Article: Parallelism at HotPar 2010
By: Mark Roulo (nothanks.delete@this.xxx.com), August 4, 2010 9:55 am
Room: Moderated Discussions
Richard Cownie (tich@pobox.com) on 8/4/10 wrote:
---------------------------
>But it seems to be almost a universal law of chip
>development that you can't survive long by selling high-end
>chips alone, against competition that has a low-end
>high-volume revenue stream. That's how we got the
>dominance of x86 over RISC. The high-end-only chips that
>survive are those that belong to systems companies who
>can make their money not just on the chips, but on selling
>complete systems, plus software and services (which isn't
>what the graphics-card market looks like, though quite a
>few companies have tried to make HPC look that way -
>and failed).
This is my reasoning also, and I draw the same conclusion. I was hoping (still am, actually!) for some insight into the relative profits between the lines, though. For example, if the bottom 80% by volume GPU sales only contribute 10% of the profits, then I can see surviving w/o the lower 80%.
For CPUs, one of the things that make a low-margin/high-volume low end useful/necessary is that the high-volume tends to attract developers, who write apps, which make the entire platform more attractive.
In GPUs, we don't have this nearly as much because the primary "app" for the GPU is the driver. And the GPU vendor supplies that. In this sense, GPUs might be more like automobiles ... BMW and Porsche don't need a low end to survive. But I don't really know.
The other thing that is going to make nVidia's life harder is what I call the 487 effect. As low to mid range GPUs get provided "for free" with chips (especially laptops), even if the nVidia solution is "better" it just won't matter ... it probably can't be better enough to compete with "good enough and already paid for."
[NOTE: AMD has an edge here, right now. My assumption is that Intel will eventually get Larrabee or something like it working well enough to have a competitive low/mid range integrated product. If nVidia can keep competing against Intel's IGP then I can see then continuing to play in the low end ...].
-Mark Roulo
---------------------------
>But it seems to be almost a universal law of chip
>development that you can't survive long by selling high-end
>chips alone, against competition that has a low-end
>high-volume revenue stream. That's how we got the
>dominance of x86 over RISC. The high-end-only chips that
>survive are those that belong to systems companies who
>can make their money not just on the chips, but on selling
>complete systems, plus software and services (which isn't
>what the graphics-card market looks like, though quite a
>few companies have tried to make HPC look that way -
>and failed).
This is my reasoning also, and I draw the same conclusion. I was hoping (still am, actually!) for some insight into the relative profits between the lines, though. For example, if the bottom 80% by volume GPU sales only contribute 10% of the profits, then I can see surviving w/o the lower 80%.
For CPUs, one of the things that make a low-margin/high-volume low end useful/necessary is that the high-volume tends to attract developers, who write apps, which make the entire platform more attractive.
In GPUs, we don't have this nearly as much because the primary "app" for the GPU is the driver. And the GPU vendor supplies that. In this sense, GPUs might be more like automobiles ... BMW and Porsche don't need a low end to survive. But I don't really know.
The other thing that is going to make nVidia's life harder is what I call the 487 effect. As low to mid range GPUs get provided "for free" with chips (especially laptops), even if the nVidia solution is "better" it just won't matter ... it probably can't be better enough to compete with "good enough and already paid for."
[NOTE: AMD has an edge here, right now. My assumption is that Intel will eventually get Larrabee or something like it working well enough to have a competitive low/mid range integrated product. If nVidia can keep competing against Intel's IGP then I can see then continuing to play in the low end ...].
-Mark Roulo
