Replies to post #28727 on Advanced Micro Devices Inc (AMD)

[BUGGI1000]

@WMBW
"
...
So while 100% performance increases may be possible on certain sequences of code, it will be far from the norm. A more realistic range may be -5% to 20%.
...
"

I wouldn't agree here. Why provide a 64BIT version which
is slowlier than the 32BIT one? If thats the case, I could
switch to the 32BIT application and have the 100%. Otherwise
I would benefit from the performance delta - delta > 0.

"
...
By the way, these performance increases are not for free. You still need a 64-bit OS, 64-bit drivers, and 64-bit applications to take advantage of 64-bits.
...
"

Thats true, but you know very well that a 64BIt OS from
MSFT is coming online with driver support for the needed
parts. Problems will of course appear by old hardware or
with hardware from Non-Tier1/2 manuf. Everyone should aware
of that. But I'm thinking of the new boards, which have
nearly all needed things onboard, with USB, Firewire, LAN,
SATA, Wireless LAN and so on. Whats needed more for the
average user? Graphicdrivers - of course. Some TV Cards and
a few PCI soundcards - have I forgotten things, that have
more marketshare than 1%?
Furthermore we have free 64BIT OS with different
LINUX distributions ... so the hole argument gets very weak.

BUGGI

[dougSF30]

It's difficult to argue about realistic "ranges" without being more specific about what your "range" means. One standard deviation? Over what domain of applications?

That said, speaking informally, your range doesn't look realistic at all, with a midpoint of 7.5% improvement. The midpoint is, IMO, more likely to be something like 15% or even 20%. And it will be the rare application indeed that runs 5% slower. I'd guess that 90% of all applications will land in the 10-30% improvement area.

Doug