AMD launches Ryzen Mobile 7 2700U & 5 2500U with Vega Graphics

[moinmoin]

Yes, it's Zen+ cores.

Zen+ as a name actually isn't officially used. There is Zen(1) on 14nm+, which we now know is 12LP, so Pinnacle Ridge. The Raven Ridge die contains rather significant improvements compared to the Zeppelin die, like Precision Boost 2, that warrant increasing the gen in the model number. But Raven Ridge still is on 14LPP.

 

[trparky]

Has anyone done the tests? Why I'm asking is that I'm interested to see if Zen+ has closed the IPC gap between AMD and Intel when it comes to single threaded performance. The AMD/Intel IPC gap with Ryzen v1.0 was the major reason why I chose not to go with Ryzen this time around and to stay with my current Core i5 3570k @ 4.4 GHz based system, I had a bad feeling that going to Ryzen (at least the current version of Ryzen) would be a downgrade as versus an upgrade. If Zen+ has truly increased IPC and closed the gap it may be time to do the upgrade when Ryzen v2.0 comes out next year.

Edit
Crap, I got this post in before moinmoin got his post in. I guess my post above doesn't mean much then. Oh well.

 

[CatMerc]

It wouldn't be the first time different architectures or even processes are mixed in the same generational name.

As for Raven Ridge specifically, as far as I'm aware, it's not Zen+. The Zen+ variant of Raven is called Picasso, and it will come in late 2018 or early 2019. That said the lines are a bit blurred as to what constitutes Zen+. Raven Ridge will have an updated SMU for Precision Boost 2 and no doubt various other IP blocks that are more up to date than Ryzen's. Memory controller might even be the one we're going to see in Pinnacle Ridge.

So, to be honest, it's a question that even after the dust settles and all chips are released and analyzed, we might disagree about. (Unless AMD officially names them)

 

[neblogai]

Has anyone done the tests? Why I'm asking is that I'm interested to see if Zen+ has closed the IPC gap between AMD and Intel when it comes to single threaded performance. The AMD/Intel IPC gap with Ryzen v1.0 was the major reason why I chose not to go with Ryzen this time around and to stay with my current Core i5 3570k @ 4.4 GHz based system, I had a bad feeling that going to Ryzen (at least the current version of Ryzen) would be a downgrade as versus an upgrade. If Zen+ has truly increased IPC and closed the gap it may be time to do the upgrade when Ryzen v2.0 comes out next year.

I think the answer to why Raven Ridge is named as 2xxx series is simple: AMD can not have new laptop chips in 2018 with last generation naming. New chips would look old. Pinnacle Ridge will be named as 2xxx- so Raven Ridge gets that naming as well. Now, regarding the IPC- it seems to be where it can be expected to be- ~5% below desktop Ryzen. That is because RR cores are about the same, upgraded only with power saving features- but L3 cash on RR is only 4MB. Ryzen desktop processors showed a ~5% performance drop when L3 was cut from 16mb to 8MB, and RR (judging from RR Cinebench results and clocks) we will have the same small drop.
As for the i5 3570K- even current Ryzen would be a good upgrade, as many new games suffer lower minimum fps with just 4c/4t. That, of course, also depends on your GPU. Pinnacle Ridge should be faster than current Summit Ridge and Raven Ridge- so it should be an even better upgrade.

 

[Yeroon]

This may seem like a really dumb question but... why the 2 in 2700u and 2500u? Wouldn't the 2 indicate Zen+? I guess my question is why the mobile chip numbers start with a 2 when the desktop versions start with a 1.

Its still the first Zen core. AMD was asked about this, and they stated that there were a lot of smaller features (mainly power control) that they had time to add to RR that weren't ready for the desktop processors when those launched, that it was worth using the 2000 series name. Its possibly in tthe AT article on RR.

 

[french toast]

It wouldn't be the first time different architectures or even processes are mixed in the same generational name.

As for Raven Ridge specifically, as far as I'm aware, it's not Zen+. The Zen+ variant of Raven is called Picasso, and it will come in late 2018 or early 2019. That said the lines are a bit blurred as to what constitutes Zen+. Raven Ridge will have an updated SMU for Precision Boost 2 and no doubt various other IP blocks that are more up to date than Ryzen's. Memory controller might even be the one we're going to see in Pinnacle Ridge.

