That's why I said different people focus on different things.
Notice that none of us here mention the i7 965 at all, yet there
will be people who'll buy it because they can afford to and they
want the best. For me, the bias edges slightly more towards
performance than price, because as I said before a system that
completes a task significantly faster can be turned off earlier,
saving power, and a 30% speed difference literally means getting
my project completed in 3 years instead of 4.
For gaming, the Phenom II is a good choice in terms of
price/performance, but for video encoding the i7 is a clear
winner. It all depends on what you use your system for. I do use
my PC for gaming aswell, but I'm not yet playing games that
require the power of the current newer CPU/gfx products - still
playing Oblivion & Stalker which run 2048 x 1536 totally smooth
with 16X AF and max-detail settings using a 6000+ (3.25GHz) and
8800GT (790 core). I have CoD4 still boxed, but existing data
suggests my current setup will run it just fine at 1280x1024 or
1600x1200.
As numerous reviews show again and again, the extra useful
performance attained with ever more expensive products (especially
SLI/CF gfx) is often minimal, with driver issues commonly making
a mockery of any theoretical performance ceilings. Those games
that benefit already run well with a good card. It does seem as
if review sites are becoming obsessed with the goal of seeing
Crysis running 2560 x 1600 max detail AA/AF, which is a bit daft
when few people are using displays of this kind yet (mine is a
good quality CRT monitor, an HP P1130, but it was cheap [78 UKP]
and 2048 x 1536 is its max). Besides, I've still yet to see any
evidence proving Crysis is even using gfx hardware fficiently.
How can we be sure Crysis isn't making gfx cards crawl simply
because it's poorly written? Few apps ever get more than a 3rd
of the peak performance out of any particular card before the
product cycle moves on to something else.
In general, it's becoming harder to find reviews of products
that cater for the majority of users. Ultra high-end reviews
of things like 3-way SLI tests are interesting, but I'd like to
see more useful results come out of these reviews. Include some
other CPU/gfx combinations, allow people to make informed
decisions. I have a 6000+; if I bought a GTX285, would my system
be able to run the current games ok? If I kept my current 8800GT
and upgraded the CPU to a Phenom II, would that make a useful
difference? And what speedup for other tasks?
To daerohn: I care. Yes, price is certainly an issue, but it all
depends your priorities. For me, the extra $300+ is worth the
expense since the performance difference is significant for my
intended core task (video encoding), especially once overclocked,
unless, as I say, it turns out to be possible to have two complete
Phenom II systems for the same price as one i7 system (jury out
on that one).
For those on a budget, or just those who are savvy enough to
look into the performance & value of products properly, the
Phenom II is clearly a worthy product. If I was buying a system
for gaming and nothing else right now, then depending on my
budget I'd either get a Phenom II or have fun with an oc'd E8400
or some such (loads of games still can't exploit a quad-core
properly). I'd only get an i7 for gaming if cost was not an issue.
To each his own. Saying the i7 is too expensive is illogical. It
depends on what you want to use it for. For video encoding and
animation rendering, the i7 is way ahead. For gaming, unless the
game(s) one wants to play will significantly benefit from the
i7's extra speed (ie. check reviews), the Phenom II is better buy.
Much of this is still hard to judge properly though since there's
as yet no direct comparison between Phenom II and the range of
older AMD/Intel CPUs that most people have, eg. the
submissions to futuremark suggest a great many people own 8800GT
systems, but nobody has included one in reviews so one can decide
if an upgrade is worthwhile. Something that annoys me about most
reviews is the assumption that users will always stay in the same
price bracket; what if one wishes to step up the ladder a little,
if not necessarily to the max?
In other words, we see lots of comparing new to new, which is
great to read, but what we need are comparisons of old to new in
those combinations that will be usefully enlightening.
Ian.