Nick M V Salmon wrote:

"Martin" <ml_news@ddnospamddml1dd.co.uk.dd> wrote

[...]

My only meagre feeble excuse is that I'd forgotton what the Israeli
 group's P-M much improved rework was based on... (:-((

An excellent design that hits a very good compromise for
efficiency. So good that it is beating the P4 in some respects!

If you edit that to "all" instead of "some" then I'll agree with you
- amazing IPC, better than Athlon (!), very low Vcore and power
consumption - a real winner..!

However, there's a fundamental difference in philosophy for keeping
Moore's Rule going for a while longer. The P-M relies on continued
improvements in fab to allow the chip to be manufactured faster. Whereas
the P4 has deeper pipelines with less delay per stage to go chasing the
marketing hyped higher GHz clock speeds (at the expense of higher
complexity and reduced efficiency, hence the hyperthreading complexity
bodge).

It's a good question as to whether the P-M's improved fab gamble will
pay earlier performance dividends than the more clunky P4 clock speeds.
Or which will be more cost efficient.

Meanwhile, AMD's Opterons are cleverly named (marketing) and technically
look good also.

There have been quite a few calls for the Israeli P-M to be released
for the desktop too but I doubt it will happen whilst Intel continue
to succeed in selling of partially dud P4s as 'Celeron'...

Yep. Its not whats good for the customer or Mankind, but what fastest
marketing trick can be pulled to make life easier for the company managers.


Regards,
Martin