Linked by Thom Holwerda on Tue 9th Sep 2008 08:52 UTC
Hardware, Embedded Systems When China launched its first microprocessor, the Godson 1 in 2002, it wasn't much of a competitor to what Intel and AMD had to offer. The 64bit Godson 2, released in 2005, still didn't worry the Western chip makers, but the chip did start to pop up here and there outside of China. Expect to see a lot more of them in the coming years, as the Godson 3 promises to be a chip that can compete head on with the big ones: quad-core, eight core version in the pipeline, and 200 extra instructions aiding in x86 compatibility.
Order by: Score:
Yeah, right...
by B. Janssen on Tue 9th Sep 2008 09:09 UTC
B. Janssen
Member since:
2006-10-11

I love MIPS (just like many others I learned Assembler on MIPS), and the article is fine, but the OSnews headline is just bunk.

Reply Score: 3

Comment by smashIt
by smashIt on Tue 9th Sep 2008 09:12 UTC
smashIt
Member since:
2005-07-06

shouldn't it be "china develops the Godson processor" instead of "intel develops the Godson processor" in the last paragraph of the news?

Reply Score: 1

The gdium is almost there
by agrouf on Tue 9th Sep 2008 09:13 UTC
agrouf
Member since:
2006-11-17

The gdium lberty 1000 netbook will have a Loongson processor: http://www.gdium.com/
Very low power consumption and no fan needed. It should have been launched at start of september, but has been delayed until the end of the month. I can't wait to put my hands on it. I want one now. This is the first time I'm excited by a new machine in years. At last a machine that does not hold us back with compatibility issues and history. Godson 3 looks very promising. I hope they don't waste too much power on x86 compatibility, because it's high time we start moving forward.

Reply Score: 5

RE: The gdium is almost there
by helf on Tue 9th Sep 2008 15:59 UTC in reply to "The gdium is almost there"
helf Member since:
2005-07-06

why? *everyone* tried killing off x86, including Intel. But it stuck around. They gave up trying to kill it and went for improving it. The new chips are perfectly fine. The design still has its downsides, but all chip designs do.
Besides, They are slowly morphing the x86 architecture into something completely different. Just give it time ;) The new nehalem core is awesome.

Reply Score: 3

RE[2]: The gdium is almost there
by agrouf on Tue 9th Sep 2008 16:58 UTC in reply to "RE: The gdium is almost there"
agrouf Member since:
2006-11-17

I believe the conditions are there now. Open source software has enough quality and momemtum to provide a very functional system on any architecture. Previously, market share did matter a lot. Nowadays, even if your architecture has less than 1% marketshare, you can have a lot of functional and quality software to run on it.

Reply Score: 2

RE: The gdium is almost there
by psychicist on Tue 9th Sep 2008 17:11 UTC in reply to "The gdium is almost there"
psychicist Member since:
2007-01-27

There are very good reasons to be excited. The processor performs reasonably well and consumes little power. The x86 compatibility will not be hard coded but provided by a combination of binary translation instructions and software.

For example SPARC compatibility is going to be provided as well, if I correctly remember what I read on Loongson 3. For now I'm happy with my Loongson 2E desktop and am looking forward to new 2F based desktop and notebooks.

The single core 2G processor is going to be taped out soon and it's likely Loongson 3 will contain four of these cores.

Reply Score: 3

RE[2]: The gdium is almost there
by collinm on Wed 10th Sep 2008 06:49 UTC in reply to "RE: The gdium is almost there"
collinm Member since:
2005-07-15

can you make some bench to show performance?

Reply Score: 2

[mmmm]
by alexis on Tue 9th Sep 2008 11:07 UTC
alexis
Member since:
2007-05-21

Looks promising as long the Chinese government don't put some user controls in it.

Reply Score: 1

Cool
by kaiwai on Tue 9th Sep 2008 11:08 UTC
kaiwai
Member since:
2005-07-06

I was having a look at a small netboot/laptop, its too bad that one can only purchase them in China. This processor sounds ideal for the NetBook and developing markets.

The new processor will also be a great addition, I only hope that more software vendors, open and closed source, get behind this platform.

Reply Score: 1

netbook, laptop, computer...
by collinm on Tue 9th Sep 2008 11:14 UTC
collinm
Member since:
2005-07-15

i hope some constructor will propose this cpu for user

x86 compatibility is fine for joe sixpack who fear to use another thing then popular cpu

hope many linux distribution will use it

bty... why nobody use sun gpl cpu?

