AMD Ryzen 5 2500U ‘Raven Ridge’ Mobile APU Spotted With Vega GPU

[Trixanity]

2 hours ago, Taf the Ghost said:

AMD's APUs are normally 35-45w parts, I believe. Though the CCX on the current 14nm process could do 800 Mhz for some ultra-low powered devices.

AMD's current APUs aren't in the same league though. Ryzen Mobile is supposed to be 50% lower power and the volume market is 15W I'd say -with 35-45W being second. So AMD needs to get down there and I do believe they've talked fanless designs which would have to be <10W to work unless they intend it to work with a giant heatsink.

[Misanthrope]

Given power usage, it might actually compete with a dedicated 1050 laptop for the far better battery life and portability this chip would afford. Seems interesting, quite a bit.

[Okjoek]

I'm not sold on an APU until they have motherboards smaller than ITX such as STX motherboards available to Intel.

[The Benjamins]

I found this, which shows the Ryzen 7 2700U being 20-30% faster.

 

and the GFX Bench shows the R7 2700U is more then 2x faster then the A12 9800E APU.

 

https://gfxbench.com/compare.jsp?benchmark=gfx40&did1=53156431&os1=Windows&api1=gl&hwtype1=iGPU&hwname1=AMD+Radeon(TM)+Vega+10+Mobile+Graphics&D2=AMD+A12-9800E+RADEON+R7%2C+12+COMPUTE+CORES+4C%2B8G

 

 

[RadiatingLight]

6 hours ago, Taf the Ghost said:

Possible. Though Intel's Mobile space is going to be a disaster for any consumer to sort out for the next ~8 months. They're going to have Cannonlake, Gemini Lake, Coffee Lake and I think some Kaby Lake parts still.

Gemini Lake?

WTF that's a new one.

[The Benjamins]

4 minutes ago, RadiatingLight said:

Gemini Lake?

WTF that's a new one.

That is a 4-6W CPU, it replaces Apollo lake.

[RadiatingLight]

30 minutes ago, The Benjamins said:

That is a 4-6W CPU, it replaces Apollo lake.

Fuck, I didn't even know Apollo lake existed.

too many lakes!

[The Benjamins]

Just now, RadiatingLight said:

Fuck, I didn't even know Apollo lake existed.

too many lakes!

honestly I just googled it.

[Doobeedoo]

Awesome, these will be quite an upgrade over last gen APUs by a lot. For laptops they will be very neat and on desktop will be great for entry builds.

[NumLock21]

First is was X399 and now it's 2500u...

[MageTank]

13 hours ago, Trixanity said:

It's intentional. Easier to follow the naming scheme of the company with the largest market and mind share. Helps consumers know how it stacks up. There is no reason for AMD to create disparity with a vastly different naming scheme unless they can come up with one that is so simple it doesn't matter but considering the amount of SKUs, it quickly becomes impossible to keep it simple, so it's easier to just follow the market leader.

It really doesn't. If anything, it's intentionally deceitful. They are mimicking the naming scheme, but using higher numbers on their products, almost as if they are trying to delude the "less than tech-savvy" into thinking the larger number means it's better. It also adds a layer of obfuscation that I myself just experienced yesterday. I went to buy a motherboard from Microcenter after returning my Z270 Apex. Told the man I needed a Z270 Taichi, he came back with an X370 Taichi from AMD. Granted, this man was a customer service rep from the returns desk, it's clearly obvious that he thought those products were either the same, or compatible.

 

This kind of mistake becomes even more tedious once you realize both AMD and Intel have lower-end B platforms (B250 Intel, B350 AMD) and with threadripper, we have X399 for AMD, X299 for Intel. You are right about this being intentional, but not in a way that "helps consumers". In my eyes, it will do the exact opposite.

 

There is nothing wrong with having your own product naming scheme and marketing for said naming scheme. To try to piggyback off the competitions marketing, while great from a competitive standpoint (forcing Intel to either risk further confusion by using the same names, or force them to change themselves), it's only going to make it more difficult for uneducated consumers to discern the difference between the two. 

