No rebrand here. It's a new design with a new architecture revision (full Zen+, unlike the "halfway to Zen+ that RR was) built on a slightly improved process node (12nm is nothing more than a refined 14nm, but it _is_ refined, after all). Is this a real generational update? Arguably not, but these are bound to carry a lot of refinements and bug fixes from the 1st-gen chips.
The statement "Furthermore, AMD uses wafers with highest yields/least amount of defective parts to build the Ryzen Pro in a bid to meet long term reliability." can not be true unless Ryzen Pro is nothing more than a rebrand of existing 2200G-2400G-2600G chips, unless all the discards are going to "athlons". They aren't going to sell enough athlon pros to justify such a wafer binning.
The mobile Pro parts are a specific binning of the mobile part, not the desktop parts. These have to run at vastly different TDPs so it really makes no sense to not have a specific die for it.
I have a Radeon Pro 2700U and I can tell you that it also has a non-pro counterpart with the same paper specs. The main difference is that the Pro gets certified by ISVs which is important for commercial software users (hint: some licenses expressly forbid running on non-certified hw). There also seems to be some indication that Pro parts can support ECC memory better than non-pro, at least for mobile. They don't need to do special wafer binning to turn things on and off like that, but they may choose to bin only the dies with the fewest defects as Pro models, and then shuffle overflow into the normal non-pro bins.
Also the PRO parts for Ryzen that I have seen compared to standard (example) Ryzen 2700 vs Ryzen 2700 PRO...same exact specs, but, the PRO out does the non pro for some reason in most things, not massive margin, but better none the less, interesting because prior to such, near 100% cases, desktop enjoyed the "quickest" speed wise and spec wise often enough were faster than PRO, Nv changed a touch with Ti stuff and Titan, AMD with Vega and prior to that Fury line and such.
It should :kind of; be that way, because the PRO parts likely have been "voltage tuned" right off the bat, so is like taking Vega 56, drop core volts done right you automatically gain speed and reduce power consumption (by large amount if you get a golden chip)
maybe AMD "noticed such" with Ryzen, and said "cool" so they voltage tune, which would make absolute sense if nothing else, it would have to be "at the factory" only thing I can see being the reason, normally PRO does not run games as well (as an example) cause they turn down clocks to keep power in check by default spec, Ryzen PRO, not so much.....voltage tuned ? for sure bin tested and moved around absolutely, but they seem to be pretty much same price sometimes less then the desktop equivalent model, but run faster and cooler .... freebee for the PRO crowd instead of other way round? (which also makes sense, gamers while in the tens or hundreds of millions are but a drop in the bucket of the computing world, significant, but just a portion none the less.
server, workstation, designers, bank institute etc is what currently forms the "world" mobile is it's own closed in beast of course
I take that back, it was 2+4, i did not notice. still 65W into laptop with 15W without boost clock change? I don't think so. Would go full throttle in first 10 sec....
65w is on-par with many Intel solutions using a 15w CPU and a discrete 50w GPU.
Intel's Cannon Lake has such a configuration.
What's important is AMD's solution has superior CPU and GPU performance at the same power envelope, but that has yet to be established so we will have to see...
People are upset that AMD is doing great. So they call AMD's new release at a lower lithography a rebrand. While apparently Intel's 5th or 6th generation on 14nm are not rebrands according to the same Intel fangirls. Give me a break.
These are no more a re-brand than the change from Ryzen 1000 to Ryzen 2000 series was - it's the same changes (process and architecture tweaks) trickling down to their mobile parts. Based on early evidence from other sites, it looks like they may have substantially reduced their idle power problems, too.
If they tidied up the idle power then these will be great. The first generation where solid in performance, but the battery drain kept them out of the high-end.
But MatebookD with 2500U has great battery life (in addition to great performance). So, the fault was not on APUs, but on laptop manufacturers and their lazy designs.
Have the same and concur, however I usually do not let it idle on battery for hours. Since it boot up really fast, there is no need for that.
But yes, battery life is great - can use it over several days (not the whole day obviously) without the need to recharge which is a fist for me with AMD based laptops.
I was thinking the same thing: they clearly had a power issue before and they don't talk about power draw at all. Which can only mean they didn't fix it.
And tbh, I really, really, really don't care about Vega being faster. For workloads that don't need a dGPU, Intel has enough HP already. Sure, I'll take better encoding/decoding capabilities, better outputs and whatnot. But all this "oh, look, Vega is so much faster" talk, does nothing for me.
