Intel Corp (INTC)

[chipguy]

IBM smacks IPF across the face with a leather glove

http://www-1.ibm.com/servers/eserver/pseries/hardware/system_perf.pdf

Very impressive results for POWER5, especially in throughput
oriented benchmarks. Why? SMT and shedloads of bandwidth.
Did I mention POWER5 has a lot of memory bandwidth?

Intel and friends will respond with Madison 9M in a month
or two but I think IBM will still hold an advantage in a lot
of important workloads until Montecito appears late next
year. How will this impact IPF uptake? That depends a lot
on pricing. How much IBM is willing to lose on hardware to
slow IPF's adoption rate? The big loser in all this? Scooter
McNealy. Big blue's mid-summer surprise and Intel et al's
inevitable response in a few months will pour a veritable
niagara of cold water on FujitSun's hot new SPARC 64 V.