Before I commit the entire month of August to headphone rambling, here's a few things in the tech world I sort of skipped over since the last relevant post.
The biggest I think is AMD ditching the ATI name. A lot of computer websites and myself included mostly referred to ATI products by AMD anyway, but occasionally "ATI" would slip back into consciousness and find its way into our writings whenever the thought arose about what actually produces these products. Well AMD hopes to dispel any such confusion in the future by erasing any question of the source of their graphics solutions, emphasizing the dissolving of any subdivisions into a unified, singular company. To some, this is a startling reminder of the extremity of the event that took place now four years ago, one that carries with it some pretty heavy significance in the history of computing, and in the minds of those who follow the industry. Those folks might have foreseen this eventuality from the outset, but the sudden immediacy of its arrival is no less sobering especially for fans of ATI in years gone by. It seems hard to imagine a Radeon without an ATI, but we must be quick to remind ourselves that it isn't the name that made the Radeon, but the company, the engineers and innovators within that kept an underdog alive, and even gave their opposition a bloody nose from time to time. Now more than ever they're on top of their game, and with NVIDIA struggling to flesh out their product line, their clockwork execution and unwavering onslaught have put them in a position few would have thought possible after the stumbling first couple years following their acquisition.
On the subject of AMD, recently some new information about their future architectures, Bulldozer and Bobcat, was disclosed. Not one to let them linger in the spotlight too long, Intel allowed the covers to be lifted on Sandy Bridge performance, just three days later. What's more encouraging to the tech observer? Some bits and pieces of architectural whitepapers, or some real-world results from its competition? You may not know how Sandy Bridge gets there, but at least now you know it does. Intel has maintained their ante-upping stride, whilst its closest rival on the bar graphs lags two generations back. However good Bulldozer might be, I doubt many think it's going to make that kind of a jump, and it doesn't help that throughout the article it's compared to current-generation Intel chips. AMD has macro-ops now. AMD has a 4-issue engine. AMD has something akin to hyper-threading. Instead of pushing the envelope with innovative new features, I'm seeing a lot of catch up, and unless all the lower-level stuff is really special, I'm not expecting this to do much better than Nehalem. I've seen enough from the first two Phenoms to know I have to keep my expectations in check.
Also on AMD, there've been some information leaks on the upcoming, presumably named HD 6870 graphics card, and this time they're looking a lot more like Intel in the CPU business: skimpy on architectural details, but with promising performance numbers, and a generation ahead of its competition. If ever there was an embarrassment to the GTX 480, it's those Crysis and Heaven results. The 6000 series is built on the same manufacturing process, due out less than a year after the heavily delayed Fermi card, not only matching its 15% average advantage over current-gen 5000 series flagships but raising it another 20%, and I doubt it runs as hot or ends up quite as large. It may be barebones feature-wise compared to NVIDIA, but all that extra engineering to make Fermi the GPGPU monster haven't really panned out for them, and it's starting to look like CUDA is a dud in the market, and with it the 400 series. AMD had their priorities straight, they had level-headed goals and they met them. Like a well-oiled machine, the 5000 series rolled out top-to-bottom in less than a year (really half a year), while NVIDIA is still trying to finish their lineup. When NVIDIA finally gets their budget cards out, AMD will be on their second-generation DX11, showing little slowdown from the 28nm delays, and still able to milk out a speed increase. I must say I've been extremely impressed with AMD's execution the last few generations, even if the technology behind them wasn't all that interesting. NVIDIA had their flaws, but they had slick and innovative designs, resembling more CPU than GPU, and had the raw performance edge and features to back it up. Unfortunately their ideals have been drifting further and further away from market realities, and they've been betting more on hope and prayers than concrete real-world trends. I don't mind a company that likes to dream, but one that doesn't seem capable of realizing those dreams is a company with their head in the clouds. It's time to get your feet back on the ground NVIDIA, because you ain't no Disney.
AMD's future looks bright, and even if Bulldozer doesn't set new records, their graphics business will be bringing home the bread, especially in upcoming heterogeneous designs, where Intel is already showing some muscle. If graphics in games don't get much more demanding, eventually GPU-on-CPU solutions might be good enough even for gamers, since it looks like competition could be pretty fierce, which means some pretty big performance jumps in a short amount of time. The next generation of consoles might help to spur on discreet GPUs, and that will be especially important to NVIDIA, as even their Tegra business is slow to start. Meanwhile AMD will continue to grow their cash reserves, and their CPU designs could get a lot more interesting going forward. It took this long for them to rework key architectural elements tracing back to the K7, but now that the ball's rolling, I expect the departures to keep growing. AMD made some seriously risky moves in recent years, but the result may be a more agile company with more IP and more innovation with a better competitive stance in their underdog position. If only I had a crystal ball...