Sounds like youa re not the market for Ryzen then.
I really, really, really care about Vega being faster. For workloads that use GPUs in more then a passing fashion, a more capable iGPU is always welcome. Some of us dont want to have to shell out $700 for a bottom tier "gaming" laptop just to have more power then intel was putting in their chips in 2013.
With this regard, the Matebook with the 2500U is really great - my kid can play Fornite @ Full HD resolution with Medium details.
The sad thing is it runs smoother than on a "four" core Trinity based Desktop with a 4 GB RX550, SSD and 16 GB RAM...I think it's really about time I upgrade my Desktop PC, as well.
AMD announced months ago that they would be expanding the powergating properties on Picasso.
In the first gen of mobile processors the iGPU never completely powered down when not in use. They have addressed that with Picasso along with other less efficient properties of the first generation.
I love that the day has come when you can truly recommend AMD over Intel for an overwhelming number of scenarios and applications.
No reason to buy Intel anymore, unless there is some specific code where Intel's arch behaves better and you care not about all the other benefits AMD brings.
One aspect that might hurt AMD here is that any prospective buyer of Ryzen Pro systems knows that better-performing 7 nm Ryzen chips are coming soon. So, in a perverse way, Intel's laziness or incompetence in moving below 14 nm works in their favor, especially with business customers: there is little risk that they will make substantial improvements in the next 12 months, so you might as well buy now.
It all depends on when you need it and what you are buying. Most companies refresh at least a fraction of their machines every year, with very few companies replacing all machine every couple of years. As the 2XXX series systems get retired from the channel, they will get replaced by 3XXX and that will be all you can buy until next spring when 7nm Ryzen Mobile is available. There's no reason to wait on for 7nm if you need to buy laptops this year. Desktops and workstations are another story though.
That depends on AMD's ambition. If they desired, they could bring a high-performance 7nm 8core CPU to mobiles this year, by creating a special, low power I/O die (with basic graphics/VPU as well), and coupling it with 7nm Zen2 chiplet. That chiplet would probably not be able to powergate as well as a chip designed specifically for mobile, but at 7nm, and with low power I/O, that should still be enough for high performance laptops to have decent idle battery life.
Very unlikely. They will be maximizing 7nm throughput for server and desktop parts that bring in all the money. Even if they had production capacity, it still wouldn't make sense. We are way too early in the design cycle for the 4XXX series parts to be even a consideration. Expect Q4 announcements with full launch early next year. With the delay going on with desktop 3XXX and Navi, there's not even remotely a chance of 7nm mobile this year. Also you are forgetting that mobile packages need to be small for wide adoption. It makes no economical sense to add an I/O die for a single CPU die, even with graphics in the mix.
Sorry- I don't get any of your arguments. 1) AMD has almost no presence in high performance/gaming laptops. It takes time to build up presence there. It would take negligible number of 7nm capacity for AMD to start that entrance. 2) Naming is the least of the issues for a company- only users make big deal out of it. These mobile chips could be mobile 3xxx H-series R7 or R9 without any problem. 3) There is no date set for 7nm monolithic APUs, only several year old slide indicates it should come in late 2019. And it also indicates that it would be power efficient 10-35W 4-core only designs, not high performance parts. 4) There is no delay with 3xxx or Navi. 3xxx was promissed to come out mid-2019- and we will be getting 3xxx and Navi announcements 27th of May. 5) AMD needs a mobile I/O die if they 8-core for want mobile and embedded markets. Packaging 2 chips together is certainly not an issue (Intel packages 3 chips for Kaby Lake-G). Investment from AMD would not be big, because most of the mobile I/O design elements could be reused from RR and from Matisse I/O. It's all on 14nm. So, it only depends on AMD's ambition, if they try to expand into high-performance mobile and embedded, or not.
