Linked by Thom Holwerda on Sun 29th Oct 2006 23:09 UTC
Morphos Genesi has announced its Open Server Workstation: "The Open Server Workstation is a six layer board with two 970MP processors, the CPC945 and Broadcom's HT-1000 and 2000 chips. Excepting those parts, the board and component cost is below USD 200. Here is the Business Plan [.pdf] we wrote for the board."
Thread beginning with comment 176939
To view parent comment, click here.
To read all comments associated with this story, please click here.
flywheel
Member since:
2005-12-28

Yes the good old x86 is very strongly represented in the cheap end of the server world, you're quite right.

Otherwise, Power (The mother of the PPC) and SPARC rules, when heavy processing are needed.

But apart from the AMD64 and AMD-V renewal of the x86 platform, nothing real new has happend since the introduction of the 386 ISA in 1985. It is old wine on new bottles, including the beloved Core and hell on earth when doing ASM programming.
The x86 ISA is very old and tired, after being constantly patched on since its birth in relation with a calculator.

Edited 2006-10-30 19:22

Reply Parent Score: 1

w-ber Member since:
2005-08-21

Saying there has been no change at all since 1985 is exaggarating...

The step from 386 to 486 wasn't big. There were a few new instructions, increase in clocking speed, and the ability to run the CPU at higher clocking speeds than the memory bus.

The step from 486 to Pentium wasn't that great either. More instructions (MMX) and a superscalar architecture (i.e. two pipelines).

Pentium Pro, however, was something else. P6 is internally a RISC processor, while previous ones are CISC. The microprogramming allows one to use CISC instructions, although they are executed as a series of RISC instructions internally. Out-of-order execution, longer pipelines.

Yes, it is fully backwards compatible, but this was also quite a feat if you look at the design.

From Pentium Pro to Pentium II was only a small step. Pentium II is essentially Pentium Pro with more (and in some models, slower) cache and a bit faster 16-bit execution mode.

And Pentium III wasn't that big either. Faster cache, SSE instructions.

However, if you look at Pentium 4, you'll see some big differences again. It's essentially a completely new CPU (using the (controversial) Netburst architecture) that just happens to be compatible with previous x86 CPUs.

Source: Wikipedia.

(Yes, I'm aware of other CPU manufacturers. However, the development in other camps is very similar to how it happened at Intel.)

Reply Parent Score: 1

flywheel Member since:
2005-12-28

I wasn't talking about the Intel implementations of the x86 i expressed myself, but about the x86 in generel.

BTW: Yes the P6 was something else. When you take a closer look at it, it looks a lot like an AMD K6, extended with a load of Cyrix 6x86 core features.

With the PII Intel added a few instructions, which sometimes makes it difficult to tell them apart. For example when you compile anything with GCC to i686, you do not compile for the P6 but the PII.

The Netburst was literally a mistake.

Reply Parent Score: 1