Article: Parallelism at HotPar 2010
By: Richard Cownie (tich.delete@this.pobox.com), August 4, 2010 11:49 am
Room: Moderated Discussions
Linus Torvalds (torvalds@linux-foundation.org) on 8/4/10 wrote:
---------------------------
>Why? Because a lot of the costs are fixed. The cost of
>things like driver development is largely fixed. The
>cost of much of the engineering is similarly largely
>fixed.
Another way of putting this is that "profit" is an
accounting fiction. If your bean-counters choose to
assign all the fixed development costs to the high-end
cards, then the low-end ones might appear to be making
a profit. But that's just shunting figures around on a
spreadsheet, not really changing anything. What's
real is cashflow: how much money comes in from selling
stuff, how much goes out in paying staff and making
products. The best approximation to figuring out whether
a particular market segment is useful is to take the
sales revenue minus the cost of making the stuff
(i.e. gross margin times revenue). That takes the
fixed costs out of the equation.
However, long-term success in a business like 3D
graphics requires that you maintain enough volume of
the right kind of products to keep game developers
interested in optimizing for your platform. That's
the kind of issue that gets tough if you retreat to a
low-volume high-end boutique business model.
---------------------------
>Why? Because a lot of the costs are fixed. The cost of
>things like driver development is largely fixed. The
>cost of much of the engineering is similarly largely
>fixed.
Another way of putting this is that "profit" is an
accounting fiction. If your bean-counters choose to
assign all the fixed development costs to the high-end
cards, then the low-end ones might appear to be making
a profit. But that's just shunting figures around on a
spreadsheet, not really changing anything. What's
real is cashflow: how much money comes in from selling
stuff, how much goes out in paying staff and making
products. The best approximation to figuring out whether
a particular market segment is useful is to take the
sales revenue minus the cost of making the stuff
(i.e. gross margin times revenue). That takes the
fixed costs out of the equation.
However, long-term success in a business like 3D
graphics requires that you maintain enough volume of
the right kind of products to keep game developers
interested in optimizing for your platform. That's
the kind of issue that gets tough if you retreat to a
low-volume high-end boutique business model.