Patches Proceed For Disabling Radeon AGP GART, Deprecating TTM AGP

[phoronix]

Phoronix: Patches Proceed For Disabling Radeon AGP GART, Deprecating TTM AGP

Several days back was the proposal to "remove AGP support" from Radeon/Nouveau/TTM. This did formulate into a set of patches that would disable the AGP mode in the Radeon driver and deprecate the AGP code in TTM memory management. However, as was pointed out in the ensuing discussion, AGP graphics cards will still be operable on Linux with this level of deprecation by using the PCI GART mode...

http://www.phoronix.com/scan.php?page=news_item&px=Radeon-AGP-GART-Disabling-Patch

[DanL]

Nitpick: The Abit AW9D mobo in the picture doesn't have an AGP slot.

[Michael]

Nitpick: The Abit AW9D mobo in the picture doesn't have an AGP slot.

Memory failed me when digging through old pictures....

[schmidtbag]

Might be interesting to see what the performance loss of PCI GART is. No matter what it is, I'd say that's better than having no support at all.

[DeepDayze]

My old IBM Thinkpad T42 has a Radeon 7500 Mobility chip which runs off AGP so this proposed patch will certainly reduce the performance of this chip so the performance hit may or may not be worth it. Benchmarks should be done if Michael has any old AGP boards and graphics cards to do such a test.

[kylew77]

I hope somebody like NetBSD keeps the AGP support so there is a high performant OS for these old cards.

[oiaohm]

Might be interesting to see what the performance loss of PCI GART is. No matter what it is, I'd say that's better than having no support at all.

My old IBM Thinkpad T42 has a Radeon 7500 Mobility chip which runs off AGP so this proposed patch will certainly reduce the performance of this chip so the performance hit may or may not be worth it. Benchmarks should be done if Michael has any old AGP boards and graphics cards to do such a test.

I hope somebody like NetBSD keeps the AGP support so there is a high performant OS for these old cards.

All of you have the same idea that it going to be loss in performance. PCI on a AGP is still a gets the clock speed boosts on transfer of AGP. In the case of Radeon 7500 Mobillity you still have the agp x4 transfer speeds in PCI mode so it not that crippling on the card performance.

Due to removing the system wide MMU conflit AGP GART causes when you use PCI GART you may in fact see higher performance with the AGP GART feature removed.

So the question is how many with AGP cards suffer performance losses and how many get performance gains and how many come in at nothing changed at all. Yes it going to be a true mixed bag outcome. There is one universal thing here the PCI GART mode is stable due to not random-ally butting in on the system wide MMU and the AGP GART is not so reduced graphic card caused issues by the change.

So stability will gain and the performance outcome is kind of going to be all over the place.

NetBSD still depends on the upstream Linux drivers for this stuff.

[Setif]

My old IBM Thinkpad T42 has a Radeon 7500 Mobility chip which runs off AGP so this proposed patch will certainly reduce the performance of this chip so the performance hit may or may not be worth it. Benchmarks should be done if Michael has any old AGP boards and graphics cards to do such a test.

Ubuntu 20.04 : April 2025

[eydee]

My old IBM Thinkpad T42 has a Radeon 7500 Mobility chip which runs off AGP so this proposed patch will certainly reduce the performance of this chip so the performance hit may or may not be worth it. Benchmarks should be done if Michael has any old AGP boards and graphics cards to do such a test.

Do you actually use it for gaming?

[starshipeleven]

My old IBM Thinkpad T42 has a Radeon 7500 Mobility chip which runs off AGP so this proposed patch will certainly reduce the performance of this chip so the performance hit may or may not be worth it. Benchmarks should be done if Michael has any old AGP boards and graphics cards to do such a test.

I'm going on a limb and take a bind stab in the dark with the following statement:
That graphics chip is complete garbage and can't saturate PCI bandwith anyway (which is what this patch falls back the bandwith to), it would be hard to notice any difference.