Cell Inside, the Future of Processor Architecture - It's the Software
(Page 7 of 7 )
It is this area that is most important to both PCs and Cell: software. Software and consumer familiarity have kept the PC alive, not good hardware. This is why making Cell hardware open source could work in its favor. IBM is probably hoping to stir some interest in the chip akin to that of the open source software community. It isn’t a new idea. There are the beginnings of an open hardware community already forming at sites like opencores.org.
And IBM may need to stimulate some creative use of programming. Telling a CPU how to use 8 SPEs may take more code, and the RISC hardware is going to put a lot more strain on software. How much boost in performace will be left after software carries out the operations that hardware once did? Perhaps the software will be Cell's largest obstacle (or downfall), as it may require more work from programmers and use more demanding software than x86 chips. Again, Mercury is expecting a jump of 3 to 4 times the performance of comparable processors to be released in the future, but we could see the final product blowing away all expectations or just else just a mediocre improvement. If AMD and Intel start on quad-cores or better, it will take away a bit of Cell's thunder as well.
IBM is trying to accelerate Cell’s development by lifting limitations and revealing its structure; in this way, talented programmers can better understand the processor and program for it more easily than otherwise. This might make it attractive for some companies that want great performance that the Cell delivers without extreme costs. IBM might also get some free hardware and software development from these opportunistic companies and from some helpful individuals. Open hardware could bring flexibility and usefulness to Cell systems that it couldn’t have reached before.
To Wrap Things Up...
Keep in mind that all these figures and expectations are speculative and have not necessarily been reached by the Cell researchers yet. The power consumption, for example, isn’t up to the planned performance. There are also some great limitations in how effectively Cells can be used in stream processing. And It may not even be as expensive as it sounds. The simplified architecture of Cell allows the chip to be far cheaper to manufacture than ordinary processors despite all the extra processing units.
It will still make the Playstation 3 awesome. If IBM and friends continue to make deals and saturate computers sector by sector, they may be starting a revolution in computer design. Eventually Cell could become viable for not just high-end media devices, but for servers and supercomputers. Hopefully it could even one day provide more rivalry for the PC than Apple has so far.
| DISCLAIMER: The content provided in this article is not warranted or guaranteed by Developer Shed, Inc. The content provided is intended for entertainment and/or educational purposes in order to introduce to the reader key ideas, concepts, and/or product reviews. As such it is incumbent upon the reader to employ real-world tactics for security and implementation of best practices. We are not liable for any negative consequences that may result from implementing any information covered in our articles or tutorials. If this is a hardware review, it is not recommended to open and/or modify your hardware. |
| Recommended by Dev Hardware |
|---|
|
/ 216 
