Shakeup in the Graphics Community

Yesterday there were rumors flying all over the internet about an ATi/AMD merger and Nvidia then turning more attention on Intel chipsets. While this potential merger would help AMD in the mobile sector with a strong source of dedicate mobile graphical chipsets, this could have killed AMD’s desktop market. Just think of who makes the majority of AMD’s performance chipsets, and now imagine ruining that friendship. Things would have gone downhill quickly for AMD. Thankfully these were just rumors and there has been no swapping of sides, so all is right in the world, for now.

”My video card has Intel Inside”

Many people do not know this, but Intel is currently the largest manufacturer of graphical processors. Just think of how many computers have onboard video? Nearly all of those will feature an Intel onboard video solution, either the GMA950, 900, or oneof the previous revisions. While Intel does onboard video very well, they have never ventured into the dedicated graphics market. However that may change, more rumors are flying around the Internet about Intel moving in on this market. If you think about it though, this wouldn’t exactly be a bad move for Intel. Their current graphical processors are installed on most notebooks, a lot of budget motherboards, and a decent portion of server motherboards. With their entire processor line dropping down to the 65nm, Intel has a few of their fabs still tooled for 90nm and even 130nm so they definitely have the manufacturing capability. Once production stops on all 90nm processors then these fabs will start to focus on chipsets and flash memory. In regards to the 130nm fabs this would be a suitable use as 130nm is a rather large process to use for graphical processors, however 90nm would be perfect. While Nvidia and ATi are currently working the bugs out of their 80nm processes, Intel has had years of experience working at 90nm and would be able to get yields up to acceptable levels quickly.

Why?

Why would Intel want to barge into another market? Profit margin would be a good start. Currently AMD and Intel are in a rather interesting predicament. Intel is in the process of rolling out three new processors in 3 months and production might not satisfy demand. Meanwhile AMD has a good three months till they have a competitive processor so they are cutting prices on all their s939 processors by up to 48%. Intel also has a huge stock of old Netburst processors that it needs to sell off quickly before Core 2 Duo swarms the market and the desire for Netburst dies. So now you have two companies selling off processors at and in some cases below cost just to get warehouses cleared out, and this will continue for a few months. Now to add to this, Windows Vista will launch soon and in most cases a dedicated graphical would be very helpful. With this problem looking to ruin their earnings for the next quarter, Intel would most likely turn to other markets for bring in revenue. While I haven’t covered everything, I would consider now to be a prime time for Intel to enter the market, they have plenty of manufacturing capacity, plenty of computer engineers, all they need is the motivation. I believe this motivation will come in the next couple of months as Intel reports falling revenue and anxious stock holders start to complain.

While this is all just speculation, you have to admit that there could be a decent chance of Intel trying to push into this market. I must admit that I wouldn’t mind seeing Intel enter the dedicated graphics scene. If Intel could carry over its manufacturing processes from CPUs to GPUs then I can’t imagine what beasts Intel would unleash on the scene. Even if the first couple rounds of GPUs are complete duds, I am sure Intel would be able to quickly work out the bugs. Even if the performance is lower than Nvidia or ATi, with so many fabs available I am sure Intel would be able to undercut both companies and more competition means more benefits for the consumer.

ATi at 80nm

Early reports showed that ATi was having trouble with their 80nm process with terrible yields, but recent news has shown quite the opposite. The new 80nm core, RV570XT, will be introduced in the new round of X1900GT chips. In theory these chips using the 80nm process instead of the 110nm will be capable of clocking higher and running cooler, being a win-win situation for the enthusiast. Technically they should also be cheaper for ATi to manufacture as each die will be smaller, but if yields aren’t up to speed then the consumer may not see this benefit for a few months. These cards are set to release in early October which will place them right before the Vista launch. While these cards will be capable of supporting Vista, there are no reports of them being DX10 capable.

DirectX 10

Supposedly the last of the DirectX series, this new revision should contain lots of graphic goodies that will impress all gamers. The new revision will also make all DX9 cards last generation cards and clear the path for the new DX10 cards. Nvidia has been working hard to perfect their G80 core which will be DX10 capable and support Shader Model 4.0. ATi’s R600 core should also be DX10 and Shader Model 4.0 capable but that will be seen once these new cards hit the streets. While I couldn’t find any definite dates, I would expect the first round of cards to hit the scene late winter, most likely around Christmas time. I will provide more information on these cards as I find it and it becomes available.

It sure has been an interesting few days with AMD/ATi rumors, Intel rumors, and Nvidia rumors; it seems like everyone wants to start graphics rumors. In the coming months a lot of products should launch and once all the dust settles there should be a new wave of budget and mid-range parts for me to pick through. In particular, Conroe should be very exciting and the launch date for this chip keeps getting bumped around. The latest I heard, July 14th would be the launch date, however I have also heard July 27th, so we will see. Good times are coming ahead, if you are holding out for a new computer, just wait a little longer.

On a more personal note, I apologize for the very lengthy delay between my posting. For the last ten to fifteen days I have literally be slammed with school work with major projects in all my classes culminating on the same week. To say I’ve been tearing my hair out would be an understatement, it has been rather stressful around here. Hopefully for the next two weeks my load will lighten and I can attempt to maintain a regular posting schedule. My autocascade project should make some headway this weekend and I will write a bit more about that. Bear with me, school sucks and isn’t getting any better.

No Comments Yet

You can be the first to comment!

Speak Your Mind

  • Comment Policy: No flaming unless directed at the author for good reason. All other flaming will be turned into a comment about ponies and rollerskates.