Nuvia 'clean-sheet CPU design' performance previewed

[HEXUS]

Compares the Nuvia Phoenix CPU against the best AMD, Apple and Intel have to offer.

Read more.

[edmundhonda]

Endless R&D dollars from ARM and Apple have produced that outstanding perf/watt at the lower levels, but this company has significantly outpaced both of them within 12 months of starting development?

I'm sure they're not just another massively over-valued tech company in SV. Definitely not.

[CAT-THE-FIFTH]

So another ARM based CPU?? Thought it might have been RISC-V based or a brand new instruction set. Also until they have production silicon,I don't believe their claims!

[kompukare]

So another ARM based CPU?? Thought it might have been RISC-V based or a brand new instruction set. Also until they have production silicon,I don't believe their claims!

Well, all the leading people there are ex Apple and have plenty of ARM experience.
They're supposed to have an ARM v9 architectural licence so wouldn't be worried about Nvidia.
Obviously take everything with a grain of salt, but if they can assemble a team like they had at Apple they might be able to pull it off.
Apple's team were able to increase their performance around 20% per year after all.
Nuvia don't have to worry about mA power budgets, lack member controllers and IO.
On the other hand, ARM themselves now show signs of making server reference designs.

None of which bodes well for the socketed DIY market, though. Unfortunately, the upgradable DIY PC market is a historical accident and throwaway stuff with built-in obsolescence is probably, sadly the future no matter how much $ trillion corporations like to pretend they care about the environment.

[CAT-THE-FIFTH]

Well, all the leading people there are ex Apple and have plenty of ARM experience.
They're supposed to have an ARM v9 architectural licence so wouldn't be worried about Nvidia.
Obviously take everything with a grain of salt, but if they can assemble a team like they had at Apple they might be able to pull it off.
Apple's team were able to increase their performance around 20% per year after all.
Nuvia don't have to worry about mA power budgets, lack member controllers and IO.
On the other hand, ARM themselves now show signs of making server reference designs.

None of which bodes well for the socketed DIY market, though. Unfortunately, the upgradable DIY PC market is a historical accident and throwaway stuff with built-in obsolescence is probably, sadly the future no matter how much $ trillion corporations like to pretend they care about the environment.

Personally I don't think this will get much traction,unless it is revolutionary in some way. ARM and MIPS based designs already exist,so companies can already buy those. Also many countries are now implementing national CPU designs around RISC V,for eventually use in import/export substitution scenarios, but these are all government backed intiatives. Not sure what market this is trying to target.

[watercooled]

If they're wanting to actually make some sales and gain a foothold in the market, ARM makes the most sense, with perhaps POWER as another option for specific markets. It's more than an uphill battle to release a CPU no-one can use because they have no software able to run on it, no well-optimised compilers, very few people able to program for it efficiently, etc. ARM itself has faced an enormous challenge to break into the server market despite being AFAIK by far the most widely used ISA on the planet.

Having said that, yeah it's easy to post charts, lets wait and see what products are actually like.

[Xlucine]

That chart is crying out for a log scale on the X axis

[yeeeeman]

I call this BS until I actually see a product which is benchmarked by phoronix or some other independent reviewer.

[yeeeeman]

So another ARM based CPU?? Thought it might have been RISC-V based or a brand new instruction set. Also until they have production silicon,I don't believe their claims!

Well, all the leading people there are ex Apple and have plenty of ARM experience.
They're supposed to have an ARM v9 architectural licence so wouldn't be worried about Nvidia.
Obviously take everything with a grain of salt, but if they can assemble a team like they had at Apple they might be able to pull it off.
Apple's team were able to increase their performance around 20% per year after all.
Nuvia don't have to worry about mA power budgets, lack member controllers and IO.
On the other hand, ARM themselves now show signs of making server reference designs.

None of which bodes well for the socketed DIY market, though. Unfortunately, the upgradable DIY PC market is a historical accident and throwaway stuff with built-in obsolescence is probably, sadly the future no matter how much $ trillion corporations like to pretend they care about the environment.

