VOGONS


First post, by WhatANerd

User metadata
Rank Newbie
Rank
Newbie

Hey everyone. Ok, so I have two 486 systems, both using VLB graphics. One of them uses a Diamond Viper Pro with 2MB of RAM, and I am able to use 16 million colors at 800x600. The other uses a SIIG (CL5429-86Q) that I just upgraded from 1MB to 2MB of RAM. I've verified that the system can see the newly added RAM, but I do not have a 16 million color option at 800x600 resolution with this card - the highest I can get is 64K at 800x600.

The SIIG manual actually says that the maximum of 2MB will only allow for 64K colors at 1024x768 (instead of 256 with 1MB). I get the same results using MS-DOS/W3.x, Windows 95, and NT 3.51. Actually, in NT I cannot get more than 256 colors at 800x600.

What's the deal? Why can the Viper do it with 2MB while the CL5429 cannot? Is there a BIOS setting I need to look into? I don't get it.
shrug.gif

  • x86: Tandy 1000RL (HD+768K), Tandy 3000HD, 486DX33 VLB, 486DX50 VLB, Packard Bell Force 1998CDT (Pentium 133)
  • 68K: Mac Plus 1MB (early), Quadra 700 (2), Quadra 950, Quadra 650

Clock multiplication is too new for me, as you can see!

Reply 2 of 12, by Anonymous Coward

User metadata
Rank l33t
Rank
l33t

Cirrus Logic made budget graphics chips. I remember the 2MB upgrade on a lot of the low end cards being almost useless. I think the ARK1000 cards are like this as well. But you're right, 2MB should be enough to do 16 million colours at 800x600. My advice would be to upgrade to a better card. The CL 5434 for example.

"Will the highways on the internets become more few?" -Gee Dubya
V'Ger XT|Upgraded AT|Ultimate 386|Super VL/EISA 486|SMP VL/EISA Pentium

Reply 3 of 12, by Matth79

User metadata
Rank Oldbie
Rank
Oldbie

http://www.vgamuseum.info/images/stories/doc/ … s/cl-gd542x.pdf
Looks like the only thing 1 to 2MB gets is 256C @ 1280x1024 it only does 16M at VGA 640x480

Unclear if this generation of cards did anything clever with extra memory

Reply 4 of 12, by konc

User metadata
Rank l33t
Rank
l33t
Matth79 wrote:

Unclear if this generation of cards did anything clever with extra memory

Until now I thought that resolution/color setting is just a matter of memory and drivers. Are we sure that it's not a matter of available drivers and that really some cards can't output (by design restrictions) every possible color setting at supported resolutions that memory permits? If this is the case, then I also want to learn more about the quoted sentence above.

Reply 5 of 12, by WhatANerd

User metadata
Rank Newbie
Rank
Newbie
konc wrote:
Matth79 wrote:

Unclear if this generation of cards did anything clever with extra memory

Until now I thought that resolution/color setting is just a matter of memory and drivers. Are we sure that it's not a matter of available drivers and that really some cards can't output (by design restrictions) every possible color setting at supported resolutions that memory permits? If this is the case, then I also want to learn more about the quoted sentence above.

That's what I've always thought, too. This is really puzzling. The strange part is that in NT 3.51, I get a maximum of 256 colors at 800x600, but in Win 95 and DOS/3.x, I get 64K colors at 800x600, even though I should be getting 16 million on all three operating systems.

So it makes me wonder, is it a driver issue, or is the OS imposing some sort of limitation on the card's memory space, or what? Why would the driver not support more colors at higher resolutions? 0onfused.gif

Matth79, thanks so much for the link! That's really good info. If it indeed is a limitation of the chip itself, that's just so ridiculous! Oh well, we take what we can get. orangeno.gif

  • x86: Tandy 1000RL (HD+768K), Tandy 3000HD, 486DX33 VLB, 486DX50 VLB, Packard Bell Force 1998CDT (Pentium 133)
  • 68K: Mac Plus 1MB (early), Quadra 700 (2), Quadra 950, Quadra 650

Clock multiplication is too new for me, as you can see!

Reply 6 of 12, by sliderider

User metadata
Rank l33t++
Rank
l33t++

My ATI Graphics Pro Turbo boasts 16 million colors at 1280 x 1024 with 4mb RAM but according to the box uses a dithering technique at that resolution. It will do it without resorting to any trickery in 1024 x 768 though. The box for my Graphics Xpression says it can do 16 million colors at 800 x 600 with 2mb.

Reply 7 of 12, by Anonymous Coward

User metadata
Rank l33t
Rank
l33t

The Windows NT colour limitation is almost certainly a driver issue. Are you using the driver that came with NT? If so, get a 3rd party driver. Maybe the pack in driver is one that supports a broad range of CL cards and only supports common features.