Reply Score: 2

RE: netbook, laptop, computer...
by kaiwai on Tue 9th Sep 2008 11:18 UTC in reply to "netbook, laptop, computer..."
kaiwai Member since:
2005-07-06

i hope some constructor will propose this cpu for user

x86 compatibility is fine for joe sixpack who fear to use another thing then popular cpu

hope many linux distribution will use it

bty... why nobody use sun gpl cpu?


*shrug* SPARC IIe was an ok processor, it had a lower power foot print, I'm surprised no one has come forward and attempted to make it even more low powered. Then again, I guess it is alot easier to do something from the ground up than taken an out of date idea and retrofit it to todays requirements.

Reply Score: 1

collinm Member since:
2005-07-15

there are great software engineer in the open source world...

could be nice if some electric engineer start to do open source hardware...

like an apn, router, netbook....

it's sure there are very greate electric engineer who read osnews...

Reply Score: 2

Laurence Member since:
2007-03-26

there are great software engineer in the open source world... could be nice if some electric engineer start to do open source hardware... like an apn, router, netbook.... it's sure there are very greate electric engineer who read osnews...


There have been open source hardware (Open source CPUs (as featured on here), Reactable, etc)
RE routers: Essentially they're just a low powered PC with a thin OS - you can buy cheap computers to build as a router (again there was an article on OSNews about it) and there's plenty of open source router OSs available to download (most of which appear to be Linux based).

Reply Score: 2

collinm Member since:
2005-07-15

i know there are opensource hardware like cpu and see... but you can't really find a product who use it easily

now there is a cellular phone opensource... greate we can buy it... but there are not enough..

an apn, gps, netbook... could be greate

there are a lot of small device with io on linuxdevices who go out everyday... there a lot of potential...

Reply Score: 2

theosib Member since:
2006-03-02

The Open Graphics Project and Open Hardware Foundation are two organizations among several who are actively developing "open source hardware". The Open Graphics Project already has an FPGA-based development board up for pre-order, and they're almost done with basic (unaccelerated) graphics hardware for it. Among other things, they have developed a simple harvard architecture MIPS-inspired controller they call HQ.

Reply Score: 1

collinm Member since:
2005-07-15

a couple opensource hardware...
http://en.wikipedia.org/wiki/Open_source_hardware

but if i go to radioshak... i will not find it

Reply Score: 2

design vs produce
by lemoncookie on Tue 9th Sep 2008 11:20 UTC
lemoncookie
Member since:
2006-03-01

AFAIK, Chinese can design chips but can't produce them.
That's the bottleneck.

Reply Score: 0

RE: design vs produce
by collinm on Tue 9th Sep 2008 11:41 UTC in reply to "design vs produce"
collinm Member since:
2005-07-15

i think the majority of the fab are in asia...

Reply Score: 2

RE: design vs produce
by agrouf on Tue 9th Sep 2008 11:42 UTC in reply to "design vs produce"
agrouf Member since:
2006-11-17

But they do produce them since 2002.

Reply Score: 2

RE: design vs produce
by Adurbe on Tue 9th Sep 2008 14:40 UTC in reply to "design vs produce"
Adurbe Member since:
2005-07-06

China is a production based economy

If they want to make it, they can, and will!

Until recently there has ben a shortage of tech skills in the country, that is FAST changing.

Wait 5/10 years and these chips will be a safe number 3 after Intel and AMD (I think they will beat Via)

Reply Score: 2

RE[2]: design vs produce
by helf on Tue 9th Sep 2008 16:01 UTC in reply to "RE: design vs produce"
helf Member since:
2005-07-06

I doubt it.

Reply Score: 1

RE: design vs produce
by psychicist on Tue 9th Sep 2008 17:20 UTC in reply to "design vs produce"
psychicist Member since:
2007-01-27

You may be right about the fact that Chinese manufacturing processes are not yet suited to the production of such sophisticated processors as Loongson.

That's the reason that ST Microelectronics has been involved with the design and production of these processors since 2004. They have a five year contract to manufacture and sell them worldwide and also acquired a MIPS64 license.

In a few years time when modern manufacturing plants will have been built to take care of production they may actually be able to produce the processors themselves. They have to go through a process during which mistakes and improvements are made based on experiences with real silicon, but will undeniably have caught up in a few years.

Reply Score: 2

v It should be easy to take on
by aperh on Tue 9th Sep 2008 12:25 UTC
List of instruction?
by renox on Tue 9th Sep 2008 13:16 UTC
renox
Member since:
2005-07-06

I wonder what those instructions are?

