VOGONS


VIA zero bytes AGP memory bug | Navigation thread

Topic actions

Reply 60 of 67, by Sly_Botts

User metadata
Rank Member
Rank
Member

Here's what I had to do.

Getting my ASUS A8V-MX to work with my Radeon x800 Pro.

It is possible to commit no errors and still lose. That is not a weakness, that is life.

Reply 61 of 67, by bloodem

User metadata
Rank Oldbie
Rank
Oldbie
Sly_Botts wrote on 2026-09-06, 14:38:

The bug described in this thread affects Windows 98 (not applicable in your case, since you used Windows XP).

2 x PLCC-68 / 4 x PGA132 / 5 x Skt 3 / 1 x Skt 4 / 9 x Skt 7 / 12 x SS7 / 1 x Skt 8 / 21 x Slot 1 / 8 x Slot A
5 x Skt 370 / 8 x Skt A / 2 x Skt 478 / 2 x Skt 754 / 3 x Skt 939 / 7 x LGA775 / 1 x LGA1155
Current PC: Ryzen 7 9800X3D
Backup: Ryzen 7 5800X3D

Reply 62 of 67, by mockingbird

User metadata
Rank Oldbie
Rank
Oldbie
bloodem wrote on 2026-09-06, 16:41:

The bug described in this thread affects Windows 98 (not applicable in your case, since you used Windows XP).

My previous challenge in this thread stands. Thus far, no one has been able to provide any empiricism in that the massive performance advantage of a PCIe video subsystem can't make up for the lack AGP texture acceleration.

I'll be glad if someone can prove otherwise.

mslrlv.png
(Decommissioned:)
7ivtic.png

Reply 63 of 67, by bloodem

User metadata
Rank Oldbie
Rank
Oldbie
mockingbird wrote on 2026-09-06, 17:15:

My previous challenge in this thread stands. Thus far, no one has been able to provide any empiricism in that the massive performance advantage of a PCIe video subsystem can't make up for the lack AGP texture acceleration.

I'll be glad if someone can prove otherwise.

We already know that PCI Express is king on Windows 98 (I have one such system running a Core 2 Duo E8600 and a Radeon X850XT PE and saying that it's fast would be a major understatement).
Still, not really relevant for the VIA AGP bug itself, is it? Or maybe I'm misunderstanding what you mean.

2 x PLCC-68 / 4 x PGA132 / 5 x Skt 3 / 1 x Skt 4 / 9 x Skt 7 / 12 x SS7 / 1 x Skt 8 / 21 x Slot 1 / 8 x Slot A
5 x Skt 370 / 8 x Skt A / 2 x Skt 478 / 2 x Skt 754 / 3 x Skt 939 / 7 x LGA775 / 1 x LGA1155
Current PC: Ryzen 7 9800X3D
Backup: Ryzen 7 5800X3D

Reply 64 of 67, by Sly_Botts

User metadata
Rank Member
Rank
Member
bloodem wrote on 2026-09-06, 16:41:
Sly_Botts wrote on 2026-09-06, 14:38:

The bug described in this thread affects Windows 98 (not applicable in your case, since you used Windows XP).

I plan on giving it a shot over the next few days I'll get back to you if I get anywhere.

It is possible to commit no errors and still lose. That is not a weakness, that is life.

Reply 65 of 67, by Dothan Burger

User metadata
Rank Member
Rank
Member
bloodem wrote on 2026-09-06, 17:25:
mockingbird wrote on 2026-09-06, 17:15:

My previous challenge in this thread stands. Thus far, no one has been able to provide any empiricism in that the massive performance advantage of a PCIe video subsystem can't make up for the lack AGP texture acceleration.

I'll be glad if someone can prove otherwise.

We already know that PCI Express is king on Windows 98 (I have one such system running a Core 2 Duo E8600 and a Radeon X850XT PE and saying that it's fast would be a major understatement).
Still, not really relevant for the VIA AGP bug itself, is it? Or maybe I'm misunderstanding what you mean.

I agree, the only GPUs better in AGP than PCI-E are GeForce FX cards.

Reply 66 of 67, by Sly_Botts

User metadata
Rank Member
Rank
Member

So my motherboard doesn't seem to have this issue.

Asus A8v-MX.
VIA K8M800

I have win98 se running with ATI x800 pro. 3dmark tests running smoothly.

It is possible to commit no errors and still lose. That is not a weakness, that is life.

Reply 67 of 67, by Geri

User metadata
Rank Member
Rank
Member
Bruno128 wrote on 2023-08-14, 12:04:

collectively known as "VIA 0 bytes AGP memory bug" or "VIA AGP GART bug".

This is not a bug. Its a feature.
Manufacturers did this intentionally: few years into the standard, nerfing basically all AGP features (not just agp memory but the rest of the special features too).
The reason for this is, they never worked as intended in the existence of agp standard, and causing stability or severe performance issues.
Eventually the manufacturers realized they have to use whitelists if they wanted to safely allow certain features.
As these features in practice bring almost nothing to the table, they stopped caring, and AGP as a standard turned into merely a bandwidth increasing standard.
Under linux, they didn't had to remove too much of them, because overwhelming majority of them never actually got implemented.
Also most of the AGP 8x motherboards will outright run 4x in secret, or even if the transfer rate is reporting 8x, all the new 8x features will be disabled, otherwise the system would be hard crash in every minute or so.
For example with the previously mentioned 775dual-vsta, you have to manually disable certain agp features in the bios (such as agp power calibration) to have agp 8x mode even work with certain newer graphics cards, or limit the agp speed to 4x. Not to mention you frequently have to test like 3-4 graphics drivers to find one which works on these newer via chipsets, because the chipset is quite buggy (the pci-e implementation is buggy here as well, but thats offtopic now).
In short, you can safely ignore the missing features and be happy it even works, and operates without stability issues.
When i used a 3850 agp in my 775dual-vsta i also had to play with every settings, and block certain features in the bios till i finally had agp8x working properly without stability issues, but agp 4x mode worked fine out of the box, in retrospect i would just simply set the speed to agp 4x and not bother with it.

TitaniumGL the OpenGL to D3D wrapper:
http://users.atw.hu/titaniumgl/index.html