Over the course of the last few days, I stumbled into the
usual flood of speculation and details that didn't have a confirmed
source. As a general rule of thumb on DailyTech, we do not post articles
on information that cannot be attributed to a source in writing -- but since we
added a blog engine maybe it's time for me to express a few of the things I've
During the show we saw some of our first NVIDIA slides with G80
information. It's much too early to speculate about specifics of G80, but
several vendors who have been reliable in the past have told me that G80 is a
dual-core GPU with 48 pixels-per-cycle per core. Given that the 90nm G71
is based on a design with a mere 24 pixels-per-cycle core, I am having a little
bit of trouble swallowing this tall order. The NVIDIA roadmap we've seen
stretches into Q4'06, and does not have G80 on the map yet. That's not to
say G80 is not a 2006 component, but it's not on the roadmap yet as such.
Details of R600 were almost non-existent at Computex, even in rumor form.
We got our first look at a few slides that seemed to show R600 uses GDDR4, but
this seems sort of a given considering vendors are already considering using
GDDR4 on the R580 core.
As with G80, it's probably worth stressing that the ATI roadmap extends into
Q4'06 and does not mention anything of R600.
The talk on everyone's tongue during Computex was the ATI-AMD prospective merger.
While I can't say I've seen evidence one way or another on this sort of deal, a
few of my colleagues seem quite adamant
on both sides of the fence.
Personally, I have a habit of not picking an opinion until I see it in
Conroe and AM2 also got their fair share of praises and slams
accordingly. Controversy still surrounds my former
employer’s benchmarks of Conroe and Woodcrest, though I am
reasonably confident the benchmarks at this time are starting to pan out in Conroe's
favor. However, the motherboards are still fairly immature and so far as
I can tell no one has really been able to push the envelope with Conroe or
Revision F outside of a few limited scenarios.
For all those AGEIA fans out there I have good news and bad news. The
good news is that PCIe versions of PhysX are on the way this year. ASUS
and BFG are the sole hardware providers in the US right now, but other tier 1
manufacturers have already confirmed to me that the exclusivity of ASUS and BFG
will expire before the end of the year. Now the bad news: it's still
going to cost $280. Engineers at all of the
AGEIA partners seemed to hint that the PCI version of the PhysX card will
come down in price at the time of the PCIe launch, but the results are
conflicting as to whether or not AGEIA will adopt a new ASIC.
Computex had its showings of UMPCs and PMPs, as I had anticipated,
but not nearly in the volume I had hoped. I had almost certainly
expected some surprise UMPC announcements, but for the most part the product
managers I've dealt with wish to avoid the UMPC project like the plague...
instead focusing on PMPs or tablet PCs instead. The majority of this animosity
seems to revolve around battery life, as the device is too small to add a
proper 6+ cell battery but also requires power hungry processors typically
found in notebooks. The Linux-based AMD
alchemy devices from ECS were actually my most favorite, as the hack scene
for such hardware is virtually unlimited.
Just to wrap up, I suppose I should mention HDCP -- considering virtually
everyone on the show floor couldn't stop talking about it. HDMI made a
fairly strong appearance during the vendor showcases, which is great.
Almost every vendor is incorporating HDMI-to-DVI adaptors in with the
HDMI-capable video cards at almost no additional cost. I see this as a
great move as the average RadioShack still charges $30+ for an adaptor.
The thing that puzzled me though was the number of HDMI-ready cards on the show
floor that didn't support HDCP. Granted, the HDCP certification is not
cheap (Fudo claims its
$15,000 -- which is true if you ship more than 10,000 units a year).
I personally do not advocate DRM, particularly when it comes to HDCP, but I
think the ship has already sailed with regard to getting TV manufacturers to
stop implementation. The adaptor manufacturers are certainly hesitant,
but you can be sure if HDCP implementation was free to them we'd already have
it in every PC and notebook.
quote: THIS ISNT FUCKING ROCKET SCIENCE