Question AMD Phoenix/Zen 4 APU Speculation and Discussion

[uzzi38]

I can finally make this thread.

https://twitter.com/x/status/1519669375283957760

Phoenix is indeed RDNA3. My advice to everyone: treat the old APU rumours as being out of date.

 

[Glo.]

21% mean with a dGPU vs competition rather than the iGPU.

EDIT: Or not according to the notes.

According to notes, its 21% over 1280p Iris G7 with 96EUs.

Its hilariously bad marketing at AMD. Hilariously. Considering that we know how Rembrandt performed compared to G7.

 

[Abwx]

AMD chose the HP Elitebook 840 G9, which actually has DDR5, and compared against their reference system equipped with LPDDR5. Moreover the HP system has 30W PL1/60W PL2 which means it will most likely throttle when pushed because HP isn't known for designing their laptops with the best possible cooling.

They should have compared it against the Lenovo Yoga Slim 9i, as it is one of the few 1280P models equipped with LPDDR5. It has PL2/PL1 of 52/35 W which means it should be possible to have it cooled better. And the results speak for themselves. There is a whopping 36% difference in CB R23 MT scores between the two models.

https://www.notebookcheck.net/HP-EliteBook-840-G9-laptop-review-The-Lenovo-ThinkPad-X1-Carbon-alternative.674583.0.html

https://www.notebookcheck.net/Lenovo-Yoga-Slim-9i-review-Arguably-the-fastest-Evo-Laptop.647754.0.html

AMD marketing at it again.

Never tired when it comes to bad mouthing...

How do you know that they didnt use the first run as basis..?..

Think about it, at 4nm process the 7040 has vastly better efficency than a 1280P, so a at 35-45W it should have no difficulty outperforming the 1280P@50W by something like 35%, after all there s a full node advantage here.

 

[tamz_msc]

How do you know that they didnt use the first run as basis..?..

Again you've demonstrated your poor reading comprehension.



The 9885 score is on the first run.



Presumably this is also the first run.

So, yeah ~36% performance advantage for Lenovo.

AMD always uses gimped configurations for the competitor when it comes to laptop comparisons. Earlier in last year's CES, they used the 12 W MX450 to compare against the 680M iGPU.

 

[IEC]

Laptop chips are notoriously hard to compare due to variance in cooling and implementation.

That's why I appreciated the opportunity to compare the Dell G15 5511 vs 5515 - basically the same laptop except Intel 11th gen vs Ryzen 5000 series. And since I didn't care about battery life and wanted Thunderbolt 4 for occasional DTR use, it made selecting the Intel part the easy choice.

If I wanted thin and light or good battery life, AMD's newest chips look promising. If there's one with good cooling and has cTDP options for CPU and GPU I could see replacing my G15. My one annoyance with it is how locked down the BIOS settings are, but I shouldn't have been surprised since it's Dell.

I await true apples to apples comparisons.

 

[Abwx]

Again you've demonstrated your poor reading comprehension.

View attachment 73977

The 9885 score is on the first run.

View attachment 73978

Presumably this is also the first run.

So, yeah ~36% performance advantage for Lenovo.

AMD always uses gimped configurations for the competitor when it comes to laptop comparisons. Earlier in last year's CES, they used the 12 W MX450 to compare against the 680M iGPU.

Because we should use a 52W rated 1280P (actually 64W PL2..) as comparison for a 35W 7040 to compare scores and then compute efficency as if it was the same power for both..?..

The runs at NBC are indicative that a laptop is using 64W CPU power while the other is at about 35W for the CB runs.

So AMD comparison is on point, that is, at equivalent powers, and in this case the 7040 has 35% better perf at said 35W real TDP.

 

[Abwx]

DP

 

[tamz_msc]

Because we should use a 52W rated 1280P as comparison for a 35W 7040 to compare scores and then compute efficency as if it was the same power for both..?..

Prove that the 7040 was running at 35W when running CB. You can't because it doesn't.

So AMD comparison is on point, that is, at equivalent powers, and in this case the 7040 has 35% better perf at said 35W real TDP.

There is no "real TDP". 35 W TDP does not it is pulling 35 W when running Cinebench. Power limits are entirely dependent on the OEM. 35 W -HS CPUs can be configured to run at 45 W. Many of them do.

AnandTech Forums: Technology, Hardware, Software, and Deals

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

www.anandtech.com

 

[insertcarehere]

They said 250+ designs from the start, this imply big quantities if each design is to break even rapidly enough.

AMD also said 200+ designs expected with Rembrandt and it took flipping ages to get significant volume out. That statement doesn't mean anything.

 

[leoneazzurro]

According to notes, its 21% over 1280p Iris G7 with 96EUs.

Its hilariously bad marketing at AMD. Hilariously. Considering that we know how Rembrandt performed compared to G7.

