No announcement yet.

Gone are the days when the CPU was the only central powerhouse

  • Filter
  • Time
  • Show
Clear All
new posts

  • Gone are the days when the CPU was the only central powerhouse

    The CPU redefined: AMD Torrenza and Intel CSI.

    AMD buys out ATI, Intel and Nvidia seem to form a strategic alliance; in short these are turbulent times in the world of chip manufacturers. If we are to believe the statements made by several manufacturers, then we are on the threshold of some major changes concerning task applications for the various chips in a PC. Though AMD has gotten a head start with Torrenza, Intel won't be far behind in implementing a similar solution.

    Gone are the days when the CPU was the only central powerhouse of a computer. The first example that springs to mind is the GPU, the cetral chip on the video card: modern 3D chips are completely programmable and offer an almost unbelievable amount of raw processing power, but it doesn't stop there. Other examples are Creative's X-Fi chip, which has been implemented in its latest soundcards, which is so powerful that by right it can actually be called an APU (Audio Processing Unit) and Ageia's PPU's (Physics Processing Units). These PPU's are used dedicated to processing the physics calculations within games. Another company to appear on the news a short while ago was AISeek. This company made tech related headlines with the announcement that it was developing a processor dedicated to handling the Artificial Intelligence calculations within 3D games. Another prime example is the professional world where there are numerous chips capable of handling various calculations very efficiently hence decreasing the workload for the CPU.


    All the afore mentioned examples show that specialised chips for specific tasks can be quite a smart move under certain circumstances. Take for example processors such as the AMD Athlon 64 or the Intel Core 2 Duo; these have to be capable of processing almost any form of code. The downside of this is that they aren't optimised for a specific task, hence their basically jack's of all trades but masters of none. Taking this into account it should come as no surprise that a modern 3D-chip is capable of creating beautiful images a hundred times faster than any CPU. The same counts for any of the aforementioned examples.
    In the past the processor was the beating heart of the computer (hence the term Central Processing Unit) but with all the different developments in abovementioned areas the CPU is becoming a less determining factor for overall processing power within the PC. Modern GPU's have already taken over many of the tasks that CPU's did in the past and as time goes by we will see more and more of these specialised chips. Several people within the IT-industry predict that in the near future we will see more specialised co-processors appearing on the market, for any form of task. If you mainly use your PC for video editing, then wouldn't be more efficient to have a specialised co-processor capable of handling the required calculations far more efficiently than a standard CPU? Or when a PC is used for certain scientific calculations, wouldn't it be far handier to have a co-processor which had been specialised for that certain area?

    With al the GPU's, APU's, PPU's and AIPU's we can definitely say that the Personal Computer has become more powerful and diverse than ever before, but the flipside of this is that we have to look critically at how we implement these various innovations within a PC: using only the PCI- and PCI-Express-bus for expansion cards with specialised processors is not exactly efficient by any means....

    Stream computing
    In terms of co-processors it's quite interesting to see what ATI has been up to lately. In the last couple of years GPU's have developed from relatively simple chips, capable of handling only a few standard processes, to veritable processing behemoths. GPU's became fully programmable with the advent of DirectX 8 and are even capable of working with 32-bit floating point numbers since DirectX 9, this has led to more and more companies using GPU's for far more than just calculating 3D images. The most notable step forward has been the recent announcement by both ATI and nVidia to implement their GPU's as PPU's to handle physics calculations in 3D-games, this as a reaction to the PPU-chip from Ageia. ATI has taken this one step further in announcing a far wider scope of possible uses for their GPU's: ATI, under the codename Stream computing, is proactively supporting the implementation of their GPU's for a myriad of calculation processes.

    During a recent presentation ATI emphasized that a GPU is far more suitable for working through various calculations than a traditional CPU. According to ATI the latest generation video cards is up to 16 times faster at financial risk calculations than a modern processor, 20 times faster at granular studies and even up to 40 times faster at [email protected] operations used by Stanford University to do research on diseases.

    ATI have released a beta-version of the [email protected] client as a proof-of-concept, which you can use with any video card from the X1900 family. Though the software doesn't offer any benchmarking possibilities, it is quite clear from our own tests that the Radeon chips are far efficient at processing the calculations required for [email protected] then processors from either AMD or Intel. It can be safe to say that this was partially developed to coincide with the new Torrenza-platform developed by AMD, ATI's new owner.

    AMD are the first ones to take a step in the right direction and have already announced the "Torrenza" platform last year. This technology entails that the CPU socket in a multiprocessor motherboard can also hold other types of chips. This can mean that in the future on your mortherboard you have one socket that actually holds a CPU, while the other sockets can hold co-processors that help the overall performance of the system. A GPU is the most likely candidate, but the Torrenza platform will be able to hold any kind of chip.

