The Linux Extremist

AMD has long been one of my favourite companies. As soon as their processors became available, I was one of their early adopters; I was sick of Intel, which in my mind was equivalent to Microsoft in their presumption of a monopoly. AMD seemed to go against the grain; they were sticking it to Intel and producing processors that were on a par or better than Intel for less cost.

When I was recently building a new desktop machine, there was no question in my mind about how a balance of speed, adaptablity and cost was going to be achieved, I was going to have to use an AMD processor. With an AMD Sempron 3100+ and Ubuntu Linux, I was able to build a very fast, cheap machine that is going to give me at least a few years of computing joy.

On the other hand, ATI is definitely in the enemy camp. Their Linux support in my experience, though reportedly improving, is very poor. I’ve had no end of trouble with their proprietary fglrx driver, particularly using it with the ATI Mobility Radeon 9200 chipset on my HP Compaq nx7000. Every time I’ve utilised fglrx, it’s been say a prayer and sing a hymn time; I was lucky on my Savrow Katana K90 - fortunately the ATI card on that machine is modern enough for fglrx’s liking, for now, anyway.

The idea that companies which such divergent approaches to Linux are merging, therefore, does not fill me with hope. My first reaction upon hearing of the ATI / AMD merger was “Why?” Why indeed? Who needs it more, ATI or AMD?

Whenever a merger of this type happens, we usually hear the same management twaddle about “leveraging synergies across the range of brands”. I suggest in this case, it’s ATI that needs it more; the fact that AMD is buying them, not the other way around hints at this. We have not yet seen a really good answer from ATI to Nvidia’s range of high-end graphics cards, such as the 7600. Speaking as a system builder, I have had far better results out of using Nvidia cards than ATI; speaking specifically as a Linux system builder, I can attest to the fact the Nvidia’s support for Linux is far, far better than ATI’s. Things have gotten so bad with ATI that any system I buy or build in the future will be devoid of their cards so far as I can help it.

Can AMD turn this around? They have a good base from which to start: as for Linux, AMD’s support is rock solid. Obviously they need to find whoever is writing Linux drivers for ATI and tell them to stop writing code that makes people want to fling their laptops away like a discus. They also need to ensure the people at ATI are not adhering to a cynical philosophy which fglrx’s lack of legacy support implies; namely, that people should be forced to upgrade their hardware to utilise the latest and greatest drivers. AMD is also going to have to bring their spirit of innovation and value to ATI’s rather stodgy range. In short, AMD’s culture is going to have to prevail over ATI’s.

Creating a unified culture for two merged companies is invariably tricky; at first, there is always some difficulty in getting two different corporate personalities to play nicely together. Daimler Chrysler found it took years to get the balance right; in the end, assertion of Daimler’s personality was the best way. Hopefully AMD won’t take as long to realise they are going to have to assert themselves once they’ve stormed the corridors of ATI.

If AMD is able to do this, then all may yet be well. The fglrx driver may no longer be nightmare, and AMD’s values of reliability and value may become ATI’s driving principles too. For the moment, rather like when one installs fglrx, it’s a matter of hoping, praying and waiting for the reboot.

Trackback URI | Comments RSS