Why don't you all commenting about how Apple and ARM managed to increase the IPC by 20% yearly don't put things into context?
Apple has basically added all the tricks in the book that X86 CPUs already use for 10 years or more. The biggest achievement of Apple is that they have similar, but not quite the same performance as Skylake core in 2015 while consuming 2-3 times less power per core. But that is achieved through various ways: one is process. Intel cores are made on 14nm, Apple cores are made on TSMC 7nm+. With that you get at least 2x perf/watt. Then, Apple doesn't have legacy stuff in the core like Intel has with X86. It doesn't have CISC decoding logic. It doesn't use MOSFETS to power the CPU, they use PMIC which is FAR more efficient. Frankly, desktop CPUs are wasting a big part of their power budget on power stages.
I think people really don't understand these facts. Apple SoCs are made for low power. They use every single trick in the book to reach as low power as possible. Desktop x86 and even laptop x86 cpus are made for performance and they lose a lot of efficiency on what I've just told.
Just look at Lakefield which is a first try to get back in competition in mobile. Being made on 10nm and 22nm, it has a big handicap, but still manages to outperform Snapdragon 8cx which is state of the art 7nm SoC from Qualcomm. So I wouldn't jump to conclusions that fast in regards to ARM/X86 like I see pretty much everyone is doing these days.

[kalniel]

Having said that, yeah it's easy to post charts, lets wait and see what products are actually like.

Well given how amatuerish the blue splodge looks on the chart, perhaps it's harder than we think..

[syristix]

So another ARM based CPU?? Thought it might have been RISC-V based or a brand new instruction set. Also until they have production silicon,I don't believe their claims!

Well, all the leading people there are ex Apple and have plenty of ARM experience.
They're supposed to have an ARM v9 architectural licence so wouldn't be worried about Nvidia.
Obviously take everything with a grain of salt, but if they can assemble a team like they had at Apple they might be able to pull it off.
Apple's team were able to increase their performance around 20% per year after all.
Nuvia don't have to worry about mA power budgets, lack member controllers and IO.
On the other hand, ARM themselves now show signs of making server reference designs.

None of which bodes well for the socketed DIY market, though. Unfortunately, the upgradable DIY PC market is a historical accident and throwaway stuff with built-in obsolescence is probably, sadly the future no matter how much $ trillion corporations like to pretend they care about the environment.

Why don't you all commenting about how Apple and ARM managed to increase the IPC by 20% yearly don't put things into context?
Apple has basically added all the tricks in the book that X86 CPUs already use for 10 years or more. The biggest achievement of Apple is that they have similar, but not quite the same performance as Skylake core in 2015 while consuming 2-3 times less power per core. But that is achieved through various ways: one is process. Intel cores are made on 14nm, Apple cores are made on TSMC 7nm+. With that you get at least 2x perf/watt. Then, Apple doesn't have legacy stuff in the core like Intel has with X86. It doesn't have CISC decoding logic. It doesn't use MOSFETS to power the CPU, they use PMIC which is FAR more efficient. Frankly, desktop CPUs are wasting a big part of their power budget on power stages.
I think people really don't understand these facts. Apple SoCs are made for low power. They use every single trick in the book to reach as low power as possible. Desktop x86 and even laptop x86 cpus are made for performance and they lose a lot of efficiency on what I've just told.
Just look at Lakefield which is a first try to get back in competition in mobile. Being made on 10nm and 22nm, it has a big handicap, but still manages to outperform Snapdragon 8cx which is state of the art 7nm SoC from Qualcomm. So I wouldn't jump to conclusions that fast in regards to ARM/X86 like I see pretty much everyone is doing these days.