    The technology which AMD is currently using for the Athlon 64 and Opteron processors is already well suited to for this task, meaning that AMD should be able to get Torrenza up and running in quite a short period of time. The most important factor is the HyperTransport bus that AMD implemented as a substitute for the conventional frontside bus. HyperTransport is an extremely fast interface, allowing for the efficient and rapid communication between two separate chips. In an Athlon 64 system the HyperTransport functions as nothing more than a substitute for the standard frontside bus and only really starts to show its potential when used in combination with multi-processor Opteron systems. The system used by Intel is based around one central point i.e. all the processors are connected through one or more busses to the chipset. In an Opteron multi-processor system all the processors are collectively connected through the use of separate HyperTransport connections. This means that all of the processors can communicate directly with each other without having to go through a central point and without having to involve all the processors in the system in doing so. This form of one to one communication is already being implemented to allow Opteron processors to communicate rapidly and efficiently with each other, but in the near future this technology will also allow processors and co-processors to work together, whilst keeping latencies as low as possible.

    Another advantage of the AMD architecture is the integrated memory-controller within the processor, as a result of which each socket is in direct connection with its own memory, something which will also be of great benefit to co-processors. The beauty of the HyperTransport architecture lies in the fact that different chips can make use of each others memory if so required: for example when a certain chips own memory proves to be insufficient when handling a certain task or when data has to be shared collectively. In traditional system architectures the memory is controlled by the central chipset, this causes additional delays because the data has to be run through the central chipset first, this can result in the processors having to wait in line to make use of the memory


    AMD's plans with Torrenza are ambitious to say the least and the large amount of support that AMD has received means that this technology has a realistic chance of success. Intel cannot afford to sit on its laurels and has already announced its own plans. The largest chip manufacturer in the world has placed its bets on two different horses; first Intel wants to release a new version of the PCI-Express bus, code named Geneseo, onto the market, which will be better suited for using expansion cards fitted with co-processors. Intel will also be developing its own counterpart to Torrenza.

    During the last Intel Developer Forum it was announced that Intel would be developing Geneseo together with IBM. This new standard should see the light of day somewhere in 2008 and will be the direct successor to the current PCI-Express 2.0 standard which will be released some time this year. Intel is currently keeping quiet regarding the technical details for Geneseo and the only statements that have been released are vague promises that co-processors on expansion cards will have higher speeds and lower latencies when communicating with the central processor and memory. Looking at the specifications that have been released, this new technology should allow for co-processors to function more efficiently: think of the possibility of being able to block certain memory sectors or making certain sectors of the memory virtual.


    Planning ahead, Intel is also working on CSI, a new chip-to-chip connection which bears an extremely strong resemblance to AMD's HyperTransport technology. In time CSI will replace the frontside bus of both the Intel Xeon and Intel Itanium. Just as with HyperTransport the processors on CSI will be connected collectively through fast point-to-point data busses with Intel promising transfer speeds of up to 6.4 Gigatransfers a second. Intel is also planning to crossover to an integrated memory controller, based on FB-DIMM, at around the same time that it will release its CSI platform. The first processor to make use of CSI will be the next generation of Itanium processors (codenamed Tukwilla) which is planned for release in 2008. The implementation of CSI for Intel's Xeon, and possibly its desktop processors, will happen in 2009.

    Just like Torrenza, Intel will license its CSI bus to manufactureres of co-processors, by which Intel will offer the exact same possibilities as AMD. Apart from the slightly faster speeds, which Intel has promised, there are very few differences between AMD's Torrenza and CSI. AMD's advantage will be that Torrenza is based on technology which is currently available and hence will be able to release its products far sooner than Intel.


    All the aforementioned developments will first be implemented in servers, but are definitely worth following to see how things progress. Its more than likely that we will see this technology becoming mainstream around 2010 with desktop systems with multiple sockets, allowing users to install dedicated co-processors to their hearts content. These co-processors will allow the performance of computers to increase exponentially for certain applications, just as has happened with GPU's and 3D applications. Intel and AMD appear to have an almost identical vision of the future, seeing as their respective Torrenza and CSI platforms are almost identical. If these plans will actually take off and which one of the two companies will be the most successful would be like reading the dregs in a teacup.

  • #2
    Re: Gone are the days when the CPU was the only central powerhouse

    Great info. Did you write this? ..or find it?


    • #3
      Re: Gone are the days when the CPU was the only central powerhouse

      Originally posted by SuperDudeBT View Post
      Great info. Did you write this? ..or find it?

      no no , i have to copy it cause i can't place any urls or photoattacments on the site.

      but.............. Luckyshot will,..... i hope taking care of.

      I have lots of new data but i cant post it .



      • #4
        Re: Gone are the days when the CPU was the only central powerhouse


        • #5
          Re: Gone are the days when the CPU was the only central powerhouse

          Boudreux got it. ;)

          Lucky Shot




          TeamSpeak 3 Server


          Twitter Feed