![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
I'm worried because AMD's survival is vital to the consumer, and for the first time I'm wondering whether they'll survive. Now some of their most recent horrifying quarterly results come from a one-time write-off they took because they over-paid for ATI.
But seriously folks, whether you're an AMD or an Intel enthusiast.. It will be better for all of us if AMD not only survives, but prospers. If AMD doesn't survive, then the prices for CPUs will most certainly go up, and probably by an appreciable amount. If AMD doesn't survive, there's a risk of a lapse in innovation. If it weren't for AMD, we probably wouldn't have x86-64 or hardware buffer overflow protection, for example, and whether or not we had multi-core by this point would be highly debatable.
Regardless of the industry or a particular company, when there is no competition, it leads to complacency and arrogance. In the case of Intel, when they had no competition, they did things like refuse (initially anyway) to fix a fundamental flaw in their processor unless you proved you used your processor for floating point (wtf? Who doesn't use their CPU for floating point?) The above isn't a criticism for them having errata in their CPU, it's a criticism for trying to get out of doing the right thing and replacing the part, no questions asked (this is ultimately what they did, but uh, why the initial hesitation? They were making a couple of billion a quarter in profits.) Look at Nvidia, who only a month ago was asking you to pay $650 for their top of the line graphics board. With competition they've already dropped the price almost 25 percent in a month.
Listen fanboys of both sides, I'm not knocking Intel. They have done a helluva job lately, and have been firing on all cylinders, pretty much since Banias. I remember seeing reviews of Dothan back in 2004. At that time, Dothan was not only in striking distance of the K8, but in some cases exceeding performance of the single core K8. I said to myself, "oh boy.. AMD had better have something up its sleeve because the onslaught is coming." It seemed clear to me at the time that the only thing keeping AMD in the game was the memory controller on the CPU. Fast forward to 2008, four years later, and we've seen only tweaks to K8 single-core performance.
Now there has been some innovation, to be sure. Multi-core was an amazing innovation, but I think it actually clouded the issue for a while. People got so excited about multi-core; sales were brisk, people were happy. But there were dark clouds on the horizon, and in 2006, the storm named Conroe arrived in force.
There was also innovation in terms of virtualization enhancements as well as floating-point enhancements.
Since then, it's been what I would call a good-old-fashioned-ass-whoopin'. We're now two years into the Core2 era, and Intel's next generation is about to be released which looks better, stronger, and faster. From the early reports, Deneb looks like a step in the right direction and would have been competitive with Penryn. The problem is it AMD needed to aim beyond Penryn. What happened AMD? Were your design teams that bad? If so, then why couldn't you just hired the Montalvo design team or even looked across town in Austin at Centaur?
It's not all gloom and doom. ATI has finally gotten their ducks in a row and 4850 and 4870 boards have become the "new 'in' thing." This certainly will help. But AMD needs to be in development of a core that is IPC competitive with whatever's coming after Nehalem. They won't last long bleeding they way they are, and if they do in fact go under, it will be bad for all of us...
But seriously folks, whether you're an AMD or an Intel enthusiast.. It will be better for all of us if AMD not only survives, but prospers. If AMD doesn't survive, then the prices for CPUs will most certainly go up, and probably by an appreciable amount. If AMD doesn't survive, there's a risk of a lapse in innovation. If it weren't for AMD, we probably wouldn't have x86-64 or hardware buffer overflow protection, for example, and whether or not we had multi-core by this point would be highly debatable.
Regardless of the industry or a particular company, when there is no competition, it leads to complacency and arrogance. In the case of Intel, when they had no competition, they did things like refuse (initially anyway) to fix a fundamental flaw in their processor unless you proved you used your processor for floating point (wtf? Who doesn't use their CPU for floating point?) The above isn't a criticism for them having errata in their CPU, it's a criticism for trying to get out of doing the right thing and replacing the part, no questions asked (this is ultimately what they did, but uh, why the initial hesitation? They were making a couple of billion a quarter in profits.) Look at Nvidia, who only a month ago was asking you to pay $650 for their top of the line graphics board. With competition they've already dropped the price almost 25 percent in a month.
Listen fanboys of both sides, I'm not knocking Intel. They have done a helluva job lately, and have been firing on all cylinders, pretty much since Banias. I remember seeing reviews of Dothan back in 2004. At that time, Dothan was not only in striking distance of the K8, but in some cases exceeding performance of the single core K8. I said to myself, "oh boy.. AMD had better have something up its sleeve because the onslaught is coming." It seemed clear to me at the time that the only thing keeping AMD in the game was the memory controller on the CPU. Fast forward to 2008, four years later, and we've seen only tweaks to K8 single-core performance.
Now there has been some innovation, to be sure. Multi-core was an amazing innovation, but I think it actually clouded the issue for a while. People got so excited about multi-core; sales were brisk, people were happy. But there were dark clouds on the horizon, and in 2006, the storm named Conroe arrived in force.
There was also innovation in terms of virtualization enhancements as well as floating-point enhancements.
Since then, it's been what I would call a good-old-fashioned-ass-whoopin'. We're now two years into the Core2 era, and Intel's next generation is about to be released which looks better, stronger, and faster. From the early reports, Deneb looks like a step in the right direction and would have been competitive with Penryn. The problem is it AMD needed to aim beyond Penryn. What happened AMD? Were your design teams that bad? If so, then why couldn't you just hired the Montalvo design team or even looked across town in Austin at Centaur?
It's not all gloom and doom. ATI has finally gotten their ducks in a row and 4850 and 4870 boards have become the "new 'in' thing." This certainly will help. But AMD needs to be in development of a core that is IPC competitive with whatever's coming after Nehalem. They won't last long bleeding they way they are, and if they do in fact go under, it will be bad for all of us...
