More is better, that's what 100's of years of marketing has
tried to make us believe. Who can argue
with having a bigger house, a gaggle of friends or more money?
If you're a gamer you're very familiar with the mantra of
more. More performance usually means
more clock speed, more memory, more video cards. More of everything except what's in your
wallet after you pay for all of it.
In the swirling vortex of the enthusiast PC market churn is
as natural as the change of seasons.
It is a "market" after all, built on a volume of repeat
sales but often it's participants are disingenuous about the merits of their
wares. It's common, for example, to see
a previous generation's graphics platform get a few tweaks and reintroduced as
something new.
Remember the Nvidia 8800 GT?
It was still around 3 years later having changed names twice before it
finally ended up as the GTS 250. Everything
old is new again I suppose.
We see a market where performance gains from 1 or even 2
generations back only rise to the level of single digit percentages. All of these "features" you hear about while useful are just window
dressing. It obscures the fact that
we're getting more fluff and less substance.
Want a recent example?
Look no further than the latest Intel family of CPU's, Haswell. It's got better power consumption, improved
embedded graphics on the chip (whether you want them or not) and native support
for USB 3.0. Yet in spite of the
marketing, real world performance is
only marginally better than its Ivy Bridge predecessor.
It strikes me that the whole push for more everything may
actually be a sign of failure. It
doesn't say much for innovation if the major players need to recycle old
technology just to fill out a lineup. It
also makes me wonder why in this day of doing more with less that performance
is still measured in having multiples of the same component with features
irrelevant to the primary function.
Copious CPU cores are great if you need them but 90% of us
don't. 2 to 4 threads is more than
enough for most people, 16 is just gross overkill. Multiple
GPU's may be required for ridiculously high resolutions and eye candy but the
average gamer doesn't need them for 1920 x 1080 gaming.
If they do, something is very, very wrong.
I may be speaking heresy here but I really do think the whole
concept of SLI or multiples of the same anything for that matter is a huge
admission that innovation has stalled. I
shouldn't need two graphics cards to do the job of driving one display or even
multiple displays (within reason). I
also shouldn't need 8 cores just to open a spreadsheet.
That we've accepted that configuration for over a decade is
even more damning of the lack of innovation.
We accept lower performance from mobile devices because their
convenience will always trump any performance shortcomings.
$700 to $1000 for a graphics card is neither cheap nor
convenient. That much money should not
require an exponent (x2) for what is ultimately a middling benefit.
Perhaps there are advantages but my point is that at this
stage you shouldn't need 2 of anything just to enjoy a little more than the
bare minimum. We should be an order of magnitude further
along than we are but instead we're forced to buy multiples of what often
amounts to just rehashed old technology.
I don't deny that those of you that have embraced the
concept of more (SLI, 8 core CPUS) are seeing some benefit. I just
believe that the "more" you should be getting has nothing to do with what
you ultimately ended up with.
Forgive me if I lapse
into a bit of gearhead tech but I see a correlation here. For a long time cars depended on a device
called a carburetor to get fuel into an engine.
Over time refinements to its design were made to improve its efficiency but in the
end you still have a largely unregulated fuel delivery system. At the end even computer controls couldn't
provide enough lipstick for that pig. It
took fuel injection to change things.
Right now, PC tech might as well be carburetors. Just refinements to the same old technologies being passed off as revolutionary.
No comments:
Post a Comment