So, to be honest, it's a question that even after the dust settles and all chips are released and analyzed, we might disagree about. (Unless AMD officially names them)

Well, zen+ is not officially a name for anything.
Rumoured picasso is based on 7nm zen 2.
Pinnacle ridge is based on a minor revision of zen 1on 12nm (14nm+ ) ...this has been coined by many including myself as "zen +"....as to what changes ( if any) will arrive to this 12nm core is up for guesswork outside of bug fixes, power saving features not included in zen 1 and clock hangups.

With any luck this "zen +" will get an IPC bump of ~5%.

 

[Arachnotronic]

AMD chips have been going into cheap laptops because they sucked. Raven Ridge doesn't appear to suck.

Do you honestly not think that this would be a good option for Apple's dGPU-less laptops, like the 13" MBP? They'd be replacing the 28W i5-7267U with Raven Ridge, getting two extra cores and a fantastic GPU in the process. Sounds like a good product to me.

The next generation 13" MBP will almost certainly have a Coffee Lake-U chip (4+3e). That brings the extra two cores (and they're better cores) and an iGPU that'll be close enough to the Raven Ridge one but with much better/more efficient media.

 

[AtenRa]

The next generation 13" MBP will almost certainly have a Coffee Lake-U chip (4+3e). That brings the extra two cores (and they're better cores) and an iGPU that'll be close enough to the Raven Ridge one but with much better/more efficient media.

And double the price

 

[CatMerc]

Well, zen+ is not officially a name for anything.
Rumoured picasso is based on 7nm zen 2.
Pinnacle ridge is based on a minor revision of zen 1on 12nm (14nm+ ) ...this has been coined by many including myself as "zen +"....as to what changes ( if any) will arrive to this 12nm core is up for guesswork outside of bug fixes, power saving features not included in zen 1 and clock hangups.

With any luck this "zen +" will get an IPC bump of ~5%.

Picasso is still listed as "Raven Ridge architecture" like Pinnacle is listed as "Summit Ridge". Matisse, the chip after Pinnacle Ridge, clearly states Zen 2. So you would expect Picasso to also list Zen 2 if it was.

The next generation 13" MBP will almost certainly have a Coffee Lake-U chip (4+3e). That brings the extra two cores (and they're better cores) and an iGPU that'll be close enough to the Raven Ridge one but with much better/more efficient media.

I haven't seen any proper comparison of media blocks efficiency. The one Raven Ridge video test appears to be software rather than hardware accelerated, most likely because the media hardware is new (newer than Vega's even), and needs to be supported in software.

As for "better cores", that's debatable. Intel cores have a higher performance ceiling, but under these power constrained scenarios we don't have a proper comparison.

 

[NTMBK]

The next generation 13" MBP will almost certainly have a Coffee Lake-U chip (4+3e). That brings the extra two cores (and they're better cores) and an iGPU that'll be close enough to the Raven Ridge one but with much better/more efficient media.

The GPU is closer, but still behind Raven Ridge. Apple have a long history of prioritising GPU performance (most recently choosing i5 + GT3e configurations, as opposed to an i7 with weaker GPU), and Raven Ridge is definitely in the "good enough" CPU range, I would say. It's still a jump in performance over the previous dual core model (even if not quite as powerful as a CFL-U CPU), it's not like they're going backwards.

 

[jpiniero]

For the GPU, I think Apple cares more about compute performance than anything else... and Iris actually doesn't do that badly there. Plus you get the edram helping out the CPU. So you can see why they had used it versus the regular U model in cTDP Up.

 

[Jan Olšan]

Apple might like having a Radeon even in the machine with iGPU, but there are factors speaking against it:

- inertia
- keeping Intel is likely less effort, less investment needed. I think Apple favours that strongly, at least in their PC business
- they probably fear that AMD would "taint" their "we're so premium and better" image in the eyes of the fans that need to believe they are getting the best and in their brandname ladder, Intel is the premium, AMD would be suspected as "budget". Note that actual performance and abilities don't matter, it's simply the perception and status image. (Which is why I never believed those forum rumours about Llano being seriously considered by Apple.)

 

[CatMerc]

Apple might like having a Radeon even in the machine with iGPU, but there are factors speaking against it:

- inertia
- keeping Intel is likely less effort, less investment needed. I think Apple favours that strongly, at least in their PC business
- they probably fear that AMD would "taint" their "we're so premium and better" image in the eyes of the fans that need to believe they are getting the best and in their brandname ladder, Intel is the premium, AMD would be suspected as "budget". Note that actual performance and abilities don't matter, it's simply the perception and status image. (Which is why I never believed those forum rumours about Llano being seriously considered by Apple.)

They've had no problems having AMD's brand coming from their GPU's. I don't think the brand would be an issue, as most just see "Apple" and don't know anymore.