"Will the highways on the internets become more few?" -Gee Dubya
V'Ger XT|Upgraded AT|Ultimate 386|Super VL/EISA 486|SMP VL/EISA Pentium

Reply 8 of 12, by obobskivich

User metadata
Rank l33t
Rank
l33t
WhatANerd wrote:
That's what I've always thought, too. This is really puzzling. The strange part is that in NT 3.51, I get a maximum of 256 color […]
Show full quote

That's what I've always thought, too. This is really puzzling. The strange part is that in NT 3.51, I get a maximum of 256 colors at 800x600, but in Win 95 and DOS/3.x, I get 64K colors at 800x600, even though I should be getting 16 million on all three operating systems.

So it makes me wonder, is it a driver issue, or is the OS imposing some sort of limitation on the card's memory space, or what? Why would the driver not support more colors at higher resolutions? 0onfused.gif

Matth79, thanks so much for the link! That's really good info. If it indeed is a limitation of the chip itself, that's just so ridiculous! Oh well, we take what we can get. orangeno.gif

Why is it puzzling? It sounds like it's behaving exactly as the documentation says it will - a maximum of 65k color output (based on your original post). Resolution/color-depth is not exclusively related to how much RAM you have - just like my GeForce FX with 128MB cannot do 30-bit color, despite having enough memory for it (and is that "so ridiculous"?).

The limit in 3.51 is likely OS/driver related though - I'd just pick a different operating system or card if you need higher resolutions and color depths for your application (either under 3.51 or in general).

Reply 10 of 12, by obobskivich

User metadata
Rank l33t
Rank
l33t
leileilol wrote:

I've used a FX5200 128mb in 1920x1080x32 before just fine. Is there something i'm missing here?

😊 I said 30-bit color, as in billions of colors ("deep color")."32-bit color" refers to 24-bit + 8-bit alpha; you're still only drawing from a palette of 16.7M colors ("true color"). Very few consumer cards/devices can deal with 30-bit, even though many of them since at least the GeForce FX era have had sufficient memory to do so - in many cases it's both a hardware and software limitation. A notable exception to this is the Matrox Parhelia, which has 10-bit controllers/RAMDACs. Various professional cards, especially more recent models, also support the feature. True 10-bit LCDs are also relatively rare, and generally quite expensive too.

My point was to use it as an example - there are cards with more memory than a Parhelia (like FX 5950, for example), but that do no support the same resolution/color-depth/etc modes due to hardware and/or software limitations.

Last edited by obobskivich on 2014-07-11, 01:48. Edited 1 time in total.

Reply 11 of 12, by WhatANerd

User metadata
Rank Newbie
Rank
Newbie

Yeah, I guess so. It's just puzzling because I've just never heard of or seen such a thing, and I've been around my fair share of systems. I had a few cheapie VLB cards in the '90s and they were all able to do 24bit at 800x600. One was a Diamond Speedstar Pro, which as a poor kid I recall being more of a low-end card. Hmm, actually I just looked it up and the Speedstar Pro VLB used a CL5428! I must have had the Speedstar Pro SE then, which used a CL5430.

I just figured that any VLB card would (should!) be able to pull it off, especially since the box brags about 16.7 million colors and supporting a 50MHz bus speed. Good thing I got this SIIG card NIB for free; had I paid for it back in the '90s I'd be one unhappy customer! "SIIGnificantly better" my butt!

I've been hunting for NT 3.x drivers for the 5429 to no avail, unfortunately.

  • x86: Tandy 1000RL (HD+768K), Tandy 3000HD, 486DX33 VLB, 486DX50 VLB, Packard Bell Force 1998CDT (Pentium 133)
  • 68K: Mac Plus 1MB (early), Quadra 700 (2), Quadra 950, Quadra 650

Clock multiplication is too new for me, as you can see!

Reply 12 of 12, by obobskivich

User metadata
Rank l33t
Rank
l33t
WhatANerd wrote:

Yeah, I guess so. It's just puzzling because I've just never heard of or seen such a thing, and I've been around my fair share of systems. I had a few cheapie VLB cards in the '90s and they were all able to do 24bit at 800x600. One was a Diamond Speedstar Pro, which as a poor kid I recall being more of a low-end card. Hmm, actually I just looked it up and the Speedstar Pro VLB used a CL5428! I must have had the Speedstar Pro SE then, which used a CL5430.

I just figured that any VLB card would (should!) be able to pull it off, especially since the box brags about 16.7 million colors and supporting a 50MHz bus speed. Good thing I got this SIIG card NIB for free; had I paid for it back in the '90s I'd be one unhappy customer! "SIIGnificantly better" my butt!

I've been hunting for NT 3.x drivers for the 5429 to no avail, unfortunately.

It's honestly not surprising to see such a limitation on a Cirrus card; I remember having a PCI-based model in the mid 1990s that had pretty limited color/resolution options (if I remember right it was limited to 15-bit (32k) over 800x600 or something odd like that). It was functional/stable, but not the greatest in terms of quality.

Don't think I can be much help with NT 3.51 - I've never had good luck with that operating system. 😒