>I wish we'd junked MIPS when workload dependency first reared its ugly head.
IBM used to stay away from MIPS. I took one of Joe Major's first courses in Capacity Planning (May 1982), and he used CPU quanta. Then, around the era of the ES/9000 series, they published a comparison document where 31.5 was the stated capacity of a 3090-200E (They didn't use any units), then when the 9672's came out they were blatent about using MIPS. The problem is that people need/want a comparison. How do you know what to upgrade to, if you don't know what you've got, what you need, and what's available? I wrote an article about it in December 2004: "Don't be Mislead by MIPS": http://tinyurl.com/7rogu But, we have senior management committing to gobs of expenditure based on something no better than a WAG. LSPR figures are also suspect, now. LSPR started out as a PCM-bashing tool, then it got reasonable. But, now they have created new workloads that they never calibrated on older technology, they have not tested all sizes, and the bigger ones are 'straight-lined', which is specious at best. And, the worst problem is that IBM no longer gives performance guarantees. So, if you buy it, you broke it. -teD Me? A skeptic? I trust you have proof! ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
