What is involved in produced a driver for Vista x64?

[Neil]

I mean, once a hardware vendor has already gone to the trouble of making one
for Vista x86. Not knowing anything about it at all, my naive thought is
that once he'd done that, he would just compile it again but with the switch
in the "this driver is going to run in a 64bit system" position rather than
in the "this driver is going to run in a 32bit system" position.

There must be more to it than that though, unless the vendors of some of my
hardware are just being bloody-minded, because I have a couple of things
which have 32bit drivers but not 64bit support.

So why?

Neil

 

[GTS]

A simple recompile like you describe would be applicable a lot of
application programs, which are written in high level languages. Device
drivers are usually written in assembly language and there are significant
differences in CPU registers and operations in the 32 and 64 bit OS.

 

[CJM]

A simple recompile like you describe would be applicable a lot of
application programs, which are written in high level languages. Device
drivers are usually written in assembly language and there are significant
differences in CPU registers and operations in the 32 and 64 bit OS.

Device drivers are usually written in a high level language like C++, and
it's the compiler that usually has to worry about the differences in
architecture. So often, it *is* simply a case of recompiling. [AFAIK, there
are exceptions where performance is of paramount importance].

Why then is it such a problem getting hold of software that works for x64?

a) Because it's not (yet) a priority for many manufacturers.
b) Because Vista x64 needs it's drivers signing, which involves certificates
etc.

 

[Neil]

Why then is it such a problem getting hold of software that works for x64?

a) Because it's not (yet) a priority for many manufacturers.

But for hardware vendors who have already produced an x86 Vista driver, the
marginal effort required to produce an x64 version (from what you're saying)
is next-to-non-existent.

b) Because Vista x64 needs it's drivers signing, which involves
certificates

Does the vendor have to pay for certification? In which case, I suppose
that could explain why a piece of hardware might have an x86 driver but no
x64 version. Though (I'd have thought) the cost of sorting out a Vista
driver at all must be mostly in the programming.

If it really is that easy, are hardware vendors just thinking, "Ah he's
installed the x64 version of Vista, he's someone who likes to get new stuff
for no apparent reason; I bet we can convince him to buy the latest version
of our thing, by only providing x86 drivers for the slightly older stuff."
Or am I too cynical?

Neil

 

[CJM]

But for hardware vendors who have already produced an x86 Vista driver,
the marginal effort required to produce an x64 version (from what you're
saying) is next-to-non-existent.

It suggest it is easier than they make out.

Does the vendor have to pay for certification? In which case, I suppose
that could explain why a piece of hardware might have an x86 driver but no
x64 version. Though (I'd have thought) the cost of sorting out a Vista
driver at all must be mostly in the programming.

I think it's around $500 for certification but that's just hearsay. I have
no actual experience of this area.

If it really is that easy, are hardware vendors just thinking, "Ah he's
installed the x64 version of Vista, he's someone who likes to get new
stuff for no apparent reason; I bet we can convince him to buy the latest
version of our thing, by only providing x86 drivers for the slightly older
stuff." Or am I too cynical?

I would bet my granny that manufacturers are dragging their heels on
producing *any* drivers for older equipment in the hope that you'll be
suffiiciently motivated to by new kit. Creative anyone? X-Fi were sorted
first, followed by Audigy 4, then 2, and then older cards....

Just because you are cynical, doesn't mean you are necessarily wrong!