Following yesterday's preliminary information regarding the Cell processor, IBM, Toshiba and Sony have released official details about the setup which are very impressive and may just mean that the PS3 will qualify as a supercomputer.
The main point of the presentation of Cell architecture was that this design is the one that will offer Intel a serious run for its money. The whole design goes well beyond the PS3 and will act as a complete alternative to current architecture, although the developers of this new platform refused to comment on compatibility, either contemporary or backward.
Intel's dominance in the processor field has led to a large number of challengers over the past twenty years, all of which have failed to cause the x86 design any serious problems. In the 90s, IBM, Motorola and Apple tried to promote the PowerPC architecture. The end-result was the polarization of the computer world into the extremely popular PC and the Apple Macintosh and some IBM workstations and servers, using Power PC designs. Recently, Transmeta created the Crusoe, a challenge to Intel's notebooks dominance. Even though it was launched over 4 years ago, the format has failed to pose a serious threat to Intel and the company has all-but conceded defeat.
According to the Cell consortium, the new chips will be made up of 64-bit processors (each with 2.5 MB of memory) and should debut at 4GHz speeds.
The question of the speed has been a very important PR point which Cell developers had to address since, even though a different architecture, Cell has to be seen as offering superior power to Intel chips. Cell will also carry a greater number of transistors compared to current Intel chips with the triumvirate announcing 234 million compared to the 125 million carried by the latest Pentium 4 CPUs.
It is however, a well known fact that, this far, chip makers double the number of transistors every 12 to 18 months.
According to its creators the main advantage of Cell architecture is its flexibility. Jim Kahle, an IBM fellow said, We support many operating systems with our virtualization technology so we can run multiple operating systems at the same time, doing different jobs on the system.
This, in essence, means that it will be feasible to use a Cell system as a game machine, a desktop computer and a Television simultaneously. IBM declined to confirm which Operating Systems, other than Linux, have been tested on Cell. The question of compatibility remains a very important issue which could determine how successful any challenger to Intel's throne would be.
If Playstation3 ends up having 4 Cell processors, as has been rumored, the console will end up being officially classified, with today's standards, as a super computer.
The comparison between Intel and Cell performance however, is not feasible at this point in time as it would be limited to current Intel chips and predictions of future Cell performance. Plans for Cell suggest that it will be used in the PlayStation console in 2006 while Toshiba plans to utilize it in high-end televisions in the same year. The really important detail is that IBM has said it will sell a workstation with the chip later this year. This would mean that we could, potentially, have 3 different multi-core products in the market at almost the same time.
Intel and AMD
Intel and Advanced Micro Devices also have multi-core plans and their offerings are expected later on in 2005. Those chips will also increase the number of simultaneously executed instructions. Only yesterday Intel announced that it has completed the first tests on its dual-core processors and confirmed plans to deliver two different dual-core Pentium chips and chipset lines in Q2, 2005. Dual and multi-core processors include two or more processors on one chip of silicon. That will make things considerably faster but will mean that all software will have to be speeds up computing, although software will have to be specially written to utilize the new setup.
Developing for Cell
One of the main concerns regarding any radical change, such as is proposed by Cell, has to do with the ability of software developers to, cost-effectively, make the move to creating content for the new format. Cell's creators played down any such concerns by claiming that they will offer extensive support and training so that no such problems arise. Just as with PS2, it is likely that special teams of programmers will work at a machine level to create extensive libraries of functions which other developers should have no problems using. There are also plans for middleware applications, a set of programms which allow content creators to work without having to learn the intricacies of the system.
According to today's information we will soon have two very capable and very different systems competing for our affections, we can only hope for cross-compatibility as it is the only case in which users will have the true ability to choose the system that best fits their needs.