1) You seriously underestimate the cost of small runs, especially on a new process node. 2) I was referring to naming because that's likely what they will be called next year, not because AMD couldn't release a faster chip under the current branding. 3) I am estimating based on what we have seen in the past for rolling out to new nodes. AMD has almost always launched the mobile parts on a new node and microarchitecture after their desktop counterparts. 4) A few different sources have stated that there was an issue in Fall of 2018 that held back the 3XXX parts from being released earlier in 2019 to mirror the launch of the 2XXX parts. It wasn't until after that that AMD came out and officially started saying later in 2019. One would assume Navi may have been impacted as well. 5) False. They would simply go with a larger monolithic package. The cost of integrating multiple chiplets onto a single package for what would be a < 300mm^2 die makes no sense. Interposers are expensive and add height to the package. Spacing of the chiplets makes the package larger in area. Both of which are tradeoffs that make no sense for mobile designs. Intel uses EMIB which allows them to do this without an interposer which means shorter packages. AMD have no publicly announced corresponding IP. I forget where Papermaster answered a question about this, but AMD have already indicated that they would not be using chipsets for this generation of mobile and desktop APUs.
1) Old 14nm process for I/O die, while half the silicon is aleady done in Zen2 chiplet. So releasing it would be cheap, and proper use of chiplet architecture to expand addressable market. 2) Not doing in now is a waste of opportunity. 3) Does not have to be later this time, because Zen2 chiplet is done, while I/O chiplet does not need any new tech, only a copy of what they have already done. 4) Rumours, and no leaks of actual schedule. 5) No interposer or EMIB needed, and no need for new Zen2 chiplet- only new 14nm I/O die, and mounting the two dies together on FP5. AMD said Matisse (AM4 package) will not have GPU die chiplets, that is all. And it makes sense, because Matisse is for desktop, and that would be unnecessary and expensive- same as making a monolithic powerful 7nm mobile APU, while Zen2 chiplet is already there waiting.
1) Old Process != cheap. Zen2 chiplets do not have the same power management features as the mobile parts. Mobile Ryzen doesn't have enough PCI-E lanes for an IF link to an external I/O die. 2) Not really. High-end parts are low volume and would not chip in high enough quantities to offset the tooling and marketing costs, for AMD or OEMs. 3) Again, Zen2 chiplets of today do not have the power-saving tech found in the mobile parts. 4) Fine. Time will tell. 5) An interposer or other integration tech is 100% required to connect the packages together. FP5 is just a normal BGA package where the traces connect the die contacts to the BGA pads. You would still need another layer to connect the packages together or you would have to route through to the motherboard and back which adds latency, power consumption, and design cost. No one would do this. Even if you were stupid enough to do so, the GPU/IO die you have proposed on 14nm would barely fit on the package and only be as fast as last gen. You underestimate how much of the APU die is GPU (hint: over half). The reason that matisse won't have a GPU chiplet version is far simpler: limited memory channels. The size of GPU that you could fit on matisse would require 2 additional memory channels just to keep the GPU fed. The current APUs are already bottle-necked by having to share two channel with a GPU on a 4C/8T system. This is why increasing memory clockspeed (bandwidth) has so much uplift on gaming benchmarks. Trying to do that with 8C/16T and a GPU that's at least as powerful as a 560 and the same number of memory channels is not remotely feasible. I can guarantee you that when the 16C/32T matisse parts get reviewed we will see two memory channels become a bottleneck and that is why AMD made the L3 cache so much larger (to reduce the impact).
1) Old process, and a lot of reused blocks (Vega, not Navi; same display, multimedia engines, etc) = way, way cheaper and could be done for this summer, compared to trying to build everything new on 7nm for next year. 2) Still needed if AMD aims to enter the high performance mobile/embedded market. Entering with an expensive design is even worse, because sales will not jump up anyway, and will only ramp up slowly, over many years. 3) With an energy efficient I/O and iGPU, Zen2 7nm chiplet with behavior tuned for efficiency would be good enough for long battery life, radically better than current desktop chips + dGPUs in certain Asus/Acer laptops. 5) Interpozer is not needed, same like it is not needed for Epyc/Threadripper, or Matisse. So, it is cheap. And you are thinking wrong about the GPU for it. For such mobile chip, mobile iGPU would be needed for extreme energy efficiency at idle, multimedia playback and video outputs, not for rendering games. So, 1-3CU in total, and no iGPU memory bandwidth issues. For Embedded and HPE- that is enough, and for gaming machines- proper dGPU would be added and switched on while gaming.
