Intel on Monday quietly marked the 25-year anniversary of x86, the basic architecture underlying the chips that power most of today’s PCs. Elsewhere, rivals Sun and Intel on Monday announced a deal set to ensure that their mobile products will work better together over the next few years.
Intel in agreement with Sun on a Microsoft platform. Ouch.
After all this is a platform which Microsoft wants to conquer and dominate badly.
Eugenia: XScale is the keyword missing in the newspost. That’s why this entire story is interesting.
…to one of the world’s youngest archaic technological beast.
But the question is, does anyone have one of the original batch which is still in a working system…
Slightly off-topic, dreaming of my ultimate PDA.
Two years ago, I worked on a project using wireless handhelds. (Before J2ME took off)
We tried getting our AWT app to work on an iPAQ. The only JVM which worked was the blackdown ARM port of Sun’s J2RE on familiar linux. (64MB RAM/16MB ROM)
Today, one could put an ipod-style HD inside a Pocket-PC. Then one could have a fully functional mobile/wireless Linux machine.
Applications? In addition to PDA functionality, limited by open source handwriting apps admittedly, carry around a network device in your pocket and host X11 apps to ‘dumb’ desktops.
Whether Sun would spend the effort porting hotspot to XScale Linux is another matter.
What about NetBSD on XScale? That’s something I’m more looking forward to. Lean, simple and efficient. I read Wasabi Systems ported it sometime ago.
I do, I have an Amstrad PC 1640 up and running. Looking at it now.
What I find funny is that Intel didn’t plan to concentrate much on 8086. Those times its main plans were connected with super processor iAPX432 that had instructions from the high level programming languages. “The processor of the future” – see: http://www.cs.washington.edu/homes/levy/capabook/Chapter9.pdf.
Somewhere in 1982 still working on iAPX432 Intel released quick and dirty improved version of 8086 – 80286, still not believing in its future.
But noone was interested in the super processor and all wanted the x86. So Intel gave it up and started real work on a new processor which later was called 80386. And then they realised what a lame architecture they have designed before. But it was too late…
And the rest is a history.
Andreas
Yes and even if they want to kill the x86 (with Itanic,xscale(dec arm)) they can’t
, there are too many apps written for
the ugly x86 and both AMD and Intella know about this : Hammer will extends the x86 to the 64bits level and maybe Intella will use the Yahmill skunk project to extend to x86-64 too.
And if at least one of folks commenting archaism of x86 architecture actually KNEW why it isn’t the brightest and can tell at least 3 of it’s faults.
Wouldn’t be it a nice day
For example say faults of P3 architecture(and miss stuff like real mode and such which is simply tax for compatibility).
Heres 4 🙂
Huge, complex instruction set.
Mixed length instructions – makes predicting future instrcutions more difficult.
Lack of registers and half of these are special pupose. Makes instruction re-ordering more complex. Doubt it eases programming much either.
Same error in SSE – small number of registers means it can’t reach it’s true potential.
They can and do get around these by throwing hardware at the problem but chips such as the Alpha don’t need to and this is why a 7 year old core can still outperform the fastest x86 even though it runs at only 1.25 GHz.
>What about NetBSD on XScale?
Only my own personal bias. I’m primarily a Java person; Sun’s SDK is somewhat better supported under Linux (x86 at least) than NetBSD.
With x86-64 AMD got it right, I believe. The overlapping features of segmentation and paging are gone, 16 registers which is enough for building eg. complex loops, relative addressing etc. The only thing I’m a bit disappointed is that paging _has_ to be enabled in 64 bit mode, which adds another layer for memory addresses.
x86-64 – x86 done right ! 🙂
And no wonder that Intel is quiet about the 25th anniversary, it just proves the lifepower of x86, which they just ‘abandoned’.
Just my 2 european cents