The only one that I could think of are 'little endian' load|store, what are the other instructions which were added to simplify x86 emulation?

80bit float computation? MMX/SSE instructions?

Reply Score: 2

why do they an x86 license
by Yamin on Tue 9th Sep 2008 13:47 UTC
Yamin
Member since:
2006-01-10

Do they really need an x86 license to implement the x86 instruction set?

Pardon my lack of knowledge about IP ;)
It just weird that the instruction set needs to be licensed. It would be like Microsoft restricting the MS Word file format and then requiring you to buy a license to make a compatible product.

Reply Score: 2

RE: why do they an x86 license
by smitty on Tue 9th Sep 2008 17:13 UTC in reply to "why do they an x86 license"
smitty Member since:
2005-10-13

It's not really any different than the idea of patenting an API, although in this case it actually is in hardware rather than purely software.

When AMD first started making x86 compatible chips, Intel sued them for violating their IP. The case was eventually settled out of court (AMD got the x86 ISA and Intel got all future IP AMD created which is why they didn't have to worry about copying AMD64 or HyperTransport), so I don't think there is any precedence one way or the other about it.

Reply Score: 3

Intel is not scared
by truckweb on Tue 9th Sep 2008 14:34 UTC
truckweb
Member since:
2005-07-06

Intel already reported that Godson3 will only reach 60-70% of the performance of Intel CPU because of the decoding required to process x86 instruction on MIPS hardware. (I read somewhere that the core where from SPARC CPU).

So, the idea sound cool, but it's nothing to scare Intel or AMD in the near future.

Reply Score: 2

RE: Intel is not scared
by werpu on Tue 9th Sep 2008 14:40 UTC in reply to "Intel is not scared"
werpu Member since:
2006-01-18

Actually Intel should be worried, it is not a tie in situation anymore, if it provides 70% of the speed of a native intel processor it probably is good enough to run windows for many users, but you can run Linux and pretty much every other non Windows OS natively on Mips and that is probably where this CPU really shines. I can see it appearing in many devices which run Linux and other operating systems (and probably have a somewhat slow Windows option as well)
good enough for many applications not all but many...
Besides that the Mips processor is superior to anything intel has put out in the x86 space...

Reply Score: 4

RE[2]: Intel is not scared
by helf on Tue 9th Sep 2008 16:06 UTC in reply to "RE: Intel is not scared"
helf Member since:
2005-07-06

70% of a native CPU... but at what speed? If it's going to run like a 600mhz p3... who wants that? ;)

Reply Score: 1

RE[3]: Intel is not scared
by werpu on Tue 9th Sep 2008 17:31 UTC in reply to "RE[2]: Intel is not scared"
werpu Member since:
2006-01-18

70% of a native CPU... but at what speed? If it's going to run like a 600mhz p3... who wants that? ;)

only in the software/hardware intel emulation, if you run in native mode, which is probably every os outside of the windows world which can do that (the mips instruction set is a very popular one among non intel processors)
then the difference in speed probably will be completely different...

Reply Score: 1

RE[4]: Intel is not scared
by dmantione on Tue 9th Sep 2008 17:54 UTC in reply to "RE[3]: Intel is not scared"
dmantione Member since:
2005-07-06

... but not threathening. The MIPS instructuction set is rather cripled, having only simple addressing modes, so you need more MHz to compensate. The promise has always been than the simpler design of the MIPS would allow this, but this has been a broken promise, nobody has managed to make a MIPS that significantly exceeds x86 clock speed to make up for the reduced performance.

The extra instructions will probably not fix the MIPS addressing mode scheme, but rather help emulating several x86 instructions which don't map well to native MIPS instructions.

Reply Score: 1

RE[5]: Intel is not scared
by helf on Tue 9th Sep 2008 19:26 UTC in reply to "RE[4]: Intel is not scared"
helf Member since:
2005-07-06

plus, didnt Windows at one point run on MIPS? I think NT did and some builds of Windows 2000. So MS could easily port the OS again.

Reply Score: 2

RE[5]: Intel is not scared
by renox on Tue 9th Sep 2008 20:18 UTC in reply to "RE[4]: Intel is not scared"
renox Member since:
2005-07-06

... but not threathening. The MIPS instructuction set is rather cripled, having only simple addressing modes, so you need more MHz to compensate.


