
We are not yet back from AMD's CTO Technology Summit held this week in California, but it did make a big impression on us. Here are some thoughts that we want to share with our readers as to the direction of AMD.
After recently attending AMD's CTO Technology Summit I came away with a much better understanding of its direction in the CPU market as well as the GPU market. While many of the specifics cannot be discussed in detail due to non-disclosure agreements in place, we can certainly share our thoughts and feelings about what has been presented to us. This editorial will also act a follow-up in some ways to the recently published State of the Silicon Union editorial.
Phil Hester, AMD CTO, gave a presentation that helped me understand AMD's direction since it acquired ATI. He was stern in his statement that, “Extending homogeneous cores is not the way to go.” He outlined his vision of the CPU and GPU coming together. He was also very direct to the point that the most important benchmark to AMD is customer experience rather than metrics which you know is something that is near and dear to our hearts here at HardOCP. A true desire of taking care of their customers' needs was expressed to be their mission. Issues such as excellent backwards compatibility with current software sets was described to be a must with all of their products while pushing forward with emerging technologies such as stream computing models and multi-core processors.
He also described the merging of CPUs and GPUs in detail. His vision sees AMD's GPU technology being totally integrated into the CPU, much like we saw the floating point processor integrated into our current CPUs. This obviously begs the enthusiast question, “What about my video cards?” Undoubtedly there are many of us that have no desire to have our main display controller integrated into our CPU as it simply does not jive with our hardware upgrade paths. While we will very likely see the “ATI” logo go away with the next-gen R700 technology, high-end discrete graphics is not falling off of AMD's road map. Phil Hester repeatedly explained that AMD is not moving away from the high end GPU business. The CPU/GPU merge model mentioned above simply does not cater to the high-end gaming market, but will likely be comparable to integrated sets of technology we see today. These models obviously appeal to the mobile markets and as AMD merges CPUs and GPUs, we expect to see them move towards a system-on-chip model where AMD will be supplying much of the silicon needs of mobile and business use systems. Don't think of these solutions as being poised to be weak though. AMD sees these configurations as being able to run HD video conferencing and handling multiple applications at the same time.
Interestingly enough, Bob Drebin, CTO of the graphics product group, brought up the point that ATI clearly saw the need of the merging of their GPU with a CPU. Bob explained how ATI was seeing the need for some type of sequential processing to be brought along side their deep parallel GPUs. Anand Shimpi asked Bob what exactly their plan was to deal with those needs had AMD not acquired ATI. While Bob responded, we never really got the answer to the question.
As we earlier posted, Barcelona technology was shown off in the form of Agena FX desktop processors running live demonstrations. Agena FX processors are basically the Barcelona server quad-core CPU that has been fitted with a non-registered DDR2 memory controller for desktop usage. High definition video encoding was shown being run while using a currently available application and all four cores of the Agena FX being utilized. Also shown was the new two processor “4x4” platform that is now truly a 4-core by 4-core system with the Agena FX. The same HD video encoding demos were run showing to take advantage of all 8 processor cores in the system. While not exactly a demonstration, AMD did allow us some hands on with newly produced 45nm process technology in the form of a 300mm wafer as well.
While I left the AMD CTO Technology Summit with a enthusiastic overall vision of where AMD wants to take their business, it does not change the fact that AMD currently has its back against the wall. The company is in a situation where it must execute flawlessly. Barcelona must make its “mid 2007” launch date into the server arena. Doing this will allow AMD to gain back some of the market share it has recently lost to Intel. I think that Barcelona will be a superior product compared to what Intel currently has in the space. Any Barcelona tardiness is going to push it out closer to Intel's upcoming Penryn, and that will put Barcelona on a much more level playing field with its competition. I am unsure which is going to be faster, but I do not think there will be a huge performance differential in the server market once Intel's Penryn makes it debut. Barcelona must push into the server space while it can exhibit dominance and leverage that performance benefit, and that firmly depends on AMD's execution being without error for the remainder of the year.
AMD's internal vision is clearly defined and it seems to be very much on track. If it is able to clear the hurdles immediately in front of it, it seems to be very much back in the position to innovate and lead the industry in design and user experience.
Discussion
Please join us in this HardForum thread dedicated to this editorial.