So here it goes: possible beginning of the end for x86

[ViRGE]

128-bit address registers would be my definition. And yes, overkill for the next 10+ years at least.

That would have to be the definition. Otherwise if the requirement is just 128bit math, Intel already has 256bit processors thanks to Sandy Bridge and AVX.

 

[drizek]

That was worth a ROFL or two

Thing is, I remember reading, and believing, exactly that back in around 04. I almost bought an AMD64 laptop because of it too. And an eMachines laptop, no less.

Of course, not nearly as bad as the GIF animations that Apple used to prove how their 450MHz G4 Velocity Engine was actually faster than a 2GHz Pentium 4.

http://images.apple.com/lae/powermaccube/images/processorphoto7081699.gif

The Velocity Engine
The secret of the PowerPC G4’s phenomenal performance is its aptly named Velocity Engine. It’s the heart of a supercomputer miniaturized onto a sliver of silicon. The Velocity Engine can process data in 128-bit chunks, instead of the smaller 32-bit or 64-bit chunks used in traditional processors (it’s the 128-bit vector processing technology used in scientific supercomputers — except that we’ve added 162 new instructions to speed up computations).

In addition, the PowerPC G4 can perform four (in some cases eight) 32-bit floating-point calculations in a single cycle — two to four times faster than traditional processors.

The PowerPC G4 with Velocity Engine works with the PowerPC architecture to accelerate the data-intensive processing required by next-generation video, voice and graphics applications. Among the G4 key features is a vector permute function capable of rearranging data in the registers — a priceless benefit when converting data from one format to another (often necessary with voice, video and graphics apps, which typically need to save data in a number of different formats).

These vector processing acceleration advantages give the G4 microprocessor a significant edge when it comes to fast visualization of large data sets, and intensive math for real-time simulations. Making the PowerPC G4 perfect for everything from digital video, graphics and 3D games to astronomy, the biosciences and predictive modeling.

http://www.apple.com/lae/powermaccube/processor.html

 

[Idontcare]

^ ah yes, back when the Steve Job's reality distortion field was really doing its magic on making the aging G4 architecture look sexy.

Gotta wonder just how different AMD could have turned out if Jerry Sanders had Steve Job's charisma and vision.

 

[drizek]

^ ah yes, back when the Steve Job's reality distortion field was really doing its magic on making the aging G4 architecture look sexy.

Gotta wonder just how different AMD could have turned out if Jerry Sanders had Steve Job's charisma and vision.

Say what you want about Jobs, but he and his company are not stupid. They knew that Intel had hit a home run with Banias/Dothan and they jumped on it as soon as they could. AMD had no comparable mobile chip. I'm not even sure they do even now, at least not until Llano comes out.

This was around the time where Intel was basically a joke on the desktop. They were just dumping whatever they could find on the Pentium 4 to try and keep it afloat (HT, EE, L3, PD). AMD got cocky but Steve Jobs knew the future was in mobile, and Intel was ahead by a mile. I think AMD just lacked vision and they built the wrong chip.

 

[dma0991]

Looks like AMD is not getting ARMed

Advanced Micro Devices shot down rumors that it is pursuing an ARM license, saying it will stick to developing chips for tablets around the x86 architecture.

http://www.pcworld.com/article/226682/amd_not_pursuing_arm_license_sticking_to_x86.html

 

[VirtualLarry]

"Chip companies making x86 processors -- Intel and AMD -- have minimal presence in the tablet market. Tablets with Intel's first dedicated tablet chip, code-named Oak Trail, are expected to become available next month. AMD next year will release tablet processors code-named Wichita and Krishna, which will include up to four CPU cores. Just like AMD's recent Fusion laptop and desktop chips, the processors integrate the CPU and graphics processor inside one chip."

How come I haven't heard anything about Wichita and Krishna on this forum before? Four cores? Low power (tablet range)? Is this a low-power LLano, or something else entirely?

 

[busydude]

Is this a low-power LLano, or something else entirely?

They are next gen APU's based on bobcat cores.

 

[Arkadrel]

"Chip companies making x86 processors -- Intel and AMD -- have minimal presence in the tablet market. Tablets with Intel's first dedicated tablet chip, code-named Oak Trail, are expected to become available next month. AMD next year will release tablet processors code-named Wichita and Krishna, which will include up to four CPU cores. Just like AMD's recent Fusion laptop and desktop chips, the processors integrate the CPU and graphics processor inside one chip."

