313
136
It is impossible to give proper replies to you because you shift goalposts with every comment. When did efficiency become the topic under consideration?
Efficency has always been the main topic once we re talking of mobile devices, and in this case it s counter productive to use the CPU at full frequency on ST since it would exceed 15W and hence reduce battery life uselessly.

Now if you dont bother for this metric you can always set a power profile that would negate this limitation, but then the advertised run times would be no more accurate.
 
  • Like
Reactions: moinmoin

tamz_msc

Diamond Member
Jan 5, 2017
3,366
3,299
136
There is no performance delta. What you are talking about is the delayed boosting behavior of mobile Zen cores which is completely negligible on longer running tasks.
You literally have the data right here posted in this very thread.
 

ASK THE COMMUNITY