Intel graphics drivers for linux

[The Natural Philosopher]

Yes. There are less bugs in the open source drivers. Sadly the Nvidia HARDWARE is much better than the competition and with the Nvidia drivers the performance is there..

[The Real Doctor]

It's written by Intel staff and made available on an Intel website. What more do you want?

Ian

[Andrew Gabriel]

You may be thinking of Microsoft - it often came as a bit of a surprise to new unix source code programmers to find lots of Microsoft copyright notices in the source code. These were parts which either came from Xenix or included modifications originally done on Xenix, which Microsoft owned for a period of time, before SCO took it on.

Intel had contributed very little to unix at that stage, and it was actually very slow do start doing so. It was not even doing so when it was developing Itanium - it relied on HP and Sun to do that for it. Ironically, it was AMD that kicked Intel into more software development. AMD launched its 64 bit Athlon and Opteron processors, which were instantly of interest to the Enterprise computer manufacturers. I did some of the first work at Sun with AMD on this at the time. We worked with AMD to get the Enterprise class error reporting into Opteron, so it could do the same type of error recovery which SPARC did at the time, enabling an OS to survive processor and other faults.

Intel started trying to catch-up with its Yamhill EM64T instruction set. It had developed Enterprise class error reporting for Itanium, which it moved across. Intel's was much more complicated to drive than AMD's, and no operating system had started using it yet even on Itanium, so there was no code for it. Intel started producing prototype code for this, and paid for 10 engineer positions in Sun's kernel engineering team to take the prototype code and integrate this in to the Solaris kernel (and I suspect they did the same for Windows with Microsoft - these were the only two x86 OS's at the time which had error recovery frameworks to allow them to continue running when parts of the processor or other hardware failed).

There were a few much earlier ports of unix to x86. These were done by Microsoft (Xenix) and Interactive Systems Corporation (Venix/86, PC/IX, and SysV/386).

[The Real Doctor]

Did you even bother to look at

That's a website run by Intel from which can be downloaded full Intel drives, written by Intel, supplied by Intel and maintained by Intel.

Didn't you spend many years on uk.d-i-y promoting magnets as a way of reducing limescale in kettles?

Ian

[dennis]

No, I knew Microsoft had done Xenix stuff bu this was done by the Intel skunk works.

Yes, this was later being the source for SVr5 on which Solaris and Unixware are based. There was a lot of it in System X exchange peripheries as I specified the Unix systems and design a lot of the software and stuff. Each exchange had clusters of bit slice processors and each cluster had a SBC based either on an Intel

[dennis]

Opps, must remember its a touch screen when I decide to wipe a hair from the screen, that way I don't press the send button half way through typing.

Anyway..

Each cluster has an SBC (its all duplicated for availability but they operate independently so I will ignore that), either a Motorola VME or an Intel MB2 in an adapter to fit the bus in the cluster which run diskless SVr5 talking to a Unix server also running SVr5.

Quite a lot of the software on the SBC was written in the STREAMS driver to give a "real time" response.

It was one of the more interesting systems I designed.

[Tim Streater]

Confirms to me that touch screens are a dead loss.

[dennis]

I use it all the time, I struggle with just a mouse. Its so easy to just prod the screen and move a document about.

The pressure sensitive pen is also useful for photo editing.

[Tim Streater]

I have an Apple Mighty Mouse, sadly no longer made. It has a small tracker ball in the top so I can scroll a document in two directions at once as well as move the mouse pointer (at the same time).

Right hand: rests on desk on mouse

Left hand: props up chin

If I had to touch the screen I'd need a very short one of those jobbies used to keep load-bearing walls from falling down before the I-beam gets put in. I'd be needing it to support my arm between shoulder and screen.

It would get in the way.

[The Natural Philosopher]

No, they are b'development' drivers. Not p[olished code.

No. Never.

[dennis]

Get a monitor and put the touchscreen flat on the desk. Duplicate the screens. Use it like a tradition graphics tablet BTGTTS.

[Tim Watts]

Worked in Star Trek (Next Gen. onwards) - and yet, you never see:

1) Greasy smeared panels;

2) Anyone polishing them every 5 minutes;

3) Anyone blow the left warp nacelle out because they "typed" 9 instead of 0!

[Tim Streater]

Then there'd be no room for the keyboard, or for papers I want to refer to as I work.

If I got a second screen it would always be for extended desktop.

Graphics tablet eh? Haven't seen anything like that since I was writing software for a vector graphics display connected to a CDC 3800 in 1970.

[The Real Doctor]

Sez you. On the basis of ... what evidence? How much of the source have you reviewed?

Ian

[Andrew Gabriel]

Solaris and Unixware were SVR4 based.

The last System V Unix standard was SVR4, after which Unix switched to the POSIX standards body.

SVR5 was a marketing name SCO used to merge their Openserver and Unixware products together - it's not a Unix standard. It was a (failed) marketing ploy to try and make the unix world think that Unixware was relevant, which it never was.

[The Real Doctor]

Incidentally, you may care to note that the Ubuntu repositories currently contain version 2.99.910 of the Intel drivers, whereas i0.org (that's the Intel-run, Intel-maintained website) has 2.99.917. You might also care to compare the AUTHORS file from the Ubuntu repository source:

with the AUTHORS file from the Intel source

Ian

[William Unruh]

Some places call everything, no matter how polished, development. For example cdrecord calls everything alpha software. That does not mean that it does not work. And many people call stuff "final, polished code" which crashes and burns at the least provocation. The name is not the thing.

[The Real Doctor]

Mr NP started by claiming that Intel provided no drivers at all, then that they provided only "reference" drivers, then that provided only "development" drivers. Which is pretty good since he also claimed that Intel don't even let their own software people know how the hardware works.

As I have pointed out, the official X.org drivers for Intel graphics hardware are written and supported by staff at ... Intel.

Ian

[news]

I've cut the xposts as the other groups are far too erudite for me.

Using mint17 when I play my xp vm I get a warning that the host doesn't support 3d. I take it this is because my zoostation 15 graphics drivers aren't compatible with Linux. Is there anything I should do to correct this?

AJH

[The Natural Philosopher]

Is that whty they dont work then?