[Agosto]

9 minutes ago, MageTank said:

This kind of mistake becomes even more tedious once you realize both AMD and Intel have lower-end B platforms (B250 Intel, B350 AMD) and with threadripper, we have X399 for AMD, X299 for Intel. You are right about this being intentional, but not in a way that "helps consumers". In my eyes, it will do the exact opposite.

 

There is nothing wrong with having your own product naming scheme and marketing for said naming scheme. To try to piggyback off the competitions marketing, while great from a competitive standpoint (forcing Intel to either risk further confusion by using the same names, or force them to change themselves), it's only going to make it more difficult for uneducated consumers to discern the difference between the two. 

I don't like X399 naming as well, I think they should have gone with X390 instead

[Trixanity]

10 minutes ago, MageTank said:

It really doesn't. If anything, it's intentionally deceitful. They are mimicking the naming scheme, but using higher numbers on their products, almost as if they are trying to delude the "less than tech-savvy" into thinking the larger number means it's better. It also adds a layer of obfuscation that I myself just experienced yesterday. I went to buy a motherboard from Microcenter after returning my Z270 Apex. Told the man I needed a Z270 Taichi, he came back with an X370 Taichi from AMD. Granted, this man was a customer service rep from the returns desk, it's clearly obvious that he thought those products were either the same, or compatible.

 

This kind of mistake becomes even more tedious once you realize both AMD and Intel have lower-end B platforms (B250 Intel, B350 AMD) and with threadripper, we have X399 for AMD, X299 for Intel. You are right about this being intentional, but not in a way that "helps consumers". In my eyes, it will do the exact opposite.

 

There is nothing wrong with having your own product naming scheme and marketing for said naming scheme. To try to piggyback off the competitions marketing, while great from a competitive standpoint (forcing Intel to either risk further confusion by using the same names, or force them to change themselves), it's only going to make it more difficult for uneducated consumers to discern the difference between the two. 

We don't know if the Ryzen Mobile lineup will be deceitful or not yet. I agree that in some instances it can be deceitful but aren't they only using a higher number on their chipsets? Ryzen lineup is 1xxx whereas Intel is 7xxx. Z270 sounds similar to X370 but it really isn't. You might as well blame the OEM for putting the same branding (eg. Taichi) on all platforms. Yes, ultimately the blame lies with AMD but in theory OEMs could salvage it themselves.

 

Honestly though I feel like motherboards should have the socket in the model name. That fixes any and all confusion.

It's odd that AMD leaned so close to Intel in their chipset names. Chipsets/motherboards are secondary to choosing the processor. You buy a processor and pick a motherboard for it; not the other way around. So ultimately it's pointless to try to one-up on that front.

 

Just for the record I wasn't even thinking about motherboards when I made the comment you quoted. I was only thinking of processors. You have a point but I feel like it's not as important as you make it out to be albeit annoying and pointless.

[Mira Yurizaki]

I can't be the only one who went "AMD, are you still trying to piss off Intel?" with that U moniker.

[MageTank]

Just now, Trixanity said:

We don't know if the Ryzen Mobile lineup will be deceitful or not yet. I agree that in some instances it can be deceitful but aren't they only using a higher number on their chipsets? Ryzen lineup is 1xxx whereas Intel is 7xxx. Z270 sounds similar to X370 but it really isn't. You might as well blame the OEM for putting the same branding (eg. Taichi) on all platforms. Yes, ultimately the blame lies with AMD but in theory OEMs could salvage it themselves.

 

Honestly though I feel like motherboards should have the socket in the model name. That fixes any and all confusion.

It's odd that AMD leaned so close to Intel in their chipset names. Chipsets/motherboards are secondary to choosing the processor. You buy a processor and pick a motherboard for it; not the other way around. So ultimately it's pointless to try to one-up on that front.

 

Just for the record I wasn't even thinking about motherboards when I made the comment you quoted. I was only thinking of processors. You have a point but I feel like it's not as important as you make it out to be albeit annoying and pointless.