I disagree about the 'more MHz' oversimplification: you can use the transistors saved by the simplicity of RISC ISAs to build CPU with more instruction level parallelism, more OOO, more cache,etc not simply 'more MHz'.

Note that now some RISC family (including MIPS) have 16 and 32 bit instructions which allow them to have near x86 code density while still having a simpler implementation.

Reply Score: 2

RE[6]: Intel is not scared
by dmantione on Wed 10th Sep 2008 06:18 UTC in reply to "RE[5]: Intel is not scared"
dmantione Member since:
2005-07-06


I disagree about the 'more MHz' oversimplification: you can use the transistors saved by the simplicity of RISC ISAs to build CPU with more instruction level parallelism, more OOO, more cache,etc not simply 'more MHz'.


Addressing modes have little to do with RISC/CISC. The ARM has more complex addressing modes than x86, and still is considered a RISC. How much die spice does an extra adder and barrel shifer occupy? Nothing.

I'm not claiming CISC is better than RISC, but complex addressing modes do have won from simple addressing modes.

Note that now some RISC family (including MIPS) have 16 and 32 bit instructions which allow them to have near x86 code density while still having a simpler implementation.


No, that's what the MIPS marking says. Now the reality:

xor eax,[4*ebx+ecx]

is 3 bytes in x86. In ARM we have:

EOR r0,r1,r2,lsl#2

... which is 4 bytes.

How many instructions and bytes and help registers do you need on MIPS?

Reply Score: 1

RE[5]: Intel is not scared
by kaiwai on Wed 10th Sep 2008 02:27 UTC in reply to "RE[4]: Intel is not scared"
kaiwai Member since:
2005-07-06

... but not threathening. The MIPS instructuction set is rather cripled, having only simple addressing modes, so you need more MHz to compensate. The promise has always been than the simpler design of the MIPS would allow this, but this has been a broken promise, nobody has managed to make a MIPS that significantly exceeds x86 clock speed to make up for the reduced performance.

The extra instructions will probably not fix the MIPS addressing mode scheme, but rather help emulating several x86 instructions which don't map well to native MIPS instructions.


When SGI owned MIPS, MIPS were more than capable of holding their own against Intel. What was lacking, however, was the R&D budget the size of Intel to turn a lot of the 'drawing board' ideas into reality. when one's ideas are crippled by the lack of funding, it is hardly true to blame the instruction set for MIPS woes.

I remember back when an SGI engineer spoke the future of processors (way before dual core) and he spoke of massively parallel, multi-core, smp configurations. Highly efficient CPU's working in parallel instead of having massively highly clocked CPU's. I remember the benchmarks of the R20K, for example. The prices of the hardware were very high (SGI) but the performance of the processor itself rivalled Intel/x86 even at the low clock speed which it was set. Gigahertz machines being beaten by a processor of around 600Mhz IIRC.

But then again, re-writing of history on this forum seems to be a popular past time.

Reply Score: 1

RE[6]: Intel is not scared
by dmantione on Wed 10th Sep 2008 06:33 UTC in reply to "RE[5]: Intel is not scared"
dmantione Member since:
2005-07-06



When SGI owned MIPS, MIPS were more than capable of holding their own against Intel. What was lacking, however, was the R&D budget the size of Intel to turn a lot of the 'drawing board' ideas into reality. when one's ideas are crippled by the lack of funding, it is hardly true to blame the instruction set for MIPS woes.


I don't deny that Intel budgets play a role, but I do maintain my position that the MIPS instruction set is too cippled to be competitive. True, the x86 instruction set sucks too with its stack based FPU and its lack of registers, but hey, this has been fixed during the past years.

The prices of the hardware were very high (SGI) but the performance of the processor itself rivalled Intel/x86 even at the low clock speed which it was set. Gigahertz machines being beaten by a processor of around 600Mhz IIRC.


Well, Intel 1Ghz machines were also beaten by 600 MHz AMD machines at that time. At that time, the x86 had still only 8 registers and SSE was only just available, most software used the x87 FPU. But we shall see, with the Chinese government behind MIPS will receive both budget and inexpensive hardware will be available, so either MIPS's flaws will be fixed too, or it will prove to be fundamentally inferior.

Edited 2008-09-10 06:44 UTC

Reply Score: 1

RE[7]: Intel is not scared
by kaiwai on Wed 10th Sep 2008 06:57 UTC in reply to "RE[6]: Intel is not scared"
kaiwai Member since:
2005-07-06

I don't deny that Intel budgets play a role, but I do maintain my position that the MIPS instruction set is too cippled to be competitive. True, the x86 instruction set sucks too with its stack based FPU and its lack of registers, but hey, this has been fixed during the past years.