Numbers are quite strange indeed, especially because there are already tests of Rembrandt beating the Iris XE 96 EU with a RDNA2 iGPU clocked lower than what Phoenix should be:

Nsb9D59WMsD9LM4iLKpajE-1200-80.jpg (1200×514) (futurecdn.net)

It could depend on the test set used, also a problem can be that these iGPUs will be (again) limited by the memory bandwidth which is also shared with the CPU.
Rembrandt moved the performance up because of DDR5/LPDDR5, Phoenix will not have this advantage except maybe the capability to use some higher rated modules.
When looking at the few 6CU Rembrandt (the Radeon 660M) tests I could see that the performance of the 680M was far from being double of its smaller sibling, this pointing to a CPU or RAM BW limitation.
In any case, very strange. But as AMD managed to even screw the cache amount on their presentation, I would wait for the tests for a real performance value lol.

 

[Abwx]

Prove that the 7040 was running at 35W when running CB. You can't because it doesn't.


There is no "real TDP". 35 W TDP does not it is pulling 35 W when running Cinebench. Power limits are entirely dependent on the OEM. 35 W -HS CPUs can be configured to run at 45 W. Many of them do.

AnandTech Forums: Technology, Hardware, Software, and Deals

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

www.anandtech.com

35W on average for those who know what this term mean, i guess that you re stuck with Intel s methodology wich is to run CPUs at 64W and then sell the thing with a 15W TDP moniker, we re talking of real numbers here not of their usual shenanigans.

To summarize things the 7040 is 2x more efficient at same perfs, so there s nothing incredible that it can perform 35% better at same power.

Or do you think that Intel s process, wich is comparable, or a little better, than TSMC s 7nm can rival TSMC s 4nm, let alone 5nm..?...

 

[tamz_msc]

35W on average for those who know what this term mean, i guess that you re stuck with Intel s methodology wich is to run CPUs at 64W and then sell the thing with a 15W TDP moniker, we re talking of real numbers here not of their usual shenanigans.

To summarize things the 7040 is 2x more efficient at same perfs, so there s nothing incredible that it can perform 35% better at same power.

Or do you think that Intel s process, wich is comparable, or a little better, than TSMC s 7nm can rival TSMC s 4nm, let alone 5nm..?...

Shifting goalposts once again. First it was "first run vs sustained", then it became "real TDP" which doesn't exist and now it's "35 W average".

 

[Kaluan]

I sure hope that +21% in gaming performance against i7-1280P is a lower bound of some sort...

Pretty sure those numbers are analogus to these from the 7045 chips:


If they were showing more Thin & Light specific stuff (they weren't, since 7040U is MIA for now), I think they would've actually covered the iGPU.
But something also makes me think they're waiting on the driver team here. If you look at whatever endnotes are available, a lot of the Phoneix testing dates as far back as August last year, I doubt they had any sort of well working RDNA3 drivers back then.

 

[Abwx]

Shifting goalposts once again. First it was "first run vs sustained", then it became "real TDP" which doesn't exist and now it's "35 W average".

A CPU that use 50W for 10s and then 30W for the next 10s to perform a bench has used 40W on average and that s its real TDP for this task, if you dont understand such basics you are not fit for technical discussions...

 

[tamz_msc]

A CPU that use 50W for 10s and then 30W for the next 10s to perform a bench has used 40W on average and that s its real TDP for this task, if you dont understand such basics you are not fit for technical discussions...

Again with the BS arguments. Neither AMD nor Intel consume power specified by the TDP for a single CB run.

Damn pro-AMD charlatan.

 

[Exist50]

A CPU that use 50W for 10s and then 30W for the next 10s to perform a bench has used 40W on average and that s its real TDP for this task, if you dont understand such basics you are not fit for technical discussions...

The chips AMD showed have a 35W-45W TDP (base power), and it's very reasonable to assume AMD was using the higher number for performance comparisons. That's ignoring the boost entirely, which was surely enabled.

 

[Abwx]

Again with the BS arguments. Neither AMD nor Intel consume power specified by the TDP for a single CB run.

Damn pro-AMD charlatan.

The charlatan is the one who ignore measurements of the very sites he s reffering to, guess that when one is short of arguments he relies on ad hominems...

FTR NBC measured your two marvels power comsumption, get back reading the articles before spaming again by here...

 

[tamz_msc]

The charlatan is the one who ignore measurements of the very sites he s reffering to, guess that when one is short of arguments he relies on ad hominems...

FTR NBC measured your two marvels power comsumption, get back reading the articles before spaming again by here...

Middle row, right column, is the CPU package power for all the performance modes on the Lenovo. So yeah, it is consuming 52 W for the first 40 s, then 35 W for the remainder of the total run which is 4 min 20s, in the highest performance mode. So average power consumption = (52x40 + 35x220)/260 = 37.6 W.

Go back and redo your schooling.

 

[Abwx]

Middle row, right column, is the CPU package power for all the performance modes on the Lenovo. So yeah, it is consuming 52 W for the first 40 s, then 35 W for the remainder of the total run which is 4 min 20s, in the highest performance mode. So average power consumption = (52x40 + 35x220)/260 = 37.6 W.

Go back and redo your schooling.

They say that PL2 is actually set to 64W and work for 15s, that s enough for CB R15.

So you are completely out of reality, what you are displaying is an already hotted laptops since the first 15s at 64W are not displayed in this pic, it can be seen on their monitorless power test using CB though...