With the CPU's themselves, their naming is fine. The chipsets are where the biggest problems lie. By having nearly identical chipset names, you run the risk of your customers buying the wrong motherboard to go with their CPU, all because you decided to piggyback off a known product naming scheme. 

 

I also do not see how you can possibly shift the blame to the OEM even slightly. You can't blame the board partners, not even a little bit, for using their normal naming schemes on their products. What is ASRock's Taichi lineup? It's simple. It's what they call their premium black and white motherboards. Same with MSI's Krait lineup. That would be like blaming ASUS for using Strix on both Intel and AMD platforms, or MSI for using "Pro Carbon" on both. Those names are used to segment the aesthetics and features within their own product stack, and are not associated with the name of the motherboard/platform chipsets AT ALL. It's literally zero fault of the board partners when AMD chooses to use X399 against Intel's X299. If you expect board partners to change their product naming because AMD intentionally put them into that situation, then I honestly don't know what to say at that point.

 

AMD could have chosen literally any combination of numbers and letters. They could have named it X400, or A/B/Z300 for Ryzen, literally anything would be better than B350 or X370 (mimicking both the B250 and Z370 Intel consumer chipsets). X400 would be a bigger number than X299, and would be much easier to differentiate the products from your competitor, giving you the "larger number" advantage without obfuscating the process.

 

I just cannot defend AMD here. If they really do go through with this 2500U nomenclature for their mobile products, then I will chalk this up as intentional deceit. The irony being, I am currently writing this post on my laptop with an Intel Core i5 5200U, lol. 

[Trixanity]

1 minute ago, MageTank said:

With the CPU's themselves, their naming is fine. The chipsets are where the biggest problems lie. By having nearly identical chipset names, you run the risk of your customers buying the wrong motherboard to go with their CPU, all because you decided to piggyback off a known product naming scheme. 

 

I also do not see how you can possibly shift the blame to the OEM even slightly. You can't blame the board partners, not even a little bit, for using their normal naming schemes on their products. What is ASRock's Taichi lineup? It's simple. It's what they call their premium black and white motherboards. Same with MSI's Krait lineup. That would be like blaming ASUS for using Strix on both Intel and AMD platforms, or MSI for using "Pro Carbon" on both. Those names are used to segment the aesthetics and features within their own product stack, and are not associated with the name of the motherboard/platform chipsets AT ALL. It's literally zero fault of the board partners when AMD chooses to use X399 against Intel's X299. If you expect board partners to change their product naming because AMD intentionally put them into that situation, then I honestly don't know what to say at that point.

 

AMD could have chosen literally any combination of numbers and letters. They could have named it X400, or A/B/Z300 for Ryzen, literally anything would be better than B350 or X370 (mimicking both the B250 and Z370 Intel consumer chipsets). X400 would be a bigger number than X299, and would be much easier to differentiate the products from your competitor, giving you the "larger number" advantage without obfuscating the process.

 

I just cannot defend AMD here. If they really do go through with this 2500U nomenclature for their mobile products, then I will chalk this up as intentional deceit. The irony being, I am currently writing this post on my laptop with an Intel Core i5 5200U, lol. 

As I said: board partners could salvage it if need be. They obviously didn't feel the need. I'm not saying what they should or shouldn't do with their branding (if they feel like Taichi denotes black and white then they're entitled to it but they could just have called it black/white - that would be simple). And I did offer a suggestion: socket name prominently displayed in the model name. That solves any problem with horrendous and confusing naming schemes. If you can buy the wrong board with the socket in the title then it's your fault but this solution is too obvious so no one will do it.

 

How is 2500U on laptops deceitful? Intel's current SKUs are in 7000 series and are about to enter 8000 series. Worst case scenario people will think they buy Sandy Bridge (a 2011 processor) at which point they'll probably skip it for being lower than 8000. There is no way AMD is gaining an advantage through deceit by calling it 2500U. And from what I can see Intel has never had a SKU called 2500U. Closest I can think of is their core i5 2500 desktop processor. Or are you saying that any name previously used should be retired?