1) You are neglecting tooling costs. Small runs on a 1000+ step process are not cheap, especially for something they won't sell more that a few tens of thousands of. 2) They don't want to or they would have by now. What you have proposed IS an expensive design that no OEM would buy. Even if they could, it takes time to roll something like that out which AMD is no doubt planning on doing this fall with OEMS for next spring. It's too late in the year for something like that and we've had no leaks whatsoever. 3) Not really. Even if you can dramatically reduce dynamic power consumption by running at lower TDPs and idle clocks, you still have to properly manage static power dissipation. Desktop parts typically have an order of magnitude higher static dissipation than their mobile counterparts. This is precisely why Ryzen Mobile 3XXX is not using the same die as the desktop version. That fact doesn't change for your imagined 7nm chiplet. When TSMC is talking about 0.5X power consumption they are talking about dynamic, not static. 5) Epyc, Threadripper, and Matisse all use interposers. Not sure where you are getting your information from, but it's flawed. Infinity Fabric links are routed on the interposer between the chiplets directly in Zen1 and now to the I/O die in Zen2, not through the package substrate or the motherboard. Only the socket-to-socket IF links actually reach the LGA pads and motherboard.
1-3 CU? I thought we were talking about a supposed high-end mobile part -_- . Now you're talking about embedded and HPE. Even Intel's low-end mobile parts have more power that that. If AMD want a high performance mobile part, they would just shut off the built-in GPU, raise the TDP to 30W+ and then bundle a dGPU in the 550 or so range. No switchable graphics needed as Polaris can get very low idle power when undervolted and underclocked. Or they could go back to the Bulldozer era and pair an equally performing dGPU to match the iGPU and have you use crossfire. The GPU on ryzen mobile is deliberately undersized for the CPU because of memory bandwidth. You're grasping at straws now mate.
Well, you were never going to get me to agree with you :p You keep arguing about ambitious and lofty goals that AMD doesn't have the kind of spare cash to justify chasing. If this were Intel we were talking about, none of this would be unreasonable. Instead it's a company that makes an order of magnitude less revenue each quarter. Don't get me wrong, I like AMD, but this just isn't how they do business.
Well, they could be making more cash by reusing their tech more widely. They actually seem to try to do that with scaling Zen and Zen2 dies for several markets at once. But yes- this would be one more project, and maybe AMD do not have the capacity to pursue everything. Simply, this is my take on what would be very reasonable for them to do, in contrast to AdoreTV rumors of Zen2 desktop APU with 14-20CUs- which do not make sense.
You can now run the regular Adrenalin drivers, battery life is great if the laptop is properly configured and built and power usage is also fine. Get 9+ hours on my 2500U based laptop.
The reported issue was idle only, so if you get a laptop to let it idle for hours on battery then maybe mobile Ryzen is not for you
Is AMD going to do this again? Because if they do, then mobile 2nd gen will be just as much of a sales failure as the 1st gen. Some of us dont want to wait a year before our products work like they are supposed to.
We’ve updated our terms. By continuing to use the site and/or by logging into your account, you agree to the Site’s updated Terms of Use and Privacy Policy.
Facebook
Twitter
RSS
40 Comments
Back to Article
TristanSDX - Tuesday, April 9, 2019 - link
nothing is better than solid rebrandValantar - Tuesday, April 9, 2019 - link
No rebrand here. It's a new design with a new architecture revision (full Zen+, unlike the "halfway to Zen+ that RR was) built on a slightly improved process node (12nm is nothing more than a refined 14nm, but it _is_ refined, after all). Is this a real generational update? Arguably not, but these are bound to carry a lot of refinements and bug fixes from the 1st-gen chips.wumpus - Tuesday, April 9, 2019 - link
The statement "Furthermore, AMD uses wafers with highest yields/least amount of defective parts to build the Ryzen Pro in a bid to meet long term reliability." can not be true unless Ryzen Pro is nothing more than a rebrand of existing 2200G-2400G-2600G chips, unless all the discards are going to "athlons". They aren't going to sell enough athlon pros to justify such a wafer binning.SaberKOG91 - Tuesday, April 9, 2019 - link
The mobile Pro parts are a specific binning of the mobile part, not the desktop parts. These have to run at vastly different TDPs so it really makes no sense to not have a specific die for it.I have a Radeon Pro 2700U and I can tell you that it also has a non-pro counterpart with the same paper specs. The main difference is that the Pro gets certified by ISVs which is important for commercial software users (hint: some licenses expressly forbid running on non-certified hw). There also seems to be some indication that Pro parts can support ECC memory better than non-pro, at least for mobile. They don't need to do special wafer binning to turn things on and off like that, but they may choose to bin only the dies with the fewest defects as Pro models, and then shuffle overflow into the normal non-pro bins.
