| Article: |
Apple's High-Water Mark? | |
| Subject: | Cell's PPE is not at all related to the POWER5 | |
| Date: | 2006-03-27 21:24:23 | |
| From: | riskin | |
|
I'm not sure where you came up with the statement that the Cell’s PPE is derived from the POWER 5, but in your own link to Anandtech's article on the cell it states "The PPE is a new core unlike any other PowerPC core made by IBM."
|
||
Showing messages 1 through 2 of 2.
-
Cell's PPE is not at all related to the POWER5
2006-03-30 17:39:12 pquam [View]
-
Cell's PPE is not at all related to the POWER5
2006-03-29 13:08:11 AdrienLamothe [View]
Digital media is where the commercial computing market is heading, it is where most of the action will be. Apple has a core group of customers, who have been with Apple a long time, who use Macs primarily for digital photos, sound and film. Cell is optimized for digital media and computational work; this makes it very attractive to people who work with digital media. You can connect the dots from here.
Regarding the Cell processor's roots in the Power5 architecture, perhaps it is best to quote an article in the IBM Journal of Research and Development, authored by the Cell architects. You can find this article at:
http://www.research.ibm.com/journal/rd/494/kahle.html
Three sentences stand out:
1. "Indications were that a completely new architecture can easily require ten years to develop, especially if one includes the time required for software development. Hence, the Power Architecture* was used as the basis for Cell."
2. "By the end of 2000 an architectural concept had been agreed on that combined the 64-bit Power Architecture* [4] with memory flow control and “synergistic” processors in order to provide the required computational density and power efficiency."
3. "The Broadband Processor Architecture extends the 64-bit Power Architecture with cooperative offload processors (“synergistic processors”), with the direct memory access (DMA) and synchronization mechanisms to communicate with them (“memory flow control”), and with enhancements for real-time management."
Regarding the difficulty of extracting Cell processor performance from a programming perspective, a large effort has been underway to provide developers with the necessary programming tools. Intel has been doing the same thing for their Core Duo architecture. A muti-core, multi-processing environment is about to become the new norm; those who refuse to adapt will suffer the same fate as others in the past who didn't change with the technology. If you want to develop relatively simple applications, then you will likely be able to continue using your current programming model. If you program on top of engines such as Apache, then you also won't likely need to worry about it; the Apache developers will port to the new environment and you won't have to deal with it.
Cheers,
Adrien



At up to 4ghz with absolutely phenomenal I/O, Cell is more than fast enough for this stuff; all of the overhead generated by techniques to try to feed the processor (like out-of-order processing) continuously actually slowed down the P4 and its sucessors, in my opinion. The consensus in the processor design community is that chips had become unnecessarily complicated with these techniques, and that they didn't pay enough attention to I/O. Cell was specifically designed to remedy that.
Cell is a specialized processor, and its specialty is in what consumers care most about: multimedia. This was also Apple's specialty until recently, which is what I assume led to this article's focus on Apple. Cell has over 200 fully programmable gigaflops, which is like 30 times more than your average processor these days (or about 3 times more than a quad G5). Cell is a consumer electronics dream.