The weekly PCSTATS Newsletter has been running a column for the last couple weeks on those two eternal foes - AMD and Intel. The full editorial article "AMD vs. Intel - Will the Best CPU Please Stand Up" is now posted, and we've been getting some great feedback from you on it.

Some of you agree with our opinions, and some of you disagree entirely. That's great! At PCSTATS we're not so high in the clouds to think that our word is the final one... With that in mind, Aashish K. writes in with his comments on the AMD vs. Intel topic. Let us know what you think about what he has to say.

"I am ashamed by the quality of your editorial. I understand it provides your view of the current state of the computer industry, but you certainly also make forward looking statements within the editorial as well. Therefore, I cannot consider it as a simple editorial, but a recommendation for which company's processor is better at a given point. In fact, your recommendation is given near the conclusion of your statement. Whether or not this was your intention is pointless, because that is how it is perceived.

Now since you are giving a recommendation and make forward-looking statements, you should give a complete picture of the PC industry. You fail to mention any substantial information about Conroe, K8-L, mobile processors, or provide concrete evidence behind your points. I believe that most users read your article to get some insight into a purchase a system. There is insufficient information about both processors to help people in their decision-making. As a respected member of the press, it is your duty to report information.

For example, it is known that AMD will substantially increase the quality of prefetching in the K8-L, which will be an AM2 socket part. A better prefetcher will better mask the latencies of memory and thus reduce the importance of DDR2 memory with tight timings. Furthermore, your work with Intel systems proves that in real-world memory intensive benchmarks, such as games, memory with better timings does not produce a significant increase in performance given that they are TWICE the price of normal memory modules from CORSAIR, CRUCIAL, OCZ, etc. Intel CPUs are known to have better prefetchers than AMD CPUs and so a similar result can be expected from K8-L. As a result, an individual using an AM2 platform system may be better off saving extra $200, that they would have spent on better memory, and instead save it toward the purchase of a K8-L next year or a better GPU.

Another example of where you come up short is on the mobile computing front. As we all know the computer industry is moving toward mobility. As evidenced by the tremendous growth in the mobile sector, you should provide insight into the offerings of Intel and AMD there as well. There was one mention of a Core Duo processor and none about AMD Turion processors. So someone looking to buy a laptop receives no meaningful information from your article is probably even more confused that before. Many of Intel's dual processor sales are from the Core Duo and we know that this CPU is much more competitive than Pentium Ds.

In reference to Intel's marketing, I believe your assertion that Intel has sold more dual core processors because of its marketing power does not even begin to paint the picture. You neglect to mention that Intel produces its chip on a 300mm wafer process. As a result of this and their financial size, Intel can sell dual core processors much cheaper than AMD. Furthermore, AMD only uses 200mm wafers right now and does not have as much capacity (Next year, AMD will have a 65nm 300mm process just like Intel, so things may be different). So, they face a very short supply of their dual core parts. There is plenty of evidence of this. For example, X2 prices have barely dropped since their introduction. AMD has an internal memo stating that they will focus on producing lower cache size cores to increase yields on the AM2 platform to increase supply.

As we all know, computing is a science of tradeoffs. Nothing is designed without making significant tradeoffs and consumers do not purchase products without making tradeoffs as well. You have no discussion of upgrade paths. Does it make sense to buy a really cheap Pentium D with the hope of upgrading to a Core Duo 2, OR building a Core Duo (Yonah) system with Intel's 975X chipset without a clear upgrade path, OR is it more practical to buy a cheap Athlon64 AM2 to upgrade to a K8-L, OR is it more practical to simple buy an Athlon64 X2 939 for the fast DDR timings without an upgrade path? Your editorial does not even begin to address this issue. I get the feeling that had your article been written 6 months ago, it would have been a better fit."