[MageTank]

24 minutes ago, Trixanity said:

As I said: board partners could salvage it if need be. They obviously didn't feel the need. I'm not saying what they should or shouldn't do with their branding (if they feel like Taichi denotes black and white then they're entitled to it but they could just have called it black/white - that would be simple). And I did offer a suggestion: socket name prominently displayed in the model name. That solves any problem with horrendous and confusing naming schemes. If you can buy the wrong board with the socket in the title then it's your fault but this solution is too obvious so no one will do it.

 

How is 2500U on laptops deceitful? Intel's current SKUs are in 7000 series and are about to enter 8000 series. Worst case scenario people will think they buy Sandy Bridge (a 2011 processor) at which point they'll probably skip it for being lower than 8000. There is no way AMD is gaining an advantage through deceit by calling it 2500U. And from what I can see Intel has never had a SKU called 2500U. Closest I can think of is their core i5 2500 desktop processor. Or are you saying that any name previously used should be retired?

It's not the board partners job to salvage it, nor should they attempt to. It will only cause further issues between choosing particular boards from various manufacturers. They would need every board partner to universally agree to do so, and that just won't happen. Also, "ASRock black and white" is nowhere near as catchy or elegant as "ASRock Taichi". Their aesthetic even uses the Yin-Yang symbol, so it's more to it than just a name. 

 

As for the whole "socket in the title" argument, you again wouldn't NEED to do this, if you would just name your product in a distinctive manner. You are suggesting that everyone else go out of their way to appease AMD's silly naming crusade against Intel, and I just don't see the point. AMD is at fault here, not the board partners.

 

As for how the 2500U is deceitful, it's rather obvious. They are further taking Intel's naming conventions (using U for ultra low-power), and continuing to use the EXACT SAME chipset names (AMD's AM4 promontory lineup of X370/B350) which will no doubt cause an issue with the upcoming coffeelake Union Point variant (which is bound to use HM370) as consumers might get confused by the chipset offerings of both mobile product stacks if they only look at random slides or video reviews.

 

Again, it's not an issue of the CPU names that bother me (aside from their blatant use of the U to mock/directly contrast Intel's U lineup), it's the chipset names. If AMD was going to use an entirely different chipset name for their mobile SKU's, it wouldn't be that big of a deal. Hopefully they come to their senses, and change that before launch. We only know now what we knew back in March, which is that AMD plans to use the exact same chipsets on mobile, as they do on desktop. 

 

At this point, there is no defending AMD's current naming practice. It's a clear dig at Intel, and we get that. From a business standpoint, it makes perfect sense to preemptively step on their toes with this current naming scheme, but it hurts AMD's reputation as the "pro consumer" company that their die-hard fans genuinely believe them to be. Furthermore, it hurts unsuspecting consumers that fail to properly differentiate the products. I imagine you will be very sore if you bought an X370 mobile Ryzen platform, because you thought it supported thunderbolt after seeing Intel's HM370 (presumably, not yet confirmed) variant supporting thunderbolt for coffeelake.

 

TL:DR? This problem in naming stems from AMD and AMD alone, nobody else. If anyone get's confused by this, it's AMD's fault for intentionally taking a dig at Intel. 

[Cookybiscuit]

900W Threadripper + Vega64 APU when?

[Taf the Ghost]

3 hours ago, Doobeedoo said:

Awesome, these will be quite an upgrade over last gen APUs by a lot. For laptops they will be very neat and on desktop will be great for entry builds.

The Raven Ridge APUs really should be the final culmination of where AMD has been heading in the laptop space. It'll do light gaming just fine, while also being a great media player and a really good CPU.

[MageTank]

Just now, Taf the Ghost said:

The Raven Ridge APUs really should be the final culmination of where AMD has been heading in the laptop space. It'll do light gaming just fine, while also being a great media player and a really good CPU.

Give me a 13 inch laptop with a 120hz 720p Freesync panel, and a Zen APU that can drive that refresh rate. Talk about the best gaming netbook, lol. 

[The Benjamins]

20 minutes ago, MageTank said:

Give me a 13 inch laptop with a 120hz 720p Freesync panel, and a Zen APU that can drive that refresh rate. Talk about the best gaming netbook, lol. 

