China's 64bit Homegrown CPU

An anonymous reader writes: "EE Times is reporting on China's BLX IC Design Corp nearing the completion of their first 64-bit CPU. Based on the MIPS instruction set the 500-MHz Godson-2 microprocessor is aimed toward distributed grid computing. To avoid MIPS patent issues, several instructions (unaligned loads and storeds in the 32 bit version) have not been implemented but with the support of over 60 software providers such as Red Flag Linux and the ability to tweak compilers to not use these instructions this should not be a problem. The Godson-1 processor (also patent free) was announced last year and was aimed at the embedded market." The Godson processor line has generally been called Dragon by the Western press.

Comments

Comment removed

[account_deleted]

Comment removed based on user account deletion

Re:Welcome to the future...

[Anonymous Coward]

To all you genius programmers: you're good. But are you good enough to outhack half a dozen Chinese guys working for half your salary?

I predict that within 10 years, half the US programming market will have gone to these overseas firms.


Been there, done that.


Ed Yourdon's "Decline and Fall of the American Programmer"

and the sequel

Rise & Resurrection of the American Programmer

Re:Welcome to the future...

[Anonymous Coward]

Right here.

Re:Nice architecture

[Bender_]

Couldn't do any better than to choose the MIPS instruction set. I looked at it years ago and was impressed with its clean design

Thats no wonder - it was refined during years of research by Henessey and Patterson.

However if you look close you will notice that the instruction set does also contain some obselete legacy. For example branch delay slots do not make any sense with OOO Architectures. It is also questionable whether wasting quite a bit of instruction space for integer arithmetic both with and without overflow trapping is worth it. Maybe the could just have used the extra space for a proper move instruction so R0 is freed.

MIPS pantent issue

[brejc8]

I did post the story but the last sentence which was cut was very important.(Original).
"Although there are no patent issues MIPS have been known to be very aggressive toward people who try to create compatible systems."

Reason for MIPS r0

[yerricde]

Maybe the could just have used the extra space for a proper move instruction so R0 is freed.

The MIPS architecture already has a proper 'move' instruction without using r0: r12 = r8 | r8, or r12 = r8 | 0 (zero specified as immediate). The r0 is frozen at 0 so you can do negations (for which ARM uses 'rsb' or reverse subtraction) and other things where zero must be the first argument.

Re:China's Chip

[Anonymous Coward]

Actually, MIPS asm _is_ an open standard and they were careful to not use the few instructions that were covered by patents.

Re:DRM?

[zmooc]

DRM functionality on a CPU does not cripple anything. It's the encryption of the media that may cripple the functionality you may have with regard to data from others. And it's the software/OS that has the option to use it.

DRM only adds functionality like controlling what recipients are to allow to do with emails - just disable any functionality to forward emails which contain confidential data. Don't want others to use your picture for other purposes than viewing it on your website? Possible. Lost your Palm with those rather private pictures on it? No problem. And ofcourse digital media will no longer be copyable directly... but digital media will become a lot cheaper sometime in the future - the price is mainly due to the expensive technology used to create them; expensive studios, 3D-software, special-fx-software, videocamera's etc. are expensive but get cheaper and cheaper. This will not only drive the price of the media down (which will definately raise the volume) but bring a lot more on the market since it'll become a lot cheaper to make things for everyone. Especially with bandwith getting cheaper.

Now the things that you DO have to fear:

• DRM incompatibilities between different systems - you may need a lot of different plugins in that case... this may happen if e.g. Real first starts adding DRM to their realmedia, MS then comes up with their own passport-based shit and then finally some standards committee comes up with an open standard which is way too late
• Closed standards - if Real of MS or whoever comes up with a closed standard which will only be available by using their software, us Linux users will be fucked. This may well happen since most average windows-using internetuser won't hesitate to install all this software and therefore market-penetraion won't be a problem as long as the software is free.
• Patents on DRM-systems - Open Source would be locked out then. At least in the USA.
• DRM becoming a requirement before about everybody has the hardware. And then still your old PC won't be able to open DRM-protected media since it a secure data-path has to be built into just about everything from the memory and the CPU all the way to the last peripherals.
• DRM forcing no-fast-forward on you so you have to watch all the commercials.
• Data-recovery not being though about - losing data due to a lost key or something would be bad. Something to solve this problem should be implemented. With regard to history in the future this will also be really important; without it the 21st century will be a very dark age in history!
• The government or some large company having master-keys.
• Expensive audits required to check for leaks driving the price of hardware (which will get a lot more complex anyway) up.
• The first DRM-hardware like speakers and LCD-monitors not using wireless transmission by default:) For a really safe data-path, the DRM-decription hardware will have to be in your speakers and monitor so let's hope a wireless receiver will be built on the DRM-chip by default so we'll get cheap wireless peripherals and won't need all those cables anymore:) (everyhing will have to be powered, though. At least it'll safe CPU-cycles:)

And then offcourse one can still record the analog output of the tv, monitor or speakers but for many applications it'd be really usefull, however.

Sounds like a winner..

[digital photo]

From the sound of it, the Godson chips will be lower powered in terms of performance to current US chips. However, I find the energy consumption to be very attractive. Ie, 5 watts and 10 watts for 266Mhz 500 Mhz respectively. Scaling up linearly, that's still just 20 watts of power consumption for a 2Ghz chip.

But what I'm thinking is that China is aiming for is low cost and low power consumption chips. Ie, can be used in portable hardware and/or massively parallel setups.

Granted, they can't SMP the chips in hardware, but with a Linux cluster of these, they could quite readily setup a powerful computing cluster.

Personally, I'm glad that they are designing their own chips. It would be nice to see more competition outside of just the big two.

The way I see it, if they produce these chips at low prices($15-$50), at such low power consumption levels, I could easily see myself building many small nodes of them. Maybe now, I can POVray just ever so faster... :)