[Info-vax] RealWorldTech on Poulson

[Neil Rieck]

On Jul 1, 10:56 pm, JF Mezei <jfmezei.spam... at vaxination.ca> wrote:

[...snip...]

> Are there any Tukwila benchmarks that show it leads over Sparc ?

No, but abandoning a technology which doesn't work (VLIW) for one that
does (OOE to start; but a slow shift back to SuperScalar) might signal
the competition that engineers are back in control of this product.

> The out of order execution will likely be restricted to the instruction
> blocks that the compiler has determined can be pipelined. This may be
> more marketing hype than a true change in architecture.

> Consider that a true change in architecture to really make it out of
> order, would likely require changes to all compilers and recompiling of
> software with new compilers, and it would also require that the new
> chips support the old executables with the EPIC mentality. Me thinks
> this would be way too big of a change to be acceptable.
>
> Intel likely needs to hype up IA64 now that it has been declared dead by
> Oracle and it will try to find any way to hype it.

I agree. Many industry people responsible for hyping VLIW/EPIC are no
longer working which means there will be no personal embarrassment
associated with an about-face. Now is the time for a young Intel "VP
of engineering" to make a statement similar to this:

"starting in 1989, HP engineers went down the wrong path when they
developed VLIW/EPIC technology. They thought that SuperScalar RISC
would be very difficult to develop but companies like DEC, IBM and Sun
proved them wrong"

In the world of politics (fear, uncertainty, denial), admitting error
is ridiculed (you are a flip-flopper!). In the world of engineering,
the majority of technical people will applaud then thank you for
putting the industry back in the direction of progress. Investors like
this too because they are convinced that everyone lies so having
anyone stand up to admit previous wrong doing (along with a plan to
future increased profitability) is too rare to ignore.

NSR