Make it 1080p and you got a deal. I prefer 1080p for every day tasks and videos.

[MageTank]

Just now, The Benjamins said:

Make it 1080p and you got a deal. I prefer 1080p for every day tasks and videos.

For videos, I'd understand, but 720p at 13 inches is actually more pixels per inch than 1440p at 27 inches, or 1080p at 21.5 inches. It would be far better for battery life (a 13 inch chassis is likely not going to have the biggest battery) and will allow the APU to hit higher framerates. Best compromise at that point, would be a 1440p panel at 13 inches, and you manually run it at 720p. There will be no int - float conversion in that setup. For example: 

Quote

LCD panels are fixed pixel devices, their native resolution as it were
1440/1080 is a float, fixed pixel devices vs. floating point numbers = rescaling and interpolation, which results in poor picture quality
1440/720 is a int, 2 in this case. No rescaling or interpolation required. A 1440p panel displays a 720p image natively, just slightly uglier (screen size vs resolution debacle)

With this compromise, you get the option of more desktop real estate at the cost of battery life, or the other option of changing it down to 720p without dealing with heavy interpolation. Can we agree on that then, and send the details to ODM's to make it happen?  

[Kamina]

Is AMD mocking Intel with all their marketing brands?

 

2500U - Intel U lineup

X399 - X299

x370 - z270

B350 - B250

Ryzen 3, Ryzen 5, Ryzen 7 - i3, i5, i7

 

I'm not the only one noticing this, right?

[Dabombinable]

41 minutes ago, MageTank said:

For videos, I'd understand, but 720p at 13 inches is actually more pixels per inch than 1440p at 27 inches, or 1080p at 21.5 inches. It would be far better for battery life (a 13 inch chassis is likely not going to have the biggest battery) and will allow the APU to hit higher framerates. Best compromise at that point, would be a 1440p panel at 13 inches, and you manually run it at 720p. There will be no int - float conversion in that setup. For example: 

With this compromise, you get the option of more desktop real estate at the cost of battery life, or the other option of changing it down to 720p without dealing with heavy interpolation. Can we agree on that then, and send the details to ODM's to make it happen?  

*looks at 720p & 1080p 10" IPS screens on own tablets*

[Scitesh]

17 hours ago, Taf the Ghost said:

AMD's Radeon division is really the best iGPU company, which is the reason the Vega is what it is. The really important detail is that Raven Ridge will be 1 CCX (so 4 cores with ultra-low latency) with a solid iGPU at a low power rate. On the assumption Intel doesn't find an illegal way to prevent it from going into the Laptop space, it should be a true gem there.

 

Other detail is the naming. "2500U" suggests Raven Ridge is actually on Zen+, which will be very interesting if true. Might mean Ryzen Refresh is almost exactly 1 year after Ryzen launched.

I am going to go out on a limb and say some non techie is going to confuse i5 2500u with R5 2500u and say i5 8500u is better coz bigger no/newer processor.. Also be careful of ultra savvy rswaps and eBay sellers

17 hours ago, Trixanity said:

I'd say that's a bit pre-mature. AMD might have gone full Intel (and you never go full Intel) and randomly change the numbers like their enthusiast chips with last year's architecture sharing SKU numbers with this year's mainstream chip or their iGPU having a different number than the CPU (it pisses me off that one is "behind" the other). So it might just be to denote that it's a newer chip (perhaps with some hardware fixes at best). Of course all speculation.

The problem with that is that the mobile market is very diverse now so that you need to denote the TDP as they range from 5W to 50W (more or less) in the mobile segment, so you need a suffix or another naming scheme to differentiate the TDP via the name alone.

Intel

Desktop mainstream :leads

Hedt:lags by one year (for 6+cores)

Laptop high end :leads mainstream by 3 months

Laptop mid end :confused

Low power :more confusion... I mean what the fuck is Apollo Lake

 

Amd:

Desktop mainstream :leads

Hedt:leads

Laptop:either skips a generation and loads of lags by an year

Ultra low power :WTF??