Which is the whole point of the design. Have a small number of primitive instructions, and use those to build up and make more high level ones. That is the whole point of the RISC design; keep it simple and highly optimised, and use those as building blocks for more complex things.

Now, that might be an issue with a crappy compiler which has not been written from the ground up with that paradigm in mind. The reality is that if a compiler were to be designed solely to compile software for that processor alone - and thus not have to be compromised for the sake of supporting other processors, I don't see how it could be a failure.

GCC, for example, is primarily designed for portability. The binaries produced by GCC is ugly at the best of times, and given that the x86 processors are pretty much happy to chew through any crap you throw at it - a crappy compiler isn't going to severely hamper performance to such a degree that a poorly written compiler would with a strictly RISC based processor.

Well, Intel 1Ghz machines were also beaten by 600 MHz AMD machines at that time. At that time, the x86 had still only 8 registers and SSE was only just available, most software used the x87 FPU. But we shall see, with the Chinese government behind MIPS will receive both budget and inexpensive hardware will be available, so either MIPS's flaws will be fixed too, or it will prove to be fundamentally inferior.


I don't know what the hell you've been smoking - the first gigahertz processor put out was the AMD Athlon - back when they were using the old slot A design for their processors. They were the first to bring gigahertz to the desktop. Intel was only fractionally behind it. The difference between gigahertz to gigahertz only really started to be noticeable when the P4 was launched and the older PIII processors when clocked at 1.3Ghz were beating the P4 clocked at 1.7Ghz.

Edited 2008-09-10 06:58 UTC

Reply Score: 1

Yamin
Member since:
2006-01-10

It's a bit worry hearing people talk about China as if they can't build anything. That all they can make are knockoffs...

Hey folks... take a trip down to silicon valley one day. It's full of Indians and other Asians including Chinese people. They're not dumb people. They've just needed the proper business environment in which to thrive. The people have the knowledge. The government and business have the will.

It wouldn't surprise me one bit if China made a processor not just equal to, but better than Intel.

Reply Score: 3

Bad News
by beosguy@gmail.com on Tue 9th Sep 2008 17:48 UTC
beosguy@gmail.com
Member since:
2008-07-17

Not good news at all... Some seem to forget how the Japanese overproduced and dumped DRAM causing a glut and driving all including Intel and AMD out of that business. 1986-88 was brutal to be in Semi. You can kiss those high paying jobs in Silicon Valley if this chips hit the shores.

BTW the Japanese tried to export to the US x86 clones Nec V20 and V30 but Intel stopped them. Much of the production comes from Singapore and Taiwan anyway.

Reply Score: 1

RE: Bad News
by apoclypse on Tue 9th Sep 2008 19:51 UTC in reply to "Bad News"
apoclypse Member since:
2007-02-17

Its not China's fault that AMD/Intel were too damn lazy to get there asses in gear when they had the chance. The cpu market needs to drop a stale platform before things get really out of hand and if it takes a Chinese company to light a fire under Intel's feet so be it. Besides all the high paying tech jobs are already slowly being shipped overseas as other nations overtake us in high tech education. You can blame your current president for that one. no child left behind my ass, things are worse than they ever been. As an American its hard to see competition knocking on our door however its something we should blame on ourselves.

Reply Score: 3

RE[2]: Bad News
by Yamin on Tue 9th Sep 2008 20:05 UTC in reply to "RE: Bad News"
Yamin Member since:
2006-01-10

There's no problem in our education system. Plenty of Americans and Western people in general are very capable of being engineers and scientists.

The only question is why would they?

It's much easier going into medicine, law, finance, education, business, management... easier and more pay.

I recall from my high school, some of the dumbest kids are just going to the Caribbean for med school. They'll be doctors pulling in 150-200K/year just using the system. All they need to do is pass the USMLE which they get an unlimited number of tries... some have already pased.

Why would anyone, besides the nerdiest people really go into the engineering/software field? They wouldn't. Other fields just pay too damn well. With our aging population, you can bed the health care field is going to get even more lucrative.

You blame George Bush. I blame the democrats for making health care and education pay too well ;) medicare/aid is welfare for doctors. They never made this much more than teachers or engineers before it. There is an adjustment in the world. Our quality of life has to go down. There's is going up. That's just reality.

