Intel Patents Thread Scheduling

[alextheblue]

[citation][nom]jprahman[/nom]It certainly does, at least as an improvement to it. A major disadvantage of Hyper-threading in the P4 was that two threads running on the same processor could interfere with each others cache lines. It is possible that the reason behind the reintroduction of Hyper-threading in the Nehalem architecture was that Intel was able to implement this technology, or something similar, and as a result was able to achieve a larger performance gain from Hyper-threading.[/citation]P4 hyperthreading was a mixed bag, that's for sure. In some cases it was actually detrimental to performance to have it switched on.

 

[upgrade_1977]

Now, if only they could patent a multicore chip that emulates a higher ghz lower core chip.
For example, take a octa core cpu say at 4ghz, but the computer see's it as an 8ghz quad core, or a 32ghz single core, so that non multi-threaded apps can use the otherwise wasted idle processing power.
This would be great for games that rely more on the processor then the graphics cards, like GTAIV.

Just an idea.

 

[blazorthon]

[citation][nom]upgrade_1977[/nom]Now, if only they could patent a multicore chip that emulates a higher ghz lower core chip. For example, take a octa core cpu say at 4ghz, but the computer see's it as an 8ghz quad core, or a 32ghz single core, so that non multi-threaded apps can use the otherwise wasted idle processing power.This would be great for games that rely more on the processor then the graphics cards, like GTAIV. Just an idea.[/citation]

Sorry, but our current CPU tech doesn't work like that. There is a currently little-used instruction set architecture that may be able to do this sort of thing (Explicit Data Graph execution, EDGE for short), but I think that only the TRIPS processors use it. Despite there being very little news on TRIPS over the last 8 years or so, 2012 is there target year for their performance goal, so we may hear more about it. If X86 CPUs use EDGE at their core like current X86 CPUs use an implementation of RISC then we may see this sort of thing become possible. There are a few other ISAs that are extremely similar to EDGE such as Wave Scalar if EDGE has licensing problems for something like this, but this stuff may become a reality some time in the future if our current methods for scaling performance ever higher hit an otherwise insurmountable ceiling.

 

[fuzznarf]

lol... we designed almost an identical thread scheduler for my architecture class back in 2001 as a group project. maybe i can get some royalties.

 

[fuzznarf]

[citation][nom]mrkdilkington[/nom]I wasn't aware you could patent an algorithm.[/citation]
absolutely... probably the most popular is RSA. if IP is anything, it is algorithms.

 

[fuzznarf]

[citation][nom]ramon1[/nom]People, look beyond face value, our civilization's evolution is effectively stagnating below the weight of **corporations patenting ideas**.Things like computing should be a joint effort, not a pissing contest.[/citation]
so when you invent something cool, after many long nights of hard work, and then someone comes in having the production ability that you lack, and eats your lunch, are you going to be cool with someone else exploiting all your hard work while you live in a cardboard box because you invested everything you owned in this idea? are you going to then say "good for you, glad this wasn't a pissing match"

 

[fazers_on_stun]

As a former game coder I disagree.
A good algorithm is simply providing the most effective or logical solution for a given situation.
As such it is not a "law of physics" but a "law of logic" than one patents.

This should not be confused with the algorithm for say the AI of a game. That's not necessarily dictated by outright logic of course.

Laws of nature (physics, chemistry etc) cannot be patented, nor can mathematical formulae, or software per se. The only 4 permitted categories of patentable inventions for a utility (useful) patent are a method, device or apparatus, composition of matter or article of manufacture. And currently method claims have to be tied to one of the other 3 categories, according to recent court decisions (In re Bilski in case anybody is interested). While you could view a software program as a sort of potential method, it cannot "do" anything without being run on a processor or other device that processes the instructions. Of course, to do that it has to be in a form the processor can understand - a paper printout usually cannot be used on a computer lacking scanning & OCR capabilities..

The patent rule book (or Manual of Patent Examining Practice) is something like 6 inches thick, and the US patent system is 222 years old now, ever since Thomas Jefferson was the first patent examiner - in fact, he was the entire Patent Office, along with being the Secretary of State 😛 - back in 1790. During that 2-1/4 centuries, various courts have interpreted and embellished the original patent clause in the US Constitution (article 1, section 8). So that's why patent law is pretty arcane and often doesn't make sense in modern times. For example, right now you cannot patent a signal, so if you invent a method and device for, say, flashing your BIOS or UEFI by wifi or bluetooth, you can claim the method and/or the device, but not the signal with the flash information encoded on it. Ditto with software - it has to be claimed as a method (software executing on a processor) or a device (a computer running the software) or as being stored on a non-transitory computer readable medium (like a CD, DVD, memory, flashdrive, etc), with the "non-transitory" part necessary to exclude anything like a signal. The idea behind the rules is that a signal is 'not real' or something that you can hold, sniff, see or hear or taste, and that software by itself does not do anything useful which is the basis for a utility patent.

I disagree with both those ideas - anybody who doesn't think a signal can be real should put his hand inside a microwave oven during operation, even tho it's just a 'carrier wave' and not a modulated signal. And software should be able to stand on its own without having to be claimed as something else. But these are just my opinions and unlikely to change the "business as usual" with the USPTO bureaucracy..