Is IT ready for AMD? (continued)
Performance
This is an interesting subject that has been the topic of more than a few debates on the Internet. In general, until AMD’s recent announcement of their newest Athlon revision with full speed cache, the P3 with PC-800 RDRAM has been slightly faster than an equivalent Athlon; however, when matched with a non-RDRAM system or with inferior PC-600 RDRAM, the system have been basically similar in performance. (see www.tomshardware.com, www.sharkyextreme.com, or www.anandtech.com if you want to see the nitty-gritty on the performance of Athlon vs. P3.)
There is one exception to this. And that is when comparing Athlon based systems doing heavy floating point calculations to equivalent P3 systems. In most of the heavy scientific tests, the Athlon systems are far superior to the equivalent P3s. This is so significant that I would urge IT professionals who support Engineering type applications to even more strongly consider an Athlon based system.
Lastly, with the advent of the new Thunderbird based Athlon processors and the move to on chip, full speed cache, the Athlon is posed to surpass equivalently clocked Pentium 3 processors in the majority of standard industry benchmarks with the Duron projected to be only approximately 10% slower at even lower price points.
Conclusion
It is probably fairly clear by this time that I think that IT really should consider an Athlon based system for corporate use. A few days before writing this, I actually went through this entire evaluation and decided to proceed with the purchase of three Athlon based systems from Midwest Micro in Ohio (www.mwmicro.com).
I feel that Athlon based systems have been on the market long enough to have proven their reliability. And, when you consider the excellent price, good availability and at least comparable performance, it seems like the perfect opportunity to try out the Athlon in my corporate environment.
So, where does that leave us? If you agree with my thought process, I suggest you also bring a few Athlon systems in house. I will report back my finding to you and hope that you will e-mail me (eric@IThell.com) with your findings so that I can add them to my follow-up article. Hopefully, in a month or so we will have enough anecdotal evidence to go with our reasonable thought processes and arrive at a good solution for our future desktop platform.