VIA Interview - November 2001

The last VIA interview we conducted was right after the release of the VIA P4X266 chipset. On the eve of our review of VIA's P4X266A chipset we've got another interview with Richard Brown, VIA's Director of Marketing. As usual, VIA's responses are indented:

1) What are VIA's thoughts for the direction AMD is headed with their recently announced Hammer architecture? The inclusion of a North Bridge and Memory controller on the CPU itself takes a significant chunk out of the tasks that were conventionally reserved for chipsets. Do you think that there is potential in what AMD is doing with Hammer? How would it affect the role of VIA as a chipset manufacturer if more processor manufacturers went down this route?

We think that the AMD Hammer processor is a great architecture for a CPU-Memory cluster in a multiprocessor server or workstation. It is a significant enhancement of the Athlon S2K bus and overcomes the limitations of that bus architecture.

We are strongly committed to fully supporting the AMD Hammer architecture, and believe that there are significant opportunities for VIA to add value as a chipset manufacturer to this architecture, for example through the development of an AGP controller bridging chip and continuous enhancements of the South Bridge. Particularly for desktop and potential future mobile platforms, these features are just as critical as the DRAM controller.

I think it's too early to speculate whether other processor manufacturers will follow AMD's route.

2) With Hammer, a significant portion of the real world performance that a chipset manufacturer has control over will have been removed. For example, a revision such as the KT266A would never have been possible had a processor with an on-die memory controller been present. What plans does VIA have to differentiate (on a hardware performance basis) their Hammer solution(s) from others that may otherwise perform identically? Can we expect to see an increased focus on AGP controller performance for example?

As I indicated in my answer to the previous question, the AMD Hammer processor with a local DRAM controller will definitely deliver significant performance benefits in a multiprocessor system.

But if you look at the history of the PC industry, DRAM speed enhancements occur more times within a given CPU generation than CPU interface enhancements. Therefore, for a CPU-Memory cluster server multiprocessor system, you will need to make a trade-off between the memory size/locality and the memory speed. For the Hammer architecture, such a trade off will definitely be worth it.

For high performance desktops, on the other hand, memory speed and access latency from components other than the CPU are also critical to the performance of the system - just as the VIA Apollo KT266A has demonstrated. A Hammer CPU with local DRAM may not be the best choice for this type of application. With the AGP card for example, it would probably be necessary to increase its onboard memory in order to reduce the dependence on the Hammer CPU's HT bus. For an integrated SMA chipset, it would be unavoidable for the chipset to have a DRAM interface that could serve as a display frame buffer or system memory.

So the way we see things, there will be no single "one size fits all" chipset solution for the Hammer processor. There are a lot of factors that will need to be carefully considered in developing chipsets for each particular market segment. One thing's for sure, however: VIA will have the first cost effective and high performance volume Hammer chipset and will be the validation partner with AMD. VIA will continue to provide the highest speed DRAM controller as a building block of the VIA VMAP architecture. We will provide multiple levels of chipset integration and cost/performance to enable Hammer for different market segments.

3) What will Hammer's integration of the North/Host Bridge and Memory Controller into its die do for motherboards and chipsets? Will there be a significant cost reduction in a VIA Hammer chipset as opposed to the VIA KT266A for example? If so, can you speculate as to what percent of a drop we could expect? What about reducing overall motherboard costs, any speculation there? Or do you believe that the opposite could occur, that chipsets would end up being more expensive and if so, how?

It's still a little early to give details on this. The critical figure to look at is the overall cost of the platform, including the motherboard and processor, rather than the cost of each individual component. I am sure that this will be very competitive.

4) Although considerably late, NVIDIA's nForce chipset is finally appearing in motherboards that should be available soon. One of the strengths of nForce is its powerful APU which arguably raises the bar for what end users should expect from "on-board audio." What plans does VIA have to remain competitive in the realm of on-chip audio processors? NVIDIA licensed their very powerful DSP from Parthus, would VIA consider a similar strategy or would it be something that would be developed in house?

We are very proud that VIA's audio technology has been completely developed in house, and are continuing to enhance this. In fact, we acquired some very strong consumer electronics audio capabilities when we purchased IC Ensemble over a year ago, and are working to integrate some of this technology into future chipsets.

There are two key issues that we look at in the area of audio. The first is striking the right balance between quality, features, and cost; the second one is making sure that our solutions meet the real usage needs of the PC user and enhance the user's overall experience.

nVidia's MCP is, for example, an excellent solution for the X-Box and its intended market segment: namely consumer electronics. Multiple DSP's are certainly necessary to take the load off the device's relatively low speed CPU and to compress multi-channel audio on-the-fly into AC3 format so that it can be hooked up to an AC3-capable receiver through S/PDIF. Such an approach is certainly very advanced in terms of technology, but it is more suited to a living room scenario where audio output will probably be delivered through some kind of home audio center.

Most PC users, on the other hand, do not generally have an AC3/DTS receiver sitting next their system. Instead, they normally connect their PC to a self-powered multi-channel speaker system through analog cables or through a direct digital link. VIA is therefore focusing on leading the industry by providing a hardware accelerated 6-channel audio solution for the former - even with the convenience of single chip codec with 6-channel output. We are also working with other companies in the industry to provide the digital link with the best cost-performance. S/PDIF path is only one of them. The goal for us is to create a balanced and high quality audio solution for PC systems. To keep on adding technologies to one subsystem without considering the overall PC environment is not our objective.

5) Have there been any significant advancements in the pending lawsuit against Intel? I realize that there is not much you can say but is a settlement close or can we expect this tension to continue?

Unfortunately, I am not allowed to comment on this matter.