Dragonstongue - Friday, April 12, 2019 - link
Also the PRO parts for Ryzen that I have seen compared to standard (example) Ryzen 2700 vs Ryzen 2700 PRO...same exact specs, but, the PRO out does the non pro for some reason in most things, not massive margin, but better none the less, interesting because prior to such, near 100% cases, desktop enjoyed the "quickest" speed wise and spec wise often enough were faster than PRO, Nv changed a touch with Ti stuff and Titan, AMD with Vega and prior to that Fury line and such.It should :kind of; be that way, because the PRO parts likely have been "voltage tuned" right off the bat, so is like taking Vega 56, drop core volts done right you automatically gain speed and reduce power consumption (by large amount if you get a golden chip)
maybe AMD "noticed such" with Ryzen, and said "cool" so they voltage tune, which would make absolute sense if nothing else, it would have to be "at the factory" only thing I can see being the reason, normally PRO does not run games as well (as an example) cause they turn down clocks to keep power in check by default spec, Ryzen PRO, not so much.....voltage tuned ? for sure bin tested and moved around absolutely, but they seem to be pretty much same price sometimes less then the desktop equivalent model, but run faster and cooler .... freebee for the PRO crowd instead of other way round? (which also makes sense, gamers while in the tens or hundreds of millions are but a drop in the bucket of the computing world, significant, but just a portion none the less.
server, workstation, designers, bank institute etc is what currently forms the "world" mobile is it's own closed in beast of course
Valantar - Tuesday, April 9, 2019 - link
Uhm, there are non-pro 3000-series mobile APUs, you know...ajp_anton - Wednesday, April 10, 2019 - link
Eh, what?These 3000U Pro parts use the same silicon as the 3000U non-pro parts. That's where the "discards" are going.
drexnx - Tuesday, April 9, 2019 - link
even RR was a big improvement over SR in terms of platform features actually working.I never see my SR 1700 turbo correctly and my RR 2200G behaves as it should
deil - Tuesday, April 9, 2019 - link
they mentioned that in RR changelog, turbo-magic.about rebrand -> 2400G have 4MB cache and this have 6.
Stuff changed, so its not just a rebrand.
deil - Tuesday, April 9, 2019 - link
I take that back, it was 2+4, i did not notice.still 65W into laptop with 15W without boost clock change? I don't think so. Would go full throttle in first 10 sec....
Samus - Wednesday, April 10, 2019 - link
65w is on-par with many Intel solutions using a 15w CPU and a discrete 50w GPU.Intel's Cannon Lake has such a configuration.
What's important is AMD's solution has superior CPU and GPU performance at the same power envelope, but that has yet to be established so we will have to see...
albert89 - Saturday, April 27, 2019 - link
People are upset that AMD is doing great. Sothey call AMD's new release at a lower lithography
a rebrand. While apparently Intel's 5th or 6th
generation on 14nm are not rebrands according
to the same Intel fangirls. Give me a break.
Spunjji - Tuesday, April 9, 2019 - link
These are no more a re-brand than the change from Ryzen 1000 to Ryzen 2000 series was - it's the same changes (process and architecture tweaks) trickling down to their mobile parts. Based on early evidence from other sites, it looks like they may have substantially reduced their idle power problems, too.gijames1225 - Tuesday, April 9, 2019 - link
If they tidied up the idle power then these will be great. The first generation where solid in performance, but the battery drain kept them out of the high-end.neblogai - Tuesday, April 9, 2019 - link
But MatebookD with 2500U has great battery life (in addition to great performance). So, the fault was not on APUs, but on laptop manufacturers and their lazy designs.Irata - Wednesday, April 10, 2019 - link
Have the same and concur, however I usually do not let it idle on battery for hours. Since it boot up really fast, there is no need for that.But yes, battery life is great - can use it over several days (not the whole day obviously) without the need to recharge which is a fist for me with AMD based laptops.
bug77 - Tuesday, April 9, 2019 - link
I was thinking the same thing: they clearly had a power issue before and they don't talk about power draw at all. Which can only mean they didn't fix it.And tbh, I really, really, really don't care about Vega being faster. For workloads that don't need a dGPU, Intel has enough HP already. Sure, I'll take better encoding/decoding capabilities, better outputs and whatnot. But all this "oh, look, Vega is so much faster" talk, does nothing for me.
