On MovieTome: The 10 worst movies of 2009 so far!
BNET Business Network:
BNET
TechRepublic
ZDNet
TalkBack 5 of 6:
Next »
« Previous
Good ...
Although Sun is faced with some financial troubles, it hasn't
given up and continues R/D to teh best of their abilities.

The industry was expecting that Itanium will take over the
entire world by the ... horns and exterminate all the pesky
RISCs.

It DIDN'T HAPPEN and it WONT. Itanium improves VERY
SLOWLY, to the point it has made RISC system vedors --
even HP -- to keep extending the life and improving on
their own RISC microprocessors.

Unfortunately, athough the Itanium 2 has started showing
promise as a serious 64-bit platform, Intel has started
running out of commitment and desire to push hard to
improve it. Itanium NOW looks good. If Intel slows down
the improvement efforts other processors soon will be
much better than Itanium.

Itanium is really a RISC microprocessor with EPIC and other
tricks that go with a RISC/EPIC. S/w tools -- COMPILERS,
LIBRARIES and other s/w -- is still NOT at the level to
ultilize fully the capacities of the chip. Everyone knows
HOW MUCH effort it takes to improve the compilers,
especially for this non-standard EPIC style.

Another shortcomming of the Itanium is that it is a fast and
FAT processor: it may run a SINGLE THREAD quickly, but
the Itanium would thrash under heavy multi-tasking
workloads. The amount of state cached within it per single
thread is what saves it the time to go outside the chip to
get/put data. A heavy interrupt storm (eg, network packets)
or task switching due to fine grain multithreded
computation will most lilkely reverse all the nice results
seen so far.

Thus as is, Itanium is good for coarse grain multiprocessor
workloads. BUT: Intel never BOTHERED to design the
Itanium inteface for SMP with # of PE > 4. Obviously, they
never expected it to go beyond the small SMP scale
(middle-tier server for business apps).

To me the future of Itanium is uncertain: it is geared to
compete with platforms based on its archaic predecessor
x86 familiy and not equiped to directly compete with the
high-end RISC/UNIX SMPs. This is strange to me.

I think Intel was hopping too idealistically to get excellent
results. But after a few hard learned reality checks it
realized that it takes more than ambition and confidence to
push the envelop effectively. I would say that the older
RISC/UNIX firms have already paid their dues to the R/D
altar. I will not be surprised if the RISC processors, which
everyone was expecting to disappear like the dinosaurs,
will make a revival to lead the direction.

This is likely with the Power 5 and maybe with UltraSparc
IV, V, etc.

Thank you Intel / MS for scaring the RISC/UNIX community
into action to improve their solutions.....

-m
Posted by: michael-t   Posted on: 02/10/04 You are currently: a Guest | Members login | Terms of Use

Alert moderator to an offensive message

Subscribe to this discussion via Email or RSS

Strategy will fail...  Mike Cox | 02/10/04
Strategy will fail...  JohnW_z | 02/10/04
Hey!  bhanes@... | 02/10/04
Windows is failing  BillMerchant | 02/11/04
Good ...  michael-t | 02/10/04
Good post  jeffs_z | 02/10/04

What do you think?

SponsoredWhite Papers, Webcasts, and Downloads

Enterprise Applications

  • Check out some of the easiest and most powerful ways to boost productivity while saving money on your application infrastructure. See ZDNet's comprehensive Enterprise Application resource center, now!
  • New Online Dashboard
  • Read about top issues IT decision-makers face every day, plus get cost effective solutions to real life IT problems. Oracle Topline