Edited 2008-09-09 20:06 UTC

Reply Score: 2

RE[3]: Bad News
by rayiner on Tue 9th Sep 2008 23:46 UTC in reply to "RE[2]: Bad News"
rayiner Member since:
2005-07-06

I disagree that there is no problem with our education system. Pretty much my first year of college was spent undoing the brain-damage that had been done to students in public schools.

That said, I agree with you about the cost/benefit issues of going into engineering. If you're smart and hard-working enough to get through engineering school, then you're smart and hard-working enough to get a good LSAT score and do well in law-school. Considering the pay difference between a good lawyer and a good engineer...

The problem along those particular lines is that engineers themselves haven't protected the value of their profession in the way doctors and lawyers have. Engineers create plenty of value in the economy, but they're quite willing to simply be worker-bees and let their company take most of the value they create. This is true even among very good engineers. There is nothing akin to, for example, the private law segment of the law profession, where partnership allows individual employees to get a stake in their firm. Engineers have allowed their profession to become commoditized, even as the rigorous demands for working in the field have not been reduced.

Reply Score: 2

RE[4]: Bad News
by Yamin on Wed 10th Sep 2008 00:03 UTC in reply to "RE[3]: Bad News"
Yamin Member since:
2006-01-10

I agree with you 100%.

Engineers could have protected their field through licensing. Then again, we probably wouldn't have had facebook ;) But especially for government, banks, operating systems...

On the other hand, if licensing protection is not your cup of tea, engineers could have at least formed partnerships / employee owned corporations instead of being corporate workerbees.

On the other hand, many engineers miss the 'easier money', to provide a constant stream of money. Google for example makes money off ads and then funnels that into all kinds of projects some of which don't have a direct revenue stream. Imagine if Engineers owned a telecom or something. We could plow the money earned from those into R&D and what not.

Nonetheless, I think we're in agreement here ;)
Suffice to say, my kids aren't going into this field.

Reply Score: 1

RE[4]: Bad News
by unclefester on Wed 10th Sep 2008 06:29 UTC in reply to "RE[3]: Bad News"
unclefester Member since:
2007-01-13

There is a huge difference between writing software (it ain't engineering BTW) and medicine. A software crash is in most cases no more than a nuisance. The simplest medical 'glitch' is highly likely to result in serious injury or death.

For the most part software architects do not act like engineers. The primary goal of an engineer is to ensure safety and reliability. An engineer (or doctor) can be sued or even imprisoned for negligence. No commercial software writer has that pressure.

Reply Score: 2

RE[5]: Bad News
by collinm on Wed 10th Sep 2008 13:00 UTC in reply to "RE[4]: Bad News"
collinm Member since:
2005-07-15

i hope it's a joke?

you don't remember the software bug who died some astronaut

you don't remeber the software bug in the medical cancer machine who burned some burned...

Reply Score: 2

RE[5]: Bad News
by Yamin on Wed 10th Sep 2008 15:26 UTC in reply to "RE[4]: Bad News"
Yamin Member since:
2006-01-10

"software crash is in most cases no more than a nuisance. The simplest medical 'glitch' is highly likely to result in serious injury or death. "

I'm going to preface this by saying I agree with your post.

Unless you're talking surgery or major prescription, an error is unlikely to result in a major problem. But even surgery, that's really no different from so much of the other work that occurs. What happens if the surgeon screws up... someone might die. What happens if the construction workers doesn't weld something correctly and the engineer misses it? The bridge could collapse killing hundreds. What happens if the software engineer writes the wrong code for your automobile or airplane? It could cause it to crash. What is the minimum wage restaurant worker screws up and infects your food? You could also die then. What happens if the software of the airline industry fails and holds up a million people?

Should we also pay the restaurant worker, welders, and civil engineers 200K/year?

There are consequences to everything that affect your life. Singling out doctors is just not that accurate.

True and that is part of the problem that we have had without licensing software.

Everything runs on software these days. From you car, to banks, to social networking sites, to telecommunications. These all have severe reliability and security considerations. Look at what has happened because qualified people have not been designing these systems. We've had identity theft, unsecure network, system crashes (for example the one on the London Stock Exchange recently).

Truth be told though, software cannot be like all other engineering. Every thing you do is new. Most of the time in other field of engineering the 'new' stuff is only done by really experienced top of the line engineers or scientists. Most regular engineers just follow the same old routine.

In software, it's not uncommon to have a new grad just out there writing the software that runs the internet.