TheinsanegamerN - Thursday, April 11, 2019 - link
Sounds like youa re not the market for Ryzen then.I really, really, really care about Vega being faster. For workloads that use GPUs in more then a passing fashion, a more capable iGPU is always welcome. Some of us dont want to have to shell out $700 for a bottom tier "gaming" laptop just to have more power then intel was putting in their chips in 2013.
Irata - Friday, April 12, 2019 - link
With this regard, the Matebook with the 2500U is really great - my kid can play Fornite @ Full HD resolution with Medium details.The sad thing is it runs smoother than on a "four" core Trinity based Desktop with a 4 GB RX550, SSD and 16 GB RAM...I think it's really about time I upgrade my Desktop PC, as well.
Intel999 - Tuesday, April 9, 2019 - link
AMD announced months ago that they would be expanding the powergating properties on Picasso.In the first gen of mobile processors the iGPU never completely powered down when not in use. They have addressed that with Picasso along with other less efficient properties of the first generation.
BigMamaInHouse - Tuesday, April 9, 2019 - link
Cinebench R15: Core i7-8565U vs. Ryzen™ 7 3700U: 619 vs. 688!IGTrading - Wednesday, April 10, 2019 - link
Wonderful piece of technology.I love that the day has come when you can truly recommend AMD over Intel for an overwhelming number of scenarios and applications.
No reason to buy Intel anymore, unless there is some specific code where Intel's arch behaves better and you care not about all the other benefits AMD brings.
eastcoast_pete - Tuesday, April 9, 2019 - link
One aspect that might hurt AMD here is that any prospective buyer of Ryzen Pro systems knows that better-performing 7 nm Ryzen chips are coming soon. So, in a perverse way, Intel's laziness or incompetence in moving below 14 nm works in their favor, especially with business customers: there is little risk that they will make substantial improvements in the next 12 months, so you might as well buy now.SaberKOG91 - Tuesday, April 9, 2019 - link
It all depends on when you need it and what you are buying. Most companies refresh at least a fraction of their machines every year, with very few companies replacing all machine every couple of years. As the 2XXX series systems get retired from the channel, they will get replaced by 3XXX and that will be all you can buy until next spring when 7nm Ryzen Mobile is available. There's no reason to wait on for 7nm if you need to buy laptops this year. Desktops and workstations are another story though.neblogai - Tuesday, April 9, 2019 - link
That depends on AMD's ambition. If they desired, they could bring a high-performance 7nm 8core CPU to mobiles this year, by creating a special, low power I/O die (with basic graphics/VPU as well), and coupling it with 7nm Zen2 chiplet. That chiplet would probably not be able to powergate as well as a chip designed specifically for mobile, but at 7nm, and with low power I/O, that should still be enough for high performance laptops to have decent idle battery life.SaberKOG91 - Tuesday, April 9, 2019 - link
Very unlikely. They will be maximizing 7nm throughput for server and desktop parts that bring in all the money. Even if they had production capacity, it still wouldn't make sense. We are way too early in the design cycle for the 4XXX series parts to be even a consideration. Expect Q4 announcements with full launch early next year. With the delay going on with desktop 3XXX and Navi, there's not even remotely a chance of 7nm mobile this year. Also you are forgetting that mobile packages need to be small for wide adoption. It makes no economical sense to add an I/O die for a single CPU die, even with graphics in the mix.neblogai - Tuesday, April 9, 2019 - link
Sorry- I don't get any of your arguments.1) AMD has almost no presence in high performance/gaming laptops. It takes time to build up presence there. It would take negligible number of 7nm capacity for AMD to start that entrance.
2) Naming is the least of the issues for a company- only users make big deal out of it. These mobile chips could be mobile 3xxx H-series R7 or R9 without any problem.
3) There is no date set for 7nm monolithic APUs, only several year old slide indicates it should come in late 2019. And it also indicates that it would be power efficient 10-35W 4-core only designs, not high performance parts.
4) There is no delay with 3xxx or Navi. 3xxx was promissed to come out mid-2019- and we will be getting 3xxx and Navi announcements 27th of May.
5) AMD needs a mobile I/O die if they 8-core for want mobile and embedded markets. Packaging 2 chips together is certainly not an issue (Intel packages 3 chips for Kaby Lake-G). Investment from AMD would not be big, because most of the mobile I/O design elements could be reused from RR and from Matisse I/O. It's all on 14nm. So, it only depends on AMD's ambition, if they try to expand into high-performance mobile and embedded, or not.
