>Typically an upgrade cycle may be triggered by the purchaser's desire to take advantage of all these "peripheral" considerations. For example, the cost of the labor together with the parts may make it more economically feasible to order a new machine. You may be running a PII 200 clocked to 464Mhz, but you are still running a 3000rpm drive, a new machine will come with a 7200, 10k, or even 15k drive that is 5 times the size PLUS a new quieter fan, which costs $10 in parts BUT $35-75 in labor to replace, etc, etc.<
Agree in general, though not on the specific. Most business PCs I was buying a few years ago were typically coming with 5600RPM IDE drives. Today, they're mostly at 7200. 10K and 15K drives are mostly in use in servers. The advance in speed of disk drives is incredibly slow compared to the advance in processor speeds.
The reason I believe this stuff is so important to Intel, is that if I perceive the bottlenecks to be someplace other than the processor, that's where I'll put my money. I'll save a bit on the processor to get a faster disk, or hold off on PC upgrades altogether so I can spend more money upgrading the network. Some people here would say "you should do all those things." They just don't live in the real world of limited budgets (curently SHRINKING budgets).
Technology spending is always driven by where you can get the biggest bang for the buck. For years, Intel could count on the processor being the BEST place to gain the most performance for the least money. Today, I no longer see that as the case. Distributed applications mean the network is often more important. The terabytes of data we've collected put a premium on storage technologies. The CPU is not the only game in town, and often not the most important one, at least for now.
I don't want to be lumped in with those who are saying "good enough" and believing they'll never need to upgrade again. I've been around long enough to recognize that at some point I will need to upgrade my CPUs too. But right now there are generally more important things to upgrade. The money that it would take to upgrade PCs "just in case" can also be used for other things, ones that will yield real performance benefits today. So that's where I recommend the money be spent.
Otellini said it quite well at IDF, quoting Charles Kettering, he noted that "I believe business will come back when we get some products that people want to buy." And I read his speech as indicating clearly that he knows "more MIPS" is no longer sufficient to make people want the product. (Unfortuantely, two of three demos went right back to CPU performance, only the one about low power capabilities really spoke to me.)
Many here seem to think that performance on esoteric benchmarks are what's going to matter. Your boss seem to know better. You should listen to him.
mg |
