in 1995-remember the announced $180 LockheedMartin R3D-100?

ATI's engineers

most of Real3D's engineers got hired by ATI when Lockheed decided to stop its forray into the consumer 3D graphics segment.

how else did you think ATI started releasing such awesome 3D hardware all of a sudden?

 

This might be a bit OT now, but I'll just say this and then be quiet, OK?

Well, in that case you are definitely way better informed than me.

It doesn't, according to my recollection. The 486 had 8 kb, but I think that was the first of the x86 processors with internal cache. (Probably external too, for that matter.)

 

Re: ATI's engineers

LOL

 

Sorry for this OT post.

http://slashdot.org/articles/02/07/22/0615221.shtml?tid=126

US fighters are too fragile, they are high tech toys, unlike Russian fighters, which are much more robust. Well, they have to be robust.

You are right with the 8KB L1 of the first i486. The 386DX could have had 128KB L2 on the motherboard.

[Edit: changed "sensitive" into "fragile"]

 

Isn't it a bit risky using a consumer gfx chip for military use? Hell, it's only been a couple of years since NASA upgraded to 486s in its space gear.

 

Interesting link, but the F-22 is not done yet. It has multiple redundancies in its computer, and the software is far from finished. I'd be highly surprised if it ever required an in-flight reboot once the aircraft finally shipped (Somehow I don't think it would go over very well with pilots in the middle of a fight...).

 

Just a small point of interest:

When I was working on a payload for the (later cancelled) Mars 2001 Lander, I was doing interface testing in Denver. The lander CPU was a MIPS based beast that took about 3 minutes to reboot.

Guess exactly how effective that would be if you had a reboot upon reentry. Perhaps an unlikely event, but still....

 

well consumer electronics have moved so much incredibley faster than anything other industry releasing new products and revisions each quarterly and in between...the military or NASA are begining to find out they are the ones technologically left behind.....remeber couple months back , there was something about NASA requiring 20 years old intel chips beacuse they are no longer in production

 

<more OT>

Most likely if one system reboots, its backup system takes over critical tasks during flight. Check out some of the comments in the slashdot thread.

But my comment about the robustness was meant overally, not about the electrical systems.

From what I've heard every morning the personal of an airbase in the US must search for small objects on the runway that could get into the air intakes of the F-16 before it could start. Russian fighters can start on almost any runway, doesn't matter how messy it is, because they have big grills on the air intakes that takes care of such objects.

Several years ago I saw a video of a SU-27 which had to land on the fuel tanks beneath the fuselage because the gear was broken. The sparks flew when the metal touched the ground, but it landed succesfully, without fire. The airplane was repairable afterwards. That convinced me that the SU-27 is a flying tank.

It is simply a different approach in plane design. Americans build their fighters like high precision, high tech machines, which are more fragile and need lots of maintenance, Russians still more like tanks which have to work in a worst case scenary with minimal requirements.

BTW, it is really bad that the SU-27 caused the death of ~60 innocent people today at an airshow in the Ukraine.

</more OT>

 

Not all american fighters are "high-tech" fighters.

The A-10 is one example. The aircraft is build with multiple redudancies on all of the avionics, and the pilot sits basically in a titanium tub in the cockpit.

There have been incidences where the craft was severely damaged (And I'm talking about it having many holes in the wings, fuselage, etc.), and still managed to limp back.

Anyway, the F-22 is a different kind of design. It's meant to be an elite superfighter, so the technology is necessary. It doesn't look like more than around 200 will be built. Basic durability is sacrificed for combat superiority. But, you still have to consider that the F-22 is not yet in full production. It may yet turn out to be quite a bit more durable than current reports indicate.

The real next-generation mass production fighter coming up is the Joint Strike Fighter (which Lockheed also just received the contract for). It is meant to be a fighter that is designed to work for all major branches of the military, with different modifications for each one. It's being designed for cheap mass production, as well as stealth and fairly so phisticated avionics.

 

In the general public, there is a slight misconception that the Hubble telescope et al are flying around with desktop Intel 486 processors in them. Your average PC doesn't have to worry about the operational radiation environment that the Hubble has to; temperature isn't a huge problem (with the average around 10~15C). The architecture (and therefore, the associated programming) follows a 486 but in reality it is constructed differently to the desktop chip with the oxide layers, interconnects, etc almost "hand-crafted". IIRC, it's Sandia that makes the radiation-hardened chips.

 

I'm kind of curious about heat dissipation in space, though...in a vacuum, there is virtually none. I imagine they have to implement some fairly sophisticated temperature controls (Also, the temp does vary quite a lot in space...).

 

No convection, but plenty of heat loss via radiation since the temperature gradient between the object and environment is pretty high.

 

The only thing is, don't they have to fully-enclose the chip? After all, if it's not insulated, it will fry when it's on the day side. I'd be interested in finding out whether or not there are any significant cooling hurdles...haven't read anything about it.

 

IIRC, the HST486 computer was incased in a damn big box! I saw a picture of it a while back but I can't remember from where. I've not seen many reports myself and ever since I stopped working at a science museum (about 6 years ago), I've not kept up-to-date with any of the general technology used in the HST. I have found one PDF report on the test mission done in 95, I think, called HOST - this was a run designed to see if the upgrade parts for the HST would be okay:

http://hstsci.gsfc.nasa.gov/host/docs/HOST_Post_Mission_Review.pdf