AMD Ryzen 7040 Series "Phoenix Point" Mobile Processor I/O Detailed: Lacks PCIe Gen 5

[Bruteque]

Way to miss the point.
We're looking for APUs without dGPUs for the last three pages and that's all you quote out of context? Context like the next entire paragraph of my post you just quoted.
Jesus.

I don't know about other regions, but these have been widely available in the U.S. for a few months now. A friend got his from Micro Center; I got mine from Amazon. It's an awesome laptop. The 6800H iGPU + 16GB LPDDR5-6400 performance is noticeably behind 6900HS iGPU + 16GB DDR5-4800, but most games are playable at 1080p on business trips.

Amazon.com

The problem with AMD is that they don't offer nearly the level of OEM support that Intel does. Case in point, the two USB-C ports on my 6800H Zenbook 14 Flip OLED have PCIe tunneling hardware and are technically USB4. However, AMD USB4 drivers for the integrated chipset still have all kinds of issues. Worse yet, AMD does not offer nearly the kind of support Intel does for OEMs on resolving these issues, so AMD driver issues in practice don't get resolved in a timely manner. In the end, ASUS is forced simply to advertises the USB-C ports on the Zenbook as 3.2, even though they are hardware USB4. You in turn get flip-flopping firmware updates which frustrate users.

If this was Intel, they would have entire OEM support teams helping ASUS get the software to work on the ASUS hardware, updating Intel drivers based on the interaction to get the job done as necessary, but to AMD, "it's just one laptop model." A lot of the time the trouble with buying an AMD laptop is that you get significantly worse software/firmware/driver support from the OEMs, which is a direct consequence of the OEMs getting significantly worse software help from AMD. Until AMD improves their OEM support, it will be extremely difficult for AMD to make big gains on laptop market share, even if they manage to release a killer product.

There is only so much headache most users will put up with for the performance increase of going from Iris Xe to 780M. For most users, having a laptop that "just works" is their top priority. If AMD laptops continue to be as much of a headache as Intel ARC-equipped laptops, then the audience will continue to be limited to the select few highly headache-resistant.

 

[Mawkzin]

The source of this news what btarunr linked can be set to english.
Here are the links:
Ryzen 9 7940HS
Ryzen 7 7840HS
Ryzen 5 7640HS

P.S. L3 is only 16MB. Less than what Dragon Range(Raphael) has.

Looking to the Ryzen 5 7640 the igpu has 8 CU this time instead of 6 CU, so maybe with rdna3 it uses 3 wgp with 4 CU each instead of 2 wgp with 6 CU each in the rembrand series.

 

[TumbleGeorge]

You can doubt it all you wish. Market trends say otherwise.

And is it possible that on the sign of your store, it is written under the name, "we only sell gaming laptops" and therefore only gamers visit you as customers?

More like €500 in germany at least.

In this website cheapest offers with Ryzen 7 6800U start from 1299€, little cheap has one or two models with Ryzen 7 6800HS. Significantly expensive than 5000U offers ~2X...

The source of this news what btarunr linked can be set to english.
Here are the links:
Ryzen 9 7940HS
Ryzen 7 7840HS
Ryzen 5 7640HS

P.S. L3 is only 16MB. Less than what Dragon Range(Raphael) has.

Maybe we has been a some difference between models for China and for...outer world where part of them will be artificially limited.


PS. I have an idea why limitations. Because official support of more RAM, maybe even more cache from Windows is more expensive. Recently, in China, they are betting on their own operating systems...

 

[Garrus]

No. 6800U with 28W is for that purpose in ultralight and portable laptops. "People just wanted" is too vague.
- there are mini-PC with 6800H and without mGPU, if you wish
- 6800H APU can come with mGPU. I have one at home from Asus.
- you can also configure some laptops, such as Lenovo, and choose 6800H without mGPU. So, there is a choice for consumers.

oh god, no point in responding, you're completely wrong

just go to bestbuy and find that stuff you are talking about, it's not there

I'm just so happy that AMD's APU will finally get the 35W it needs to shine. Hopefully we get more than a SINGLE Asus model at Bestbuy without dGPU after 2 years. :/

 

[john_]

For some use cases, not all. For other use cases an IGP is the preferred choice. It depends on what someone wants. My earlier point was that the complains about systems having a DGPU are silly. ALL large laptop makers have more IGP based models than models with a DGPU. Only small makers(Razer, etc) have more DGPU models. There is no shortage of IGP based laptops.