That said, it's doubtful we'd have had the innovation in the software world we it so regulated. So it's been an interesting trade off.
For me personally, one of the reasons I've been frustrated in the field is the lack of quality people and products. It's tiring working a bridge that's been duck taped and super glued together ;)

Edited 2008-09-10 15:31 UTC

Reply Score: 1

RE[3]: Bad News
by StychoKiller on Sun 14th Sep 2008 04:22 UTC in reply to "RE[2]: Bad News"
StychoKiller Member since:
2005-09-20

Medicare/Medicaid is welfare for sick people -- and Doctors are expected to sacrifice the payments due them because these people are entitled to subsidized medical care. If younger people voted in larger numbers than the AARP crowd, Congress would be forced to change the status-quo.

Reply Score: 1

unclefester
Member since:
2007-01-13

The Chinese government wants a totally independent software/hardware platform. Tn other words no x86 or Windows.

The Chinese government hase no interest in whether this is financially viable. It will be done regardless of cost.

They will achieve this by mandating the use of Godson/Linux in certain areas such as junior schools within a certain time frame. This is an instant market of 100 million machines.

CPUs will be sold at or below cost. They will be sold for virtually nothing at the low end (<$5) and probably 1/10th the price of any performance comparable Intel/AMD CPU. This will destroy AMD and severely weaken Intel. MS will eventually be locked out as the move from x86 is completed.

Reply Score: 2

psychicist Member since:
2007-01-27

That's exactly what it's all about. I've been working with the main vendor of Loongson based computer systems ever since I heard about the release of the first systems one and a half years ago.

Suffice it to say that many companies and individuals (including me) both inside and outside of China have been working very hard to get all software ported over to the Loongson processor.

That's because it's finally a credible alternative to the complete dominance of the computer landscape by Intel and Microsoft that would otherwise be the case, not only for China but the world at large.

In a way it's about breaking Intel's monopoly on microprocessors the same way GNU/Linux has broken MS Windows' dominance in operating systems. GNU/Linux runs on many architectures, it's just closed source that doesn't because of its inherent limitations.

That's not a reason or excuse to keep an otherwise very good operating system from being used on desktops and notebooks based on non-x86 platforms. You simply can't take closed source software into account during the development of a free operating system, only make it work as well as is realistically possible.

Getting x86 software for Windows and Linux to work on these processors is nothing but an afterthought. That may be hard to believe for some people but it's what's necessary to make a clean break from the past and take matters into your own hands.

I hope a European alternative will be established too, possibly based on the ARM Cortex A9 architecture. If people really believe in free market principles the additional competition should only be encouraged in a market that's been dominated by a few proprietary companies for far too long.

Reply Score: 2

agrouf Member since:
2006-11-17

The loongson is produced by STMicroelectronics, a European company, in collaboration with the Chinese Academy of Science.
All this talk about the free market and americans feeling threaten by the free market makes me laugh though. Your real problem is not the chinese or that they build or may build cheaper and better processors. No, the real problem is your twisted view of the world as a scene of war where competition is always better than collaboration and the more you compete the better, no matter how much you destroy, except when you loose. The problem is in your mind, really.

Edited 2008-09-10 05:37 UTC

Reply Score: 1

psychicist Member since:
2007-01-27

I haven't said that American companies should stop to exist and I wouldn't want to since I'm happily using Intel and AMD processors here, but seriously, what's wrong with another big party to give the customer an additional choice for a microprocessor?

It doesn't even matter to me that it's a Chinese or an American processor, it's just an alternative to the obvious assortment of x86 processors that are on the market.

You are putting words into my fingers that I haven't even uttered. If you knew more about the situation you would know that there is a greater landscape of semiconductor companies that have a right to sell product than your precious Intel or AMD, the latter of which which isn't even second or third on the list of largest semiconductor companies.

It's too bad for them Intel doesn't license the x86 architecture so in order to have a competitive product on the market, they have to resort to licensing another one. Other architectures such MIPS, ARM, PPC and SPARC are being licensed, so what other choice does a starting (fabless) semiconductor company have than to license one of these?

For that matter it is Intel that is dominating the market in an unhealthy way, otherwise you would have a choice of an x86, SPARC, MIPS, PPC, ARM or another implementation for your general computing needs.

I run all of these, but don't make the mistake that I'm waiting for American companies to die or to be destroyed, that's laughable since four of these were invented in America. It's about a return to the healthy situation we had 10 years ago when you had a choice of systems from various vendors before Wintel destroyed that landscape.

