VOGONS


First post, by 386SX

User metadata
Rank l33t
Rank
l33t

Hello,

I'm testing early 3D accelerators and even using latest S3 Virge FastD3D drivers, it looks like Quake based games like Half Life or Thief titles miss most textures with white flat walls and floors. Is this something expected or it should work in Direct3D even without a wrapper where OpenGL is supported? Thief for example is a Direct3D only game and shows the same problem of Half Life while it looks like it might run at acceptable frame rate at 400x300.
Are there any better drivers than the 4.10.01.2122

Thanks?

Reply 1 of 16, by leileilol

User metadata
Rank l33t++
Rank
l33t++

Virge doesn't have blending functions so yes, broken graphics are expected, such as with completely unblended lightmaps. This was 'fixed' in Trio3D 2X IIRC.

Last edited by leileilol on 2022-01-14, 11:35. Edited 1 time in total.

apsosig.png
long live PCem

Reply 2 of 16, by 386SX

User metadata
Rank l33t
Rank
l33t

Interesting, thanks! It means I'll find again a Trio3D/2X card to see how it runs.. 😁 It has been my first 3D accelerator of the 1998 and lasted only a couple of week for this reason that forced me to play most games in software rendering. Back then I suppose I needed a driver upgrade I could not have not having even internet.

Reply 6 of 16, by 386SX

User metadata
Rank l33t
Rank
l33t

Thanks. 😉

I remember the wrapper and I wonder how it worked if S3 Direct3D support was actually designed for early 3D rendering and not supporting such feature. I mean beside the very low frame rate (I think to remember it ran GLQuake something like 4-5fps if I'm correct) as expected by any wrapper, still the general GLQuake final accuracy was enough good and in fact first time I saw that I couldn't believe that it was actually working. GLQuake translated to S3D did texturing in a different way without the blending functions compared to the later game that used the same engine supporting D3D? Did S3D API had more functions than the Direct3D Virge/Trio3D drivers could support?

Reply 7 of 16, by leileilol

User metadata
Rank l33t++
Rank
l33t++

The S3D wrapper tries to approximate the lightmap's blending by reprocessing it as an alpha texture, similar to how PowerVR Extreme's MiniGL works for PowerVR PCX2 cards (which don't have blending functions either, but Videologic relatively had a lot more care than S3)

As for Half-Life... Half Life 1 textures on S3 Virge

apsosig.png
long live PCem

Reply 9 of 16, by Putas

User metadata
Rank Oldbie
Rank
Oldbie
386SX wrote on 2022-01-14, 12:24:

Does the native Direct3D with console tweaks results in better or worse performances compared to the S3D wrapper both with textures appeared?

This is hard to parse.

Reply 10 of 16, by 386SX

User metadata
Rank l33t
Rank
l33t
Putas wrote on 2022-01-15, 04:07:

This is hard to parse.

I meant with the textures visible it's clearly not my first language and I was using the verb "appear" to say that in both cases with Direct3D with the console tweaks or in S3D with the wrapper the textures should be visible at last but I was wondering the speed impact. 😉

Reply 11 of 16, by leileilol

User metadata
Rank l33t++
Rank
l33t++

There'd still be an impact since it's trying to draw both the textures and the lightmap (and picking the unblended lightmap). ViRGE doesn't have overdraw mitigation like multitexturing, tiled rendering and HSR.

apsosig.png
long live PCem

Reply 12 of 16, by 386SX

User metadata
Rank l33t
Rank
l33t

Interesting thanks! I'm not surprised by these limitations from such an early design for some of the lighter early games and might have needed more S3 native games/patches to run better but I do respect this chip history. It'd have been interesting if at least reached the Rage IIc rendering results. While just as slow this last surprised me for the compatibility but maybe more comparable to a Trio3D/2x.

Reply 13 of 16, by 386SX

User metadata
Rank l33t
Rank
l33t

I was looking some youtube test of the Trio3D and I wasn't expecting it to be quite faster and compatible of the Virge previous chips. Which were the fastest versions of the Trio3D, like any chip revisions, brands and clocks?

Reply 15 of 16, by LubieCipy

User metadata
Rank Newbie
Rank
Newbie
386SX wrote on 2022-01-15, 16:13:

I was looking some youtube test of the Trio3D and I wasn't expecting it to be quite faster and compatible of the Virge previous chips. Which were the fastest versions of the Trio3D, like any chip revisions, brands and clocks?

Are you asking about Trio3D or Trio 3D/2X?

Reply 16 of 16, by 386SX

User metadata
Rank l33t
Rank
l33t
LubieCipy wrote on 2022-01-15, 19:56:
386SX wrote on 2022-01-15, 16:13:

I was looking some youtube test of the Trio3D and I wasn't expecting it to be quite faster and compatible of the Virge previous chips. Which were the fastest versions of the Trio3D, like any chip revisions, brands and clocks?

Are you asking about Trio3D or Trio 3D/2X?

I've read the great vintage3d.org review on this chip I never tested enough when I had it and understood the /2X SDRAM 8MB 100Mhz versions should be mostly the one I'll look for. I've found the exact model I had in the 1998 and it was a Trio3D standard 4MB probably SGRAM but I'll go for the latest version to test.