They do have iGPU models, probably Intel. To be honest I haven't look at Intel and in any case I don't even consider most Intel iGPU only models an option, when needing something with more than good enough CPU and more than average GPU. Intel options are probably excellent for media playback/transcoding. I only look at AMD options and there, the models with Ryzen 6000 APUs and no discrete GPUs are limited, at least in my country.

More like €500 in germany at least.

I am saying in the same post you are quoting, that I got mine at 500 euros. That's why I also say that they got pricier compared to the time I bought mine.

You can always buy a laptop with 6800U or HS APU, which more often do not come with mGPU.
Plus, some vendors like Lenoco allow you to pre-configure your laptop and even choose 6900HX without mGPU.


Lenovo allows you configure your own laptop. You can get even 6900HX wihtout desrete graphics card. Where is the problem?

Those laptops are usually expensive.
Is configuring a laptop available in all regions, or only US? And is the price (almost) the same as a typical laptop, or higher considering it obviously needs more steps to be done from the manufacturer before delivering?


Anyway. If there are enough models without discrete GPUs at price points much lower than those equivalents with discrete GPUs, everything is fine. But there aren't always. People pointing at "no discrete GPU" models, shouldn't be counting Atom based models. Yes they are without discrete GPUs, but they are crap. We are compaining about strong iGPUs getting pared with what someone would describe as mid to low end discrete GPUs. Putting a GTX 1650 or an RTX 3050 next to a 6600U, does make sense for people not having another system, but when the "no discrete GPU" options are limited, or even non existed(in some markets), that's at least annoying. I was ready to get a 6600U/H in last summer, with no discrete GPU and at a price point not much over 700 euros, but there was nothing available. Gone with a 5600U option. But as I said in a previous post, here 6000 series are either with a discrete GPU, or just expensive, meaning the strong iGPU advantage is non existing.

 

[ToTTenTranz]

Some of the cache has evaporated overnight, or something...

https://twitter.com/x/status/1610974871575744514

On one hand it's strange that they'd make such a big mistake on something like an official presentation slide, and then repeat that same mistake across all 3 models.
On the other hand, that slide only mentions "total cache", and doesn't specify whether it's L3+L2+L1 or something else.

At the same time, they haven't really said much about that RDNA3 iGPU with 6 WGPs running at ~3GHz (despite being the highest-clocked RDNA3 GPU to date and the only one at N4).
The Radeon 780M seems to have >150% of the processing throughput of the 680M in Rembrandt while the total RAM bandwidth only goes up from 6400MT/s to 7500MT/s.
This is only a 17% boost which is hardly enough to keep up with the faster CPU cores which will undoubtedly demand more bandwidth by themselves.

One thing that could make sense is that the new RDNA3 iGPU actually has 16MB of Infinity Cache all to itself. It would explain the 40MB "total cache" typo, and provide an adequate boost to effective bandwidth to the substantially more powerful iGPU.

Or that really was just a mistake in those slides and this iGPU is going to be massively bottlenecked by memory bandwidth.

 

[trsttte]

On one hand it's strange that they'd make such a big mistake on something like an official presentation slide, and then repeat that same mistake across all 3 models.
On the other hand, that slide only mentions "total cache", and doesn't specify whether it's L3+L2+L1 or something else.

At the same time, they haven't really said much about that RDNA3 iGPU with 6 WGPs running at ~3GHz (despite being the highest-clocked RDNA3 GPU to date and the only one at N4).
The Radeon 780M seems to have >150% of the processing throughput of the 680M in Rembrandt while the total RAM bandwidth only goes up from 6400MT/s to 7500MT/s.
This is only a 17% boost which is hardly enough to keep up with the faster CPU cores which will undoubtedly demand more bandwidth by themselves.

One thing that could make sense is that the new RDNA3 iGPU actually has 16MB of Infinity Cache all to itself. It would explain the 40MB "total cache" typo, and provide an adequate boost to effective bandwidth to the substantially more powerful iGPU.

Or that really was just a mistake in those slides and this iGPU is going to be massively bottlenecked by memory bandwidth.

I think it was a simple copy paste mistake because of Dragon Range mobile cpus that are based on the desktop silicon, it's the simplest explanation

 

[THANATOS]

Looking to the Ryzen 5 7640 the igpu has 8 CU this time instead of 6 CU, so maybe with rdna3 it uses 3 wgp with 4 CU each instead of 2 wgp with 6 CU each in the rembrand series.