I also wish for several of the companies that design the processors to collaborate, as they're already doing to a large extent. After that there is a choice of having completed designs manufactured at several manufacturing plants such as those from IBM, TSMC and UMC. Can you tell me what's wrong with that (it's today's reality anyway)?

Edited 2008-09-10 12:23 UTC

Reply Score: 2

apoclypse Member since:
2007-02-17

As an American I totally agree. We have been trained in arrogance as a nation that we must be the best or destroy every other competition, that we must be the leader and not the follower. Unfortunately for us the world had changed drastically, some of it due to things put in motion by our own government and now you have other nations who can do it cheaper and better all working together and here we want to be king of the hill or take our ball home. The kicker is that taking the ball home in this instance can gravely affect our already ailing economy if we don;t get our asses in gear.

Reply Score: 2

bnolsen Member since:
2006-01-06

Actually that's not true. In the world american business has been looked at as being a good model of integrety. In the outsource countries I know of the premium prize jobs are with American companies, followed by german ones with France near the bottom.

I'm not a big fan of them "just adding" a whole bunch of new instructions just to accomodate x86. The point of a good processor design is to keep it simple, not junk it up with debateable instructions. I would have preferred it if they left the original instruction set alone and came up with a very specialized VM type co processor. Or else use the extra silicon space to add more cores and use those for possible on the fly recompilation.

Edited 2008-09-11 05:19 UTC

Reply Score: 2

agrouf Member since:
2006-11-17

Actually that's not true. In the world american business has been looked at as being a good model of integrety. In the outsource countries I know of the premium prize jobs are with American companies, followed by german ones with France near the bottom.

The problem is this nationalist attitude. The world has moved on. It doesn't matter that the company is american of french or german or chinese. There are those people who preach the free market as the ultimate end of history and yet they are upset when a non-american organisation does things that is not under american control. That makes me laugh a lot. The american companies don't give a shit about that. If there are profit to be made that way, they will move to China or France or Bangladesh.

Edited 2008-09-12 10:54 UTC

Reply Score: 2

bnolsen Member since:
2006-01-06

Okay so I've got high performance multithreaded software I'd like to sell in the chinese market. There ARE buyers available. Is there any way to get information or even sample hardware (more than one core) to work on? The only immediate problem *could* be that we've coded some IO assuming little endian, but that's very easy to isolate and fix.

Reply Score: 2

psychicist Member since:
2007-01-27

The problem is that Loongson 3 hasn't even taped out, so unless you get hold of a multiprocessor system based on Loongson 2E/2F, which apart from the KD50 (prototype of a supercomputer) I haven't heard of yet, there is nothing to do but wait until the hardware is being sold.

Loongson is a little-endian MIPS III compatible processor so I don't think you'd have to change that much. But it would of course be good to make the changes required to make your code run on big-endian architectures anyway.

Reply Score: 2

herodoug Member since:
2008-09-10

The Chinese government wants a totally independent software/hardware platform. Tn other words no x86 or Windows.

The Chinese government hase no interest in whether this is financially viable. It will be done regardless of cost.

They will achieve this by mandating the use of Godson/Linux in certain areas such as junior schools within a certain time frame. This is an instant market of 100 million machines.

CPUs will be sold at or below cost. They will be sold for virtually nothing at the low end (

Please state your source, I'd love to look at it myself.

Edited 2008-09-10 13:06 UTC

Reply Score: 1

gdium ...
by collinm on Wed 10th Sep 2008 18:02 UTC
collinm
Member since:
2005-07-15

a netbook with godson 2 cpu will go out soon

http://www.gdium.com/

mandriva will create a special release for this netbook...

Reply Score: 2

"godson" processors?
by thebackwash on Thu 11th Sep 2008 15:06 UTC
thebackwash
Member since:
2005-07-06

"Godson"

Wow, they're a cocky lot, comparing themselves to Jesus. Maybe this will be the second coming of computing.

Reply Score: 2

RE: "godson" processors?
by agrouf on Fri 12th Sep 2008 10:39 UTC in reply to ""godson" processors?"
agrouf Member since:
2006-11-17

It's a chinese word. It means the "dragon chip".

Reply Score: 2

Some Loongson computers
by twitterfire on Thu 11th Sep 2008 18:42 UTC
twitterfire
Member since:
2008-09-11

Here you have a nice selection of Loongson based desktops and laptops: http://www.lemote.com/english/

Reply Score: 1