Lakefield does not outperform the 8cx in all areas? Some single thread but multi threaded the 8cx wins in most scenarios. Arm is getting the headlines as x86 is under attack in so many areas by Arm based designs and that's where the industry momentum is.
Apple have gone over to ARM, graviton 2 (AWS) and Ampere designs in the datacentre space are giving x86 a torrid time and now even x86 in desktop is being challenged with the likes of 8cx.
When Intel stopped innovating and MS failed to succeed in the mobile space the writing was on the wall. In my mind it's not if but when the transition to ARM occurs.
The longevity of x86 in mobility for everyday computing hinges on how long it will take apps and software to be either ported to ARM 64 or natively written for it and better price parity with x86.

[CAT-THE-FIFTH]

ARM will eventually hit the same problems as X86 CPUs too,ie,backwards compatability will have to be a thing. So they will eventually start to get more bloated. In fact this is one of the reasons why RISC-V was started,ie,to have less bloat.

[kompukare]

Why don't you all commenting about how Apple and ARM managed to increase the IPC by 20% yearly don't put things into context?
Apple has basically added all the tricks in the book that X86 CPUs already use for 10 years or more. The biggest achievement of Apple is that they have similar, but not quite the same performance as Skylake core in 2015 while consuming 2-3 times less power per core. But that is achieved through various ways: one is process. Intel cores are made on 14nm, Apple cores are made on TSMC 7nm+. With that you get at least 2x perf/watt. Then, Apple doesn't have legacy stuff in the core like Intel has with X86. It doesn't have CISC decoding logic. It doesn't use MOSFETS to power the CPU, they use PMIC which is FAR more efficient. Frankly, desktop CPUs are wasting a big part of their power budget on power stages.
I think people really don't understand these facts. Apple SoCs are made for low power. They use every single trick in the book to reach as low power as possible. Desktop x86 and even laptop x86 cpus are made for performance and they lose a lot of efficiency on what I've just told.
Just look at Lakefield which is a first try to get back in competition in mobile. Being made on 10nm and 22nm, it has a big handicap, but still manages to outperform Snapdragon 8cx which is state of the art 7nm SoC from Qualcomm. So I wouldn't jump to conclusions that fast in regards to ARM/X86 like I see pretty much everyone is doing these days.

While all that is true, let's remember a few other things too:

1. Intel lost in mobile because they wanted to ignore high-volume, low margins and didn't take the market seriously. (Also don't really believe Nvidia with their we couldn't be bother to bid for consoles because of the margins spiel.
2. Intel's Atom and constant chipset changes were meant to keep old fabs busy when core had moved the latest process.
3. Intel loosing in mobile enabled TSMC to get where they are.
4. The volume and profits of mobile enabled lots of companies to gain experience at making chips, and some now want to use this experience to enter the server market.
5. Intel seemed to have forgotten that volume is what enabled x86 to eventually destroyed all the workstation vendors, if ARM's volume now does the same for Intel's server cashcow it will be very ironic.
6. x86 only happened because IBM were stupid enough to choose Intel in the first place, if they had chosen Motorola 68K there never would have been any 64KB segments, full 32 bit registers from the beginning etc. Not that 68K wasn't very much CISC, but the overhead of x86's complexity is pretty much Intel's (and AMD's) problem and the far cleaner design of ARM is of course a major advantage.
7. Ergo Intel's poor x86 design held back computing for first decade of the PCs life. The 486 was probably Intel's first decent design.
8. Intel's x86-everywhere is crazy but on the other hand everything they touch not in their main x86 market they seem to mess up. And when you are as big as Intel, those mess-ups can be costly: Itanium, Mcafee, Larabee, $billions in Atom dumping (no, sorry call it contra revenue to avoid the gaze of regulators, and so on.

So yes, of course ARM's performance rise means using some of the tricks the bigger CPU cores have been using for years, but then the reverse is also true. A lot of the power saving features were first pioneered by ARM.
And like I said, while I would be happy to get rid of x86 I am very much attached to the DIY market and don't want a closed-system Chromebook or Apple ARM computer.

[Strawb77]

Having said that, yeah it's easy to post charts, lets wait and see what products are actually like.

Well given how amatuerish the blue splodge looks on the chart, perhaps it's harder than we think..

that`s there for the hard of thinking