WGP is a Dual Compute unit(CU). Phoenix has 6 WGPs or in other words 6 dual compute units, people like to say 12CU. There is no difference compared to Rembrandt.
There should be 2 shader engines(arrays) with 3 WGPs per shader engine, just like what you see in that block diagram. Ryzen 5 7640 will lose a WGP per shader engine, which leaves 4WGPs or 8CUs.
It looks like in Rembrandt the 6CU version has half of shader engines deactivated, that's why only 1/2 of CU is left.

Maybe we has been a some difference between models for China and for...outer world where part of them will be artificially limited.


PS. I have an idea why limitations. Because official support of more RAM, maybe even more cache from Windows is more expensive. Recently, in China, they are betting on their own operating systems...

Phoenix has only 16MB of L3 cache physically, the same as Rembrandt.
There is no artificial difference based on region. AMD marketing team did a poor job by not noticing they wrote the wrong amount of cache in some press slides, which caused this mess about L3 cache.

On one hand it's strange that they'd make such a big mistake on something like an official presentation slide, and then repeat that same mistake across all 3 models.
On the other hand, that slide only mentions "total cache", and doesn't specify whether it's L3+L2+L1 or something else.

At the same time, they haven't really said much about that RDNA3 iGPU with 6 WGPs running at ~3GHz (despite being the highest-clocked RDNA3 GPU to date and the only one at N4).
The Radeon 780M seems to have >150% of the processing throughput of the 680M in Rembrandt while the total RAM bandwidth only goes up from 6400MT/s to 7500MT/s.
This is only a 17% boost which is hardly enough to keep up with the faster CPU cores which will undoubtedly demand more bandwidth by themselves.

One thing that could make sense is that the new RDNA3 iGPU actually has 16MB of Infinity Cache all to itself. It would explain the 40MB "total cache" typo, and provide an adequate boost to effective bandwidth to the substantially more powerful iGPU.

Or that really was just a mistake in those slides and this iGPU is going to be massively bottlenecked by memory bandwidth.

There was no mention anywhere about Phoenix having Infinity cache. 40MB is a typo.

That processing throughput increase is in reality only 25% higher (3GHz vs 2.4GHz). The rest of that increase comes from dual issue CUs, and for that to work, you need VOPD instructions and even then you can't use It for everything, there are limitations and also shaders need to be optimized for It. Even AMD says RDNA3 architecture is only 17.4% faster per clock than RDNA2. There is a reason why they say N31 has only 6144 stream processors, which is only 20% more than Navi21.
BTW, If AMD really couldn't meet the required BW, then they wouldn't use 6 WGPs, If they performed so much better.

Chip & Cheese made a great RDNA3 architectural analysis. Go, check It out.

 

[TumbleGeorge]

WGP is a Dual Compute unit(CU). Phoenix has 6 WGPs or in other words 6 dual compute units, people like to say 12CU. There is no difference compared to Rembrandt.
There should be 2 shader engines(arrays) with 3 WGPs per shader engine, just like what you see in that block diagram. Ryzen 5 7640 will lose a WGP per shader engine, which leaves 4WGPs or 8CUs.
It looks like in Rembrandt the 6CU version has half of shader engines deactivated, that's why only 1/2 of CU is left.


Phoenix has only 16MB of L3 cache physically, the same as Rembrandt.
There is no artificial difference based on region. AMD marketing team did a poor job by not noticing they wrote the wrong amount of cache in some press slides, which caused this mess about L3 cache.


There was no mention anywhere about Phoenix having Infinity cache. 40MB is a typo.

That processing throughput increase is in reality only 25% higher (3GHz vs 2.4GHz). The rest of that increase comes from dual issue CUs, and for that to work, you need VOPD instructions and even then you can't use It for everything, there are limitations and also shaders need to be optimized for It. Even AMD says RDNA3 architecture is only 17.4% faster per clock than RDNA2. There is a reason why they say N31 has only 6144 stream processors, which is only 20% more than Navi21.
BTW, If AMD really couldn't meet the required BW, then they wouldn't use 6 WGPs, If they performed so much better.

Chip & Cheese made a great RDNA3 architectural analysis. Go, check It out.

Maybe Xilinx IP will help for better training how to set hardware parameters for better presence depending of game or other app on this PC.