SGI: Discussion

unixmuseum wrote:

Timberoz wrote: For one it might see some new software come to the platform. Or better still support stay for some of the older products that left. Maya, Discreet apps just to name a few.

Or see ongoing development for existing ports.

Timberoz

You'd get all this by porting IRIX to x86?!?!?! How??? I know IRIX is pretty magical, but it will be able to turn old MIPS binary into shiny new x86 code?

Porting IRIX to x86 was a good idea. Like in the late nineties, when it was obvious that they wouldn't be able to continue to compete with Intel/AMD on performance. They then could have been first out of the block with 64bit Opteron machines. Software vendors were still supporting them, Linux hadn't yet established as strong a foothold. Apple has successfully managed these types of transitions twice - and that's what has kept them alive today. If SGI had done this.......

But it's too late now.....

Nah, it wouldn't have saved them.

The problem is that cheap, commodity workstations were going to kill them because consumer graphics were getting huge and taking over.

So they went for the workstation market head on specifically. That's like attacking a walled city; not recommended.

They should've went the other route and got into consumer graphics. They have done consumer graphics devices before with the N64. Not that the N64 was in any way competitive in the 90s, that's not my point, my point is they had the resources to invest in developing something new and competitive on AGP or something. It's not like they couldn't transfer their patents and know-how to another target market.

SGI had enough coolness in graphics that plenty of people would have loved to have an SGI graphics card. I would've. I was familiar with the name SGI since I was a small kid and my brother showed me an article in his EGM magazine about them. Everyone knew SGI and had a strong association, SGI with graphics.

SGI for storage servers? Whiskey tango foxtrot?

It sounds very conservative and "safe" to think they should abandon the "flashy" graphics business and focus on those workstations, but we all know how well that went in practice.

Obviously hindsight's always 20/20, but I don't think it was a bright idea at the time either. If they could invest all they did to develop IA64 systems and completely reinvent their product line to compete with something everyone knew would kill them, surely they could spend that money to get into a business that (for very related reasons!!!) everyone knew would be huge.

You can't say they didn't know NVidia and the consumer graphics companies were going to be big; that's what got the ball rolling on all this.

Frapazoid wrote: They should've ...

It's a dead horse. You could fill an Encyclopaedia Brittanica-sized bookshelf with "they should have" but ... they didn't.

That's why they pissed away the billion dollars they had and went bankrupt.

I find it interesting because I am seeing the same thing in smaller foreigner companies all over China now - absolute imbeciles in charge. But they certainly get their oversized salaries .... wonder where this will lead, eventually ?

Frapazoid wrote: So they went for the workstation market head on specifically. That's like attacking a walled city; not recommended.

They should've went the other route and got into consumer graphics. .

What makes you think that if they couldn't compete with commodity workstations they could compete with commodity graphics?

The company fundamentally isn't capable of competing in commodity anything - their business practices and model just won't support it. IRIX on x86, if done at the right time, could've kept people buying professional workstations. Some people still do - IBM, HP, SUN all offer some form of professional workstation, but the market isn't really big and they aren't that much better. The whole problem with the commodity vs. professional workstation wasn't the price - it was the fact that the commodity worksations were actually faster ! Now, plenty of people wiould have been willing to pay for the advantages of professional gear if they had kept pace with the commodity hardware, never mind if the prof. kit was even faster.....

unixmuseum wrote:

zizban wrote: Heck port Irix to x86.

What would be the point of that?

Making grown men cry.

VenomousPinecone wrote: Making grown men cry.

:-D

Making grown men cry.

Sad but true.

What makes you think that if they couldn't compete with commodity workstations they could compete with commodity graphics?

Caaause making commodity workstations they would be selling essentially the exact same thing as any of their competitors.

Consumer graphics - and I'm talking about gaming class and above - is very competitive and they could actually put their technology to use. They've made GPUs for game consoles before, that's what ATI and Nvidia are doing now in addition to gaming cards for consumers.

And cards for workstations. The proprietary workstation may have gone the way of the dinosaur but workstation class parts for PCs are still on the market from ATI, Nvidia and a few others I can't think of right now.

SGI had a lot of technology years ahead of everyone else. Technology matters there.

Low class, made for MS Word type chips from Intel and SiS aren't what I was refering to.

It's a dead horse. You could fill an Encyclopaedia Brittanica-sized bookshelf with "they should have" but ... they didn't.

I know, I just get tired of hearing it's what the had to do .

It was a completely unambitious decision made from fear. Basically "RUN AWAY! RUN AWAY! CHEAP GRAPHICS ARE BETTER THAN US!!!"

I don't respect that at all and I don't think that sort of decision should be praised. It's wasn't wise, safe, mature, smart.

Making grown men cry.

>SGI had a lot of technology years ahead of everyone else.
-Which is an apt description, but nobody seems to (have a willingness) to read these words well and with an adequate sense of reality...

>Technology matters there.
-And it is that what SGI does not have anymore, except for their
alliance/allegiance with Intel. IA-64 is rather nice and SGI's convergence
with Linux and their own ProPack are doing the job acceptably (to the level of SGI/Irix standards[?]). So time/everything else moves on...

Silicon Graphics is like an old love, fond memories/nostalghia, but new ones appear on the horizon...

Thus, enjoy the view in the museum of the origins/octanes, whatever
and get to grips with the fact the SGI existing today is not the Silicon
Graphics we all revere (at least some of us do).
In the meantime happy dreaming...

AvS

SGI had a lot of technology years ahead of everyone else.

Emphasis added for anyone who's going suggest I was implying they have anything relevant now