SaberKOG91 - Tuesday, April 9, 2019 - link
1) You seriously underestimate the cost of small runs, especially on a new process node.2) I was referring to naming because that's likely what they will be called next year, not because AMD couldn't release a faster chip under the current branding.
3) I am estimating based on what we have seen in the past for rolling out to new nodes. AMD has almost always launched the mobile parts on a new node and microarchitecture after their desktop counterparts.
4) A few different sources have stated that there was an issue in Fall of 2018 that held back the 3XXX parts from being released earlier in 2019 to mirror the launch of the 2XXX parts. It wasn't until after that that AMD came out and officially started saying later in 2019. One would assume Navi may have been impacted as well.
5) False. They would simply go with a larger monolithic package. The cost of integrating multiple chiplets onto a single package for what would be a < 300mm^2 die makes no sense. Interposers are expensive and add height to the package. Spacing of the chiplets makes the package larger in area. Both of which are tradeoffs that make no sense for mobile designs. Intel uses EMIB which allows them to do this without an interposer which means shorter packages. AMD have no publicly announced corresponding IP. I forget where Papermaster answered a question about this, but AMD have already indicated that they would not be using chipsets for this generation of mobile and desktop APUs.
SaberKOG91 - Tuesday, April 9, 2019 - link
Edit: chiplets, not chipsets*neblogai - Wednesday, April 10, 2019 - link
1) Old 14nm process for I/O die, while half the silicon is aleady done in Zen2 chiplet. So releasing it would be cheap, and proper use of chiplet architecture to expand addressable market.2) Not doing in now is a waste of opportunity.
3) Does not have to be later this time, because Zen2 chiplet is done, while I/O chiplet does not need any new tech, only a copy of what they have already done.
4) Rumours, and no leaks of actual schedule.
5) No interposer or EMIB needed, and no need for new Zen2 chiplet- only new 14nm I/O die, and mounting the two dies together on FP5. AMD said Matisse (AM4 package) will not have GPU die chiplets, that is all. And it makes sense, because Matisse is for desktop, and that would be unnecessary and expensive- same as making a monolithic powerful 7nm mobile APU, while Zen2 chiplet is already there waiting.
SaberKOG91 - Wednesday, April 10, 2019 - link
1) Old Process != cheap. Zen2 chiplets do not have the same power management features as the mobile parts. Mobile Ryzen doesn't have enough PCI-E lanes for an IF link to an external I/O die.2) Not really. High-end parts are low volume and would not chip in high enough quantities to offset the tooling and marketing costs, for AMD or OEMs.
3) Again, Zen2 chiplets of today do not have the power-saving tech found in the mobile parts.
4) Fine. Time will tell.
5) An interposer or other integration tech is 100% required to connect the packages together. FP5 is just a normal BGA package where the traces connect the die contacts to the BGA pads. You would still need another layer to connect the packages together or you would have to route through to the motherboard and back which adds latency, power consumption, and design cost. No one would do this. Even if you were stupid enough to do so, the GPU/IO die you have proposed on 14nm would barely fit on the package and only be as fast as last gen. You underestimate how much of the APU die is GPU (hint: over half). The reason that matisse won't have a GPU chiplet version is far simpler: limited memory channels. The size of GPU that you could fit on matisse would require 2 additional memory channels just to keep the GPU fed. The current APUs are already bottle-necked by having to share two channel with a GPU on a 4C/8T system. This is why increasing memory clockspeed (bandwidth) has so much uplift on gaming benchmarks. Trying to do that with 8C/16T and a GPU that's at least as powerful as a 560 and the same number of memory channels is not remotely feasible. I can guarantee you that when the 16C/32T matisse parts get reviewed we will see two memory channels become a bottleneck and that is why AMD made the L3 cache so much larger (to reduce the impact).
neblogai - Thursday, April 11, 2019 - link
1) Old process, and a lot of reused blocks (Vega, not Navi; same display, multimedia engines, etc) = way, way cheaper and could be done for this summer, compared to trying to build everything new on 7nm for next year.2) Still needed if AMD aims to enter the high performance mobile/embedded market. Entering with an expensive design is even worse, because sales will not jump up anyway, and will only ramp up slowly, over many years.
3) With an energy efficient I/O and iGPU, Zen2 7nm chiplet with behavior tuned for efficiency would be good enough for long battery life, radically better than current desktop chips + dGPUs in certain Asus/Acer laptops.