 

[french toast]

Kind of peeved and didn't just create their own 45w hbm Apu instead of letting intel into that sector.
They could easily have charged 250$ for such a chip and make a killing...ate into intel and Nvidia sales and built a solid brand.

Instead intel gets all the kudos and high margins, intel don't have the IP to.compete and yet amd just helped them.
I hope they bring out such an Apu next year...what a waste otherwise.

 

[CatMerc]

Kind of peeved and didn't just create their own 45w hbm Apu instead of letting intel into that sector.
They could easily have charged 250$ for such a chip and make a killing...ate into intel and Nvidia sales and built a solid brand.

Instead intel gets all the kudos and high margins, intel don't have the IP to.compete and yet amd just helped them.
I hope they bring out such an Apu next year...what a waste otherwise.

An HBM APU at 250$ is wishful thinking for the 2017-2018 timeframe. I don't see how you can get any margins on that.

 

[Dayman1225]

Another thing stopping RR in Macbooks at least (Laptops) would be the lack of support of LPDDR, something Apple exclusively used for their laptops, no?

 

[Gikaseixas]

Kind of peeved and didn't just create their own 45w hbm Apu instead of letting intel into that sector.
They could easily have charged 250$ for such a chip and make a killing...ate into intel and Nvidia sales and built a solid brand.

Instead intel gets all the kudos and high margins, intel don't have the IP to.compete and yet amd just helped them.
I hope they bring out such an Apu next year...what a waste otherwise.

I see this situation as "take this and in return you give me this right here"
Intel has many interesting things in the oven right now. Let's see how this deal plays out for each side

 

[french toast]

Apple might like having a Radeon even in the machine with iGPU, but there are factors speaking against it:

- inertia
- keeping Intel is likely less effort, less investment needed. I think Apple favours that strongly, at least in their PC business
- they probably fear that AMD would "taint" their "we're so premium and better" image in the eyes of the fans that need to believe they are getting the best and in their brandname ladder, Intel is the premium, AMD would be suspected as "budget". Note that actual performance and abilities don't matter, it's simply the perception and status image. (Which is why I never believed those forum rumours about Llano being seriously considered by Apple.)

Yes but you are forgetting one important thing.
- Apple like money...they like to have multiple sources of components and encourage competition to drive down BOM costs.

Going with AMD would be far cheaper than intel..plus it serves intel notice that they have to be more reasonable with their prices next round as apple will swap chips if not.

 

[Dayman1225]

Going with AMD would be far cheaper than intel..plus it serves intel notice that they have to be more reasonable with their prices next round as apple will swap chips if not.

Well we cannot say that much for sure as we do not know how much Intel make Apple pay for these chips, nor do we know if they subsidize it even further.

 

[Gikaseixas]

Another scenario, possibilities - Next MBP lineup could have both AMD and Intel cpus at different performance tiers.

Edit: That could apply some pressure on both vendors and offer Apple better pricing

 

[scannall]

Another thing stopping RR in Macbooks at least (Laptops) would be the lack of support of LPDDR, something Apple exclusively used for their laptops, no?

They target a specific power and performance budget for the whole system. If Raven Ridge can fit that, then they could go that route. Intel has been disappointing on the GPU side of things for a while. Not to mention late with their product lines. That doesn't mean Apple WILL go with AMD, just that they could.

 

[Dayman1225]

They target a specific power and performance budget for the whole system. If Raven Ridge can fit that, then they could go that route. Intel has been disappointing on the GPU side of things for a while. Not to mention late with their product lines. That doesn't mean Apple WILL go with AMD, just that they could.

That's true, but doesn't Intel have much more efficient media decoders in their IGPUs (at least for 4k)?While their IGPUs in pure performance are MUCH weaker. Really curious to see CFL - U with GT3E

Old roadmap so things may have changed (such as dates)

 

[scannall]

That's true, but doesn't Intel have much more efficient media decoders in their IGPUs (at least for 4k)?While their IGPUs in pure performance are MUCH weaker. Really curious to see CFL - U with GT3E

Old roadmap so things may have changed (such as dates)

OS X uses the GPU to drive much of their interface, so it does matter quite a bit. As far as media decoders go their own A series chips are the best around at that. And they have been using them in Macbooks already for a while.

 

[Dayman1225]

And they have been using them in Macbooks already for a while.

...Have I missed something?

 

[scannall]

...Have I missed something?

They use a low power A series CPU to drive their touchbar, and assorted other functions. Increasing the load, or shuffling decode for example to a more powerful chip wouldn't be all that difficult.