3 ms·

Why is there "corporate disinterest" in the SGI niche? Or who is occupying it today? One would think that a high-end graphics workstation (with the looks to sc

Why is there "corporate disinterest" in the SGI niche? Or who is occupying it today? One would think that a high-end graphics workstation (with the looks to scream "creative") running Linux would be a viable proposition? I suppose game and movie studios, architecture, design and engineering shops etc. all somehow cope with more generic solutions and some combination of local and cloud compute but clearly cutting corners in the overall experience and productivity. There is usually an upmarket segment that offers that something extra (dully compensated) but somehow not in this area...

jamesfmilne 3y ago
PCs are "good enough" to do the job, the same reason they killed off these beasts back in the first place. The software I work on started life on SGI, but has been on Linux PCs for the last 20 years. For the top end, we use big servers with 6-8 GPUs, but they sound like jet aircraft nowadays so they live in a machine room. https://www.filmlight.ltd.uk/products/baselight/overview_bl.php https://www.filmlight.ltd.uk/products/baselight/overview_bl....
- nologic01 3y ago
  Yes, that is my exactly my experience too, but the task of innovators is to prove that "good enough" is just not good enough and that we have a need that we didn't know - for which we are willing to pay a premium :-) The underlying driver always involves the economics of cost of development versus market size and in the past this was ruthless for specialized vendors. But today, e.g., NVIDIA is riding high on what are extremely specialized chips, presumably on the assumption of massive adoption. So its not clear if we are still in the regime of commoditized computation that destroyed the Unix workstation era or maybe the category will be reinvented.
- jeffbee 3y ago
  At the moment that the PC killed off Unix workstations, the PCs were already radically better. We were all just waiting for application vendors to port to Windows NT. When my SGI was replaced by a dual Pentium Pro it was a little shocking how much faster it was.

jamesfmilne 3y ago

PCs are "good enough" to do the job, the same reason they killed off these beasts back in the first place. The software I work on started life on SGI, but has been on Linux PCs for the last 20 years. For the top end, we use big servers with 6-8 GPUs, but they sound like jet aircraft nowadays so they live in a machine room. https://www.filmlight.ltd.uk/products/baselight/overview_bl.php https://www.filmlight.ltd.uk/products/baselight/overview_bl....

nologic01 3y ago
Yes, that is my exactly my experience too, but the task of innovators is to prove that "good enough" is just not good enough and that we have a need that we didn't know - for which we are willing to pay a premium :-) The underlying driver always involves the economics of cost of development versus market size and in the past this was ruthless for specialized vendors. But today, e.g., NVIDIA is riding high on what are extremely specialized chips, presumably on the assumption of massive adoption. So its not clear if we are still in the regime of commoditized computation that destroyed the Unix workstation era or maybe the category will be reinvented.
jeffbee 3y ago
At the moment that the PC killed off Unix workstations, the PCs were already radically better. We were all just waiting for application vendors to port to Windows NT. When my SGI was replaced by a dual Pentium Pro it was a little shocking how much faster it was.