This statement could have only have arisen out of sheer
reckless ignorance or outright lying.
This is a constant core count system level performance
comparison of EV7 and Montecito systems, i.e. two Alpha
chips for every IPF chip.
At 64 cores the IPF system outperformed the EV7 system
by 63% and cost 41% less.
At 8 cores the IPF system outperformed the EV7 system
by 79% and cost 72% less.
At 4 cores the IPF system outperformed the EV7 system
by 71% and cost 83% less.
The performance/price advantage of IPF over Alpha is 2.73x
at 64 cores, 6.39x at 8 cores, and 10.0x at 4 cores. IPF sure
looks like a huge win over Alpha (as it should be) unlike your
specious claim.
Thanks for proving my point.It has taken IPF 4 more years to beat Alpha systems by 60-70% in 3 tests : SpecFP, SpecInt and TPC.
How much did x86 improve in the same timeframe ? Or Power ?
300% ? 400% ?
What about other database and transactional tests ? I'm pretty sure IPF doesn't beat Alpha core/core in some , but since HP never bothered to release a full spectrum of Alpha EV7 benchmarks, we're only left to guess.
Why weren't Alpha servers tested ? Possibly because they trashed the Integrity supposed to replace them ?
Thirdly , why bring cost in the equation ? It's completely meaningless.
Market Data 
Markets 