How come I haven't heard anything about Wichita and Krishna on this forum before? Four cores? Low power (tablet range)? Is this a low-power LLano, or something else entirely?

The "Bobcat" replacement.

So 9watts - 18watts (on 40nm), will probably be same or less on 28nm.

The C-50 has 1.0ghz Dual core + 80 shaders GPU@280mhz ~9watt TPD (on 40nm)
The E-350 has 1.6ghz Dual core + 80 shader GPU@400mhz ~18watt TPD (on 40nm).

AMD where supposedly able to Take away a few features, but keep the C-50 CPU/GPU the same, and hit around 5watt TPD on 40nm (for tablet market).


So a Wichita will probably be like: (Just guessing)

E xxx with ~2.2ghz quad core /w Turbo + 160 Shaders GPU@500mhz ~18watt TPD.
(so GPU kinda like a Radeon 6450)

And go into some small netbook with 5-8hours of battery life.

 

[smartpatrol]

Jon Stokes at Ars Technica wrote a great article about this not too long ago. Check it out here:
http://arstechnica.com/business/news/2011/02/nvidia-30-and-the-riscification-of-x86.ars

To sum it up:

- ARM processors will not outperform x86 in the near future, but that probably won't matter
- x86 beat Alpha, PowerPC, and the other RISC architectures not because it was faster. x86 won because it was far cheaper commodity hardware, and it was good enough performance-wise.
- ARM may do the exact same thing to x86. If ARM chips can deliver a decent fraction of x86 performance, and do it cheaper, then it will displace x86.

There's a lot more to it than that. Everybody should read that article.

 

[GLeeM]

x86 has at least 10 years to live, possibly more

x86 won because it was far cheaper commodity hardware, and it was good enough performance-wise

I think x86 will not leave soon.
Things move along much slower for commodity hardware than ten years.

When I started to need a computer at work in '98, it was a PII 266. That got replaced in '04 by a P4 2.8 with integrated graphics which I am still using. Last year we got an HD video camera that I thought I could put to a secondary use by making training videos. When I put in for a computer and software that could handle HD video all I got was the software - it was not time yet for the computer to be replaced! Needless to say, I haven't made many training videos.

 

[Cerb]

the other thing is that with lightweight operating systems that don't have all the baggage windows carries around you don't need as much processor power to have the same end-user experience. windows on ARM sounds like a disaster to me. but who knows? iOS is a stripped down version of OSX so maybe windows can do the same.

The issues are really interfaces and battery management. OS X needed to be stripped down for two main reasons: RAM, and cache. ARM CPUs are now getting enough of both, better RAM controllers, etc.. By the time Windows 8 comes around, the A9 will be old. And, if we can take tests in Linux as a base, A9 SoCs seem generally comparable to Atoms, which are quite capable of running full blown Windows.

that was back when nobody used windows NT. That's what the N stood for. Nobody's Technology.

Did you know that most games, even claiming to require Win9x and DirectX 7, ran fine in NT? Some of us skipped the 9x OSes. Most of what we missed was buggy USB support.

Also remember that just because Microsoft is putting Windows on ARM doesn't mean it'll be worth using that way.

Real Windows on ARM would make it easier to have x86 and ARM software support, and also to have drivers for both hardware platforms. Even if MS doesn't screw up the UI, it will take time for it to flourish, if it even can, but with mobile being an entirely different OS, with vast version-to-version incompatibilities, there's no chance of it.

 

[podspi]

Real Windows on ARM would make it easier to have x86 and ARM software support, and also to have drivers for both hardware platforms. Even if MS doesn't screw up the UI, it will take time for it to flourish, if it even can, but with mobile being an entirely different OS, with vast version-to-version incompatibilities, there's no chance of it.

It is going to be interesting seeing how Microsoft handles having two version of Windows aimed at consumers. I wonder if they'll go with some sort of fat binary ala OSX.

 

[gdansk]

I wonder if they'll go with some sort of fat binary ala OSX.

Would be a very good idea, merge the x86 and x64 binaries in there too. Then later the ARM-64...