5) Interpozer is not needed, same like it is not needed for Epyc/Threadripper, or Matisse. So, it is cheap. And you are thinking wrong about the GPU for it. For such mobile chip, mobile iGPU would be needed for extreme energy efficiency at idle, multimedia playback and video outputs, not for rendering games. So, 1-3CU in total, and no iGPU memory bandwidth issues. For Embedded and HPE- that is enough, and for gaming machines- proper dGPU would be added and switched on while gaming.
SaberKOG91 - Thursday, April 11, 2019 - link
1) You are neglecting tooling costs. Small runs on a 1000+ step process are not cheap, especially for something they won't sell more that a few tens of thousands of.2) They don't want to or they would have by now. What you have proposed IS an expensive design that no OEM would buy. Even if they could, it takes time to roll something like that out which AMD is no doubt planning on doing this fall with OEMS for next spring. It's too late in the year for something like that and we've had no leaks whatsoever.
3) Not really. Even if you can dramatically reduce dynamic power consumption by running at lower TDPs and idle clocks, you still have to properly manage static power dissipation. Desktop parts typically have an order of magnitude higher static dissipation than their mobile counterparts. This is precisely why Ryzen Mobile 3XXX is not using the same die as the desktop version. That fact doesn't change for your imagined 7nm chiplet. When TSMC is talking about 0.5X power consumption they are talking about dynamic, not static.
5) Epyc, Threadripper, and Matisse all use interposers. Not sure where you are getting your information from, but it's flawed. Infinity Fabric links are routed on the interposer between the chiplets directly in Zen1 and now to the I/O die in Zen2, not through the package substrate or the motherboard. Only the socket-to-socket IF links actually reach the LGA pads and motherboard.
1-3 CU? I thought we were talking about a supposed high-end mobile part -_- . Now you're talking about embedded and HPE. Even Intel's low-end mobile parts have more power that that. If AMD want a high performance mobile part, they would just shut off the built-in GPU, raise the TDP to 30W+ and then bundle a dGPU in the 550 or so range. No switchable graphics needed as Polaris can get very low idle power when undervolted and underclocked. Or they could go back to the Bulldozer era and pair an equally performing dGPU to match the iGPU and have you use crossfire. The GPU on ryzen mobile is deliberately undersized for the CPU because of memory bandwidth. You're grasping at straws now mate.
neblogai - Thursday, April 11, 2019 - link
Yea, I also feel like I'm wasting time..SaberKOG91 - Thursday, April 11, 2019 - link
Well, you were never going to get me to agree with you :p You keep arguing about ambitious and lofty goals that AMD doesn't have the kind of spare cash to justify chasing. If this were Intel we were talking about, none of this would be unreasonable. Instead it's a company that makes an order of magnitude less revenue each quarter. Don't get me wrong, I like AMD, but this just isn't how they do business.neblogai - Thursday, April 11, 2019 - link
Well, they could be making more cash by reusing their tech more widely. They actually seem to try to do that with scaling Zen and Zen2 dies for several markets at once. But yes- this would be one more project, and maybe AMD do not have the capacity to pursue everything. Simply, this is my take on what would be very reasonable for them to do, in contrast to AdoreTV rumors of Zen2 desktop APU with 14-20CUs- which do not make sense.TheinsanegamerN - Thursday, April 11, 2019 - link
So is AMD going to support these ones properly, or are they going to let them wither on the vine and require OEMs to update driver for AMD?That is what killed first gen mobile ryzen, along with poor battery life and power usage in general. A tale that is as old as time for AMD.
Irata - Friday, April 12, 2019 - link
You can now run the regular Adrenalin drivers, battery life is great if the laptop is properly configured and built and power usage is also fine. Get 9+ hours on my 2500U based laptop.The reported issue was idle only, so if you get a laptop to let it idle for hours on battery then maybe mobile Ryzen is not for you
TheinsanegamerN - Friday, April 12, 2019 - link
Sure, its fine NOW, over a year after release.Is AMD going to do this again? Because if they do, then mobile 2nd gen will be just as much of a sales failure as the 1st gen. Some of us dont want to wait a year before our products work like they are supposed to.
Soulkeeper - Tuesday, April 16, 2019 - link
I wish the ecc memory I purchased worked with my 2700UIt is a busines "pro" scam afterall ...