So much for schooling advices, actually you didnt even read the article extensively, just picked a graph randomly and then came with fantasies, as usual...

 

[tamz_msc]

They say that PL2 is actually set to 64W and work for 15s, that s enough for CB R15.

So you are completely out of reality, what you are displaying is an already hotted laptops since the first 15s at 64W are not displayed in this pic, it can be seen on their monitorless power test using CB though...

So much for schooling advices, actually you didnt even read the article extensively, just picked a graph randomly and then came with fantasies, as usual...

Go on. Try to make yourself look even more stupid than you really are.

 

[TESKATLIPOKA]

21% better performance than i7-1280P with 96 EU is pretty meh especially when you count in 25% higher graphics clock speed.

That's not IGP performance. Do you think Phoenix IGP regressed?

Ryan confirmed the L3 cache is 16MB not 32MB as shown on slide....come on, AMD shown some professionalism

That's not good. This can be hardly called professionalism. It's not like they did It a few hours before the presentation. Truly an embarrassment.

 

[Heartbreaker]

That's not IGP performance. Do you think Phoenix IGP regressed?

Where is that stated that it wasn't IGP?

According to note 7, it's AMD integrated 780M vs Intel Iris Xe

AMD Extends its Leadership with the Introduction of its Broadest Portfolio of High-Performance PC Products for Mobile and Desktop

— AMD Ryzen™ 7000X3D Series Desktop processors bring the power of AMD 3D V-Cache™ technology to gamers and creators — — AMD Ryzen 7000…...

ir.amd.com

The new AMD Ryzen 7040HS Series Mobile processors offer:

• Up to 34% faster multithreaded performance over the competition9
• Up to 21% faster gaming performance over the competition7

7 - Based on testing by AMD as of 12/23/2022. Testing results demonstrated in Borderlands 3, Cyberpunk 2077, Rainbow Six Siege, Assassin's Creed: Valhalla, World of Tanks Encore, League of Legends, Far Cry 6, Grand Theft Auto V, Shadow of the Tomb Raider, F1 2021, Strange Brigade, Total War: Three Kingdoms Battle. Ryzen™ 9 7940HS system: AMD reference motherboard configured with 4x4GB LPDDR5, Samsung 980 Pro 1TB SSD, Radeon 780M Graphics, Windows® 11 64-bit. Core i7-1280P system: HP Elitebook 840 G9 configured with 16GB DDR5-4800, 1TB SSD, Intel Iris Xe, Windows 11 64-bit. System manufacturers may vary configurations, yielding different results. PHX-9

I was definitely hoping for more than 20%.

 

[TESKATLIPOKA]

Again you've demonstrated your poor reading comprehension.

View attachment 73977

The 9885 score is on the first run.

View attachment 73978

Presumably this is also the first run.

So, yeah ~36% performance advantage for Lenovo.

AMD always uses gimped configurations for the competitor when it comes to laptop comparisons. Earlier in last year's CES, they used the 12 W MX450 to compare against the 680M iGPU.

I can only write this. ComputerBase
TDP gemessen is the measured TDP.

R9 6900HS 35/45W manages 11,434-12,438 points.
I7-1280P managed 13,504 points, which is 9-18% faster.

I think R9 7940HS should manage at least 20% higher score than 6900HS, that would be 13,721-14,926 points depending on TDP. This would mean 2-11% higher score.

BTW R7 7700X limited to 45W manages 15,167 points in CB R23.
Does anyone believe R9 7940HS would score worse?

 

[exquisitechar]

Where is that stated that it wasn't IGP?

According to note 7, it's AMD integrated 780M vs Intel Iris Xe

AMD Extends its Leadership with the Introduction of its Broadest Portfolio of High-Performance PC Products for Mobile and Desktop

— AMD Ryzen™ 7000X3D Series Desktop processors bring the power of AMD 3D V-Cache™ technology to gamers and creators — — AMD Ryzen 7000…...

ir.amd.com

I was definitely hoping for more than 20%.

Am I mistaken, or doesn't Rembrandt's iGPU already have approximately that level of performance or even more? Either Phoenix Point's performance improvement is nil or AMD made yet another mistake.

 

[jpiniero]

Am I mistaken, or doesn't Rembrandt's iGPU already have approximately that level of performance or even more?

More yes.

 

[Glo.]

Am I mistaken, or doesn't Rembrandt's iGPU already have approximately that level of performance or even more? Either Phoenix Point's performance improvement is nil or AMD made yet another mistake.

It has VASTLY higher gaming performance, than AMD touts for Phoenix.

iGPU-Vergleich: AMD Radeon 680M vs. Intel Iris Xe Graphics im Test

Wie schnell sind die schnellsten mobilen iGPUs von AMD und Intel in Spielen? Fünf Radeon und Iris Xe Graphics im Test.

www.computerbase.de

AMD Radeon 680M: Die Ryzen-6000-iGPU im Test: Benchmarks der Radeon 680M in Spielen

AMD Radeon 680M im Test: Benchmarks der Radeon 680M in Spielen / Leistung in FHD mit mittleren Details / Leistung in FHD mit hohen Details

www.computerbase.de