Application compatibility mattered a lot back then. There was no Microsoft Word or Excel for IRIX and Google Docs / Google Sheets didn't exist, for example. Many of the prosumers wouldn't have been able to use it as a daily driver. I recall that Adobe ported Photoshop to IRIX at one point, but I think it was gone by '97.
I wonder how much of this is a chicken-egg problem though; IRIX didn't have enough of a user base to justify porting over a lot of these applications, and because they didn't have these applications it couldn't develop a user base big enough, etc.
I think what you're saying is totally correct, if you're going to be spending a lot of money on a computer, you need to make sure it can actually do the stuff you need it to, and in the 90's that does more or less imply Microsoft Office compatibility.
Still, I do kind of wonder if SGI had actually tried to penetrate this market, that maybe they could have made this work. They could have licensed and ported some of these applications over themselves, and if they had gotten big enough then maybe some of these companies would have ported these things over.
Well, they tried producing high end Windows NT machines at one point, that would have run Office, had OpenGL support, etc. But no one bought them. So its not clear that doing the same but with Irix would have gone any better.
The last-gasp part of SGI I would have like to have seen develop was Cellular IRIX. If I recall it was a sort of distributed OS where the scheduler was node aware and could parcel out tasks within a single process to other nodes. What NFS did for the local/remote file system split, Irix would do for compute. Never came to fruition.
That sounds pretty cool, I guess the idea was that the scheduler could figure out smartly where to put applications and minimize latency in the process?
Yes, much like you would in an HPC system (much of Cellular IRIX was supposed to have been influenced by what Cray and SGI learned from each other when they merged).
More details (PDF) here. Remember this is from 1998. Reading it again in 2024 there is of course quite a bit of foreshadowing of what we ended up with in cloud. And you won't be surprised to learn that the author ended up passing significant time working at both Microsoft and AWS after their SGI days.
Comments
Application compatibility mattered a lot back then. There was no Microsoft Word or Excel for IRIX and Google Docs / Google Sheets didn't exist, for example. Many of the prosumers wouldn't have been able to use it as a daily driver. I recall that Adobe ported Photoshop to IRIX at one point, but I think it was gone by '97.
I wonder how much of this is a chicken-egg problem though; IRIX didn't have enough of a user base to justify porting over a lot of these applications, and because they didn't have these applications it couldn't develop a user base big enough, etc.
I think what you're saying is totally correct, if you're going to be spending a lot of money on a computer, you need to make sure it can actually do the stuff you need it to, and in the 90's that does more or less imply Microsoft Office compatibility.
Still, I do kind of wonder if SGI had actually tried to penetrate this market, that maybe they could have made this work. They could have licensed and ported some of these applications over themselves, and if they had gotten big enough then maybe some of these companies would have ported these things over.
It's tough to say.
Well, they tried producing high end Windows NT machines at one point, that would have run Office, had OpenGL support, etc. But no one bought them. So its not clear that doing the same but with Irix would have gone any better.
The last-gasp part of SGI I would have like to have seen develop was Cellular IRIX. If I recall it was a sort of distributed OS where the scheduler was node aware and could parcel out tasks within a single process to other nodes. What NFS did for the local/remote file system split, Irix would do for compute. Never came to fruition.
Archive.org has a relatively detailed Tech Report from SGI on Cellular Irix.
https://web.archive.org/web/19970706222424/http://www.sgi.co...
That sounds pretty cool, I guess the idea was that the scheduler could figure out smartly where to put applications and minimize latency in the process?
Yes, much like you would in an HPC system (much of Cellular IRIX was supposed to have been influenced by what Cray and SGI learned from each other when they merged).
More details (PDF) here. Remember this is from 1998. Reading it again in 2024 there is of course quite a bit of foreshadowing of what we ended up with in cloud. And you won't be surprised to learn that the author ended up passing significant time working at both Microsoft and AWS after their SGI days.
https://cug.org/5-publications/proceedings_attendee_lists/19...