Intel Submits Patent Covering Itanium Instructions

chris.bitmead writes: "Rather than submit garden-variety claims to the U.S. Patent and Trademark Office (USPTO), Intel is trying to patent the functions carried out by specific instructions. In doing so, the company appears to be, in effect, trying to patent the IA-64 instruction set itself." Is this an attempt to block out even reasonable competition, or is this just "business as usual" as at least one voice in the story says?

Comments

Re:oh great.

[RedWizzard]

Unfortunately modular design is very difficult when it comes to supporting multiple CPU families. The problem is, of course, that you need to swap the north bridge (what Intel now call the MCH) when you swap the CPU type because the bus protocols will be different. There has been some speculation that you might be able to drop an Athlon into a Alpha motherboard (and vice versa), as they use the same CPU bus protocol, but I've never heard it being done.

One interesting design was that of the PIOS One, which put the north bridge and the CPU(s) on a daughterboard that plugged into a PCI bus. Unfortunately it never got passed the prototype stage but it would have made for an easy way of changing CPU families and even changing the number of processors (up to four).

been there done that

[harryseldon]

Intel is hardly the first to patent aspects of an instruction set. Nobody ever cloned the VAX, because essential aspects of the instruction set were patented. The IBM 370 arch was widely cloned because it was not protected. There are intel patents on the x86 arch and microarch, but (apparently) not sufficent to prevent cloning. MIPS has patents around a few instructions but one embedded CPU company (Lexra?) got around it by not implementing those load instructions. Aspects of the Alpha arch are patented IIRC. No alpha clones, sorry. In short - looking for 'good guys'? keep looking. No modern processor instruction sets are unprotected.

not speaking for my employer. whoever that may be.

Patent App

[pjrc]

Abstract

A computational system in which an operational code (opcode) consisting of a sequence of numerical data instructs an Arithmetic Logic Unit (ALU) to perform an operation on two operands, storing the resulting arithematic or logical result into either of the operands.

What is claimed is:

1. A computational system comprising an ALU (Arithematic Logic Unit), c a operational code decoder, memory bus interface, and microcoded control logic, wherein,

Arithematic Logic Unit further comprises circuitry to perform mathematic operations of addition, subtraction, increment, decrement, multiply, divide, and logic functions of AND, OR, XOR, right and left bit shift;

2. Operational code decoder comprises circuitry that extracts encoded information to direct the activity of the Arithematic Logic Unit (claim 1);

3. Memory bus interface that transfers the operands required by the ALU (claim 1) and instruction opcodes needed by the decoder (claim 2);

4. Microcoded Control Logic which sequences the timing and produces control signals with the proper timing to direct the activities of the ALU (claim 1), Opcode Decoder (claim 2), and Memory Bus Interface (claim 3).

Description Of The Invention

The present invention related to the operation of a computational device, used to execute computational tasks, wherein the computational task to be performed by be programmed by creating a list of operational codes.

.

.

.

maybe someone else wants to continue this... it's late and I'm getting tired...

What Intel is doing...

[trims]

After reading this article, and looking at one of the patent mentioned, I'm going to hazzard a reasonable guess at exactly what's going on here:

1. They aren't patenting IA-64 codes. That is, the aren't trying to patent JMP for jump, FOO1A for super-duper multimedia instruction, etc. I don't think they're fool enough to believe that anything like that could ever get patented, or for that matter, copyrighted or trademarked.
2. It is unclear if they are trying to protect functionality or implimentation. From the patent application itself, it appears as though they are only trying to protect their implimentation of how to do interrupt context returns, not the concept of interrupt context returns itself. The article, however, sounds like they're trying for the latter.

The problem with this kind of patenting is that the "concept" is closely tied to the "implimentation". That is, the concept may be so narrowly circumscribed that any implimentation is an 'infringing' one.

Also unclear in this whole mess is how a software trap would fit in. Suppose Intel was granted the "broader" patent which covered not just the specific transistor layout of the interrupt return handler, but the more general case of returning interrupt context for IA64. Does this preclude software implimentations of that IA64 instruction (which would be particularly relevant to code-morphers like Transmeta, but even to AMD et al which do translation to microcode)?

I'm by far no Patent Lawyer. If the scope is the narrower one, I see no problem, and indeed, is well within the goal of patents. I'm alot let sanguine if the patent covers more than the circuit design, however.

-Erik

Intel IA-64 Patents not totally illegitimate

[OnanTheBarbarian]

Firstly, I would suggest that anyone who hasn't read the IA-64 architecture book, or at least a decent summary of the contents, should turn the volume down a few notches. Thank you.

This article is sort of silly. "In effect, trying to patent the instruction set itself" is a fairly vague notion; in fact, what Intel is doing is patenting some of their software techniques (expressed usually in small groups of instructions) for prefetching and control/data speculation. Right or wrong, this happens all the time. If some company has a nifty new caching algorithm, they will patent the idea; not the gatelist and implementation.

For example, if you could implement a IA-64 clone by (say) ignoring all prefetch instructions, and just fetching the data when it was needed (effectively turning the chk instructions into the actual loads, for those who are aware of this stuff - you could do it with binary translation). While this may not be a very good idea, it wouldn't infringe their prefetching patent, even if you used the same instruction mnemonics and produced a chip that could run the same binaries.

Personally, I think these patent are potentially disturbing, but put it in perspective with common practice. Read the back of Microprocessor Report sometimes; there are lists upon lists of patents being granted for techniques in exactly the same fashion as above.

As for the patents, I haven't read them, but I suspect that they'll have a tough time with them. IA-64 didn't spring out of nowhere, and a lot of the ideas that went into it follow a fairly predictable (no pun intended) path of development in academia and industry. A fairly stacatto summary of these paths can be found at Historical background for HP/Intel EPIC and IA-64 - if you don't already know something about computer architecture, don't expect to be illumined. The point is, Intel (or more accurately Idea or whatever the Intel/HP collab. is called) hasn't necessarily added that much to prior art here, so the patent may be too broad and subject to either legal attack, or too narrow and easily worked around.

Oh, and to the people cheering on the failure of IA-64, I beg to differ. Some of us write compilers and binary optimizers and code generators, and the death of the x86 architecture would make us very, very happy. The fact that the first IA-64 is going to be a dog isn't really that suprising - it's a huge engineering task and the first chip was always going to be a reference chip more than a serious performance model.

Well of course they want to block competition

[91degrees]

They're a big corporation. They'd be stupid to do anything else. The real problem is with the USPTO.

As soon as anyone gets a patent they immediately receive a virtual monopoly on that product. It lasts for a lot longer than the product could practically be useful for. It needs to be sorted out, but by whom? The US Government ain't gonna do it. They need the corporate contributions.