VOGONS


The myth of the vertical retrace interrupt on EGA/VGA

Topic actions

Reply 60 of 73, by Scali

User metadata
Rank l33t
Rank
l33t
SirNickity wrote:

Results are mixed enough that you can't count on it being supported, so it wouldn't make much sense to write software that relies on it.

That is the takeaway.
Until I wrote these experimental tools and had people run them on a wide variety of hardware, I had no idea of how broad the support was for this feature.
Now we know that it is indeed not something you can generally count on. I mean, if it was say one in 20 video cards that wouldn't support it, especially if those would be the cheap brands like Trident or Realtek, I could justify using it.
But it seems that even many popular and high-end cards don't support it (or the chips do, but it's not connected on the PCB).

The only thing that's still missing is the results of a real IBM PS/2.
I mean, with 8088 MPH, our justification was that at least all our code worked perfectly on any true blue IBM (5150, 5155 and 5160).
If this feature works relibably on a PS/2, then at least I could create some kind of '8088 MPH' for the PS/2, and use this as one of its defining features.

http://scalibq.wordpress.com/just-keeping-it- … ro-programming/

Reply 61 of 73, by Error 0x7CF

User metadata
Rank Member
Rank
Member

I could run it on my IBM PS/1 Model 2011 sometime Friday if you'd like. I'm pretty sure it's VGA and not MCGA.

What I can tell you right now is it doesn't generate interrupts on my Thinkpad 755CS.

Old precedes antique.

Reply 62 of 73, by Error 0x7CF

User metadata
Rank Member
Rank
Member

Zero IRQs out of my PS/1. Then again, it was a value machine so it may not be connected for similar reasoning as IBM's ISA card.

Old precedes antique.

Reply 63 of 73, by Scali

User metadata
Rank l33t
Rank
l33t
Error 0x7CF wrote:

Zero IRQs out of my PS/1. Then again, it was a value machine so it may not be connected for similar reasoning as IBM's ISA card.

Thanks for testing, and yes, perhaps you're right.
I suppose the most interesting test would be a true PS/2, ideally the first model to ever introduce VGA (a model 30, 50, 60 or 80 apparently).

http://scalibq.wordpress.com/just-keeping-it- … ro-programming/

Reply 64 of 73, by SirNickity

User metadata
Rank Oldbie
Rank
Oldbie

I just won an auction for a Model 30 286. I'll give it a try when I can. Will probably have to wait a couple weeks for shipping, cleaning, and PSU recapping.

Reply 66 of 73, by elianda

User metadata
Rank l33t
Rank
l33t

The only thing that's still missing is the results of a real IBM PS/2.

Done

I have tested it on my IBM Model 80 8580 and it shows:
IRQs: 0

Retronn.de - Vintage Hardware Gallery, Drivers, Guides, Videos. Now with file search
Youtube Channel
FTP Server - Driver Archive and more
DVI2PCIe alignment and 2D image quality measurement tool

Reply 67 of 73, by Scali

User metadata
Rank l33t
Rank
l33t
elianda wrote:

Done

I have tested it on my IBM Model 80 8580 and it shows:
IRQs: 0

Okay, now it REALLY gets interesting!
Did IBM deliberately disconnect the IRQ line? Or is there some kind of jumper or BIOS configuration option or such to enable it?

http://scalibq.wordpress.com/just-keeping-it- … ro-programming/

Reply 68 of 73, by SirNickity

User metadata
Rank Oldbie
Rank
Oldbie

Just got my Model 30 286 and finished the PSU recap. Hard drive is failing (shocking...), but the floppy drive works fine. I'll give this a try as well.

Reply 69 of 73, by canthearu

User metadata
Rank Oldbie
Rank
Oldbie

I have one of these cards.

https://computer-retro.de/Bilder/Grafikkarten … Taiwan-1991.jpg

IRQ2 is not actually connected to the trident chip, but the pin on the ISA connector is routed to the top, along with the IRQ pin from the trident. The traces end at 2 separated via's next to the top middle jumper. So IRQ2 must have been thought of during the design of this board, but decided fairly late on they weren't going to connect it and just broke the trace into 2 parts.

Putting a tiny jumper between the 2 vias reconnects IRQ2 to the trident chip and suddenly IRQs work again.

Reply 70 of 73, by SirNickity

User metadata
Rank Oldbie
Rank
Oldbie

So I tried this on my PS/2 Model 30 286, with onboard VGA. By default, no interrupts. But, I couldn't remember if there were any CLI switches, so I ran it as vretirq /? and, while it didn't show a help screen, it did start counting interrupts and the white bar scrolled up screen about at a rate of about one roll from bottom to top per second. Re-running without the switch went back to 0 interrupts. The switch brought it back. Not sure what to make of that, but hope it helps.

Reply 71 of 73, by mills26

User metadata
Rank Newbie
Rank
Newbie

I'm interested in this to use it in a demo, is there any source code?. I tried to add the interrupt using online docs, but it did not work 🙁.

Thanks.

Reply 72 of 73, by AlexZ

User metadata
Rank Member
Rank
Member

It sounds to me like with the advent of VGA vertical retrace interrupt ceased to be a useful feature and assigning IRQ for VGA card was considered unnecessary when most of the time it would be unused. There was no need to have it enabled by default and as support for it was inconsistent it ended up not being used.

Pentium III 900E, ECS P6BXT-A+, 384MB RAM, NVIDIA GeForce FX 5600 128MB, Voodoo 2 12MB, 80GB HDD, Yamaha SM718 ISA, 19" AOC 9GlrA
Athlon 64 3400+, MSI K8T Neo V, 1GB RAM, NVIDIA GeForce 7600GT 512MB, 250GB HDD, Sound Blaster Audigy 2 ZS

Reply 73 of 73, by xorlof

User metadata
Rank Newbie
Rank
Newbie

Necroing this thread, but if you're still interested in some PS/2 testing, I have an MCGA Model 25 and a VGA Model 25 that I could dig out and test on. I'd expect the VGA to have a similar result to the Model 30 VGA mentioned above, but I guess it couldn't hurt to get confirmation on the behavior...again, if you're still interested. I did try your utility with the most recent dosbox-x release in its default config and the counter remains at 0.