BFG nVIDIA GeForce 7800GS OC on Win98

[Goodmaneuver]

7 hours ago, Feamane said:

Not sure I exactly follow you, are you thinking both motherboards/CPUs/sets of RAM are defective in the same way?

It could be the MBs have a fault that was brought on by a common video card that was not working properly. Power is regulated to many different parts of the MB via the circuitry for example the AGP voltage is 0.8 volts for vs 3.0. An easy check of a single core x64 CPU is as explained. 77.72 is the better driver over the 81.98 as PCI 6200 nVidia cards work with it.

[Feamane]

8 hours ago, deomsh said:

Of cause, but my idea was IF the 77.72 driver gave better performance than with the 8x.xx drivers (both for your 6800), it could be worthwile to 'dig deeper' with the 7xxx AGP cards on Win98 regarding drivers /Registry. 

That makes sense.  I will have to give that a try.

Thanks,

DJ

[Feamane]

5 hours ago, Goodmaneuver said:

It could be the MBs have a fault that was brought on by a common video card that was not working properly. Power is regulated to many different parts of the MB via the circuitry for example the AGP voltage is 0.8 volts for vs 3.0. An easy check of a single core x64 CPU is as explained. 77.72 is the better driver over the 81.98 as PCI 6200 nVidia cards work with it.

OK, interesting theory.  But what kind of hardware fault on the MBs could result in only poor performance on Win98 but good performance on XP with neither showing any stability problems?  For example, I'm used to having to replace bad caps on some of my old MBs--things just start randomly crashing and freezing.  This actually happened last year, we were playing Serious Sam TSE on my LAN and after about six hours of playing one PC started crashing.  After about four times of rebooting and playing ten minutes before another crash we gave up for the night.  Next day I opened up the box and sure enough, caps going bad. Replace the caps and can play many hours of Serious Sam with no more crashing and freezing.  So I don't understand what kind of hardware fault would be so very precise and selective....

I will try 77.72, but I don't have any versions of Windows after XP, I'd have to find some demo version or something....

Thanks,

DJ

[RainyShadow]

Just a wild guess - are you using IRQ sharing in Win98?

If i.e. the video and the network card share the same IRQ, every time there is a network transfer it will slow down the video card.

And maybe XP grouped the video with another less active hardware, or left it alone...

[Feamane]

3 hours ago, RainyShadow said:

Just a wild guess - are you using IRQ sharing in Win98?

If i.e. the video and the network card share the same IRQ, every time there is a network transfer it will slow down the video card.

And maybe XP grouped the video with another less active hardware, or left it alone...

How can I check that?  Although, I'm not doing anything on the network when I'm running the benchmarks, so not sure it would apply.  Still, might be worth looking into.

Thanks,

DJ

[RainyShadow]

3 hours ago, Feamane said:

How can I check that?  Although, I'm not doing anything on the network when I'm running the benchmarks, so not sure it would apply.  Still, might be worth looking into.

The network was just an example, it could be any other device that use an IRQ.

See this, the video shares its IRQ with the USB controller:

[Goodmaneuver]

I sort of made my point but there may be parts of my other posts that may interest you. https://msfn.org/board/topic/181107-windows-98se-and-nforce-2-change-chipset/?do=findComment&comment=1177938 and this topic https://msfn.org/board/topic/181380-question-on-what-to-update-w98-with/?do=findComment&comment=1183509 

 

 

[Feamane]

On 10/3/2020 at 7:25 AM, deomsh said:

Of cause, but my idea was IF the 77.72 driver gave better performance than with the 8x.xx drivers (both for your 6800), it could be worthwile to 'dig deeper' with the 7xxx AGP cards on Win98 regarding drivers /Registry. 

OK, so to summarize the results so far:

ASRock 775V88            P4-550(3400MHz)            GF6800    Win98    FW81.98    3DMark03=7432
ASRock 4CoreDual-SATA2 R2.0    Dual-Core E5300(2600MHz)    GF6800    Win98    FW81.98    3DMark03=4775
ASRock 4CoreDual-SATA2 R2.0    Dual-Core E6800(3330MHz)    GF6800    Win98    FW81.98    3DMark03=5753
ASRock 4CoreDual-SATA2 R2.0    Dual-Core E6800(3330MHz)    GF6800    Win98    FW77.72    3DMark03=7438
ASRock 4CoreDual-SATA2 R2.0    Dual-Core E6800(3330MHz)    GF7800    Win98    FW82.69    3DMark03=5020
ASRock 4CoreDual-SATA2 R2.0    Dual-Core E6800(3330MHz)    GF7800    XP    FW94.24    3DMark03=14436

So, ya, everything else being the same, ForceWare 77.72 is faster than 81.98.  It is just as fast as the old rig with the P4, GF6800 and FW81.98.  But it is still half as fast as the GF7800+FW94.24 on XP.  I will have to try with the 82.69 drivers and the GF6800, I forgot to try that combo.  Interesting....

Thanks,
DJ

[Feamane]

On 10/4/2020 at 12:58 AM, RainyShadow said:

The network was just an example, it could be any other device that use an IRQ.

See this, the video shares its IRQ with the USB controller:

 

  Reveal hidden contents

 

 

 

Ya, the video is sharing the IRQ with the USB and ACPI.  Could be a problem, worth a shot to give the video an IRQ all its own.  Unfortunatly there is no option to turn off ACPI in the BIOS!  I install with the /p i flag set, shouldn't that prevent Win98 from implementing ACPI?

Thanks,
DJ

[Feamane]

On 10/4/2020 at 7:53 AM, Goodmaneuver said:

I sort of made my point but there may be parts of my other posts that may interest you. https://msfn.org/board/topic/181107-windows-98se-and-nforce-2-change-chipset/?do=findComment&comment=1177938 and this topic https://msfn.org/board/topic/181380-question-on-what-to-update-w98-with/?do=findComment&comment=1183509 

 

 

I have not had time to read these links yet but I will as soon as I get a chance.  Maybe there is yet a chance to figure this out and be able to stick with Win98 and still get a 3DMark03 score somewhere near 14436.

Thanks,
DJ

[RainyShadow]

1 hour ago, Feamane said:

Ya, the video is sharing the IRQ with the USB and ACPI

The ACPI entry is just a placeholder, ignore it.

First see if this really matters for your test scores. Disable the USB controller in BIOS settings, check if another device didn't take the same IRQ as your video card, and run the benchmark again.

[Feamane]

I ran the following two tests with the USB disabled so no IRQ sharing with the video:

ASRock 4CoreDual-SATA2 R2.0    Dual-Core E6800(3330MHz)    GF6800    Win98    FW77.72    3DMark03=7435
ASRock 4CoreDual-SATA2 R2.0    Dual-Core E6800(3330MHz)    GF6800    Win98    FW82.69    3DMark03=7589

Thanks,
DJ

[Goodmaneuver]

I do remember installing Dotnet Framework 2.0 having an effect on video. On one ME machine it stopped Glory of the Roman Empire but this was way back in the very early days with not so good hardware. It did seem to accelerate everything and so did Intel(R) Integrated Performance Primitives RTI4.0. It does not matter if you then go to an AMD machine RTI4.0 does not hurt an AMD build. I can not locate the actual file at the moment but it was installed before Framework 2.0 was introduced and gives a good speed improvement.

[deomsh]

12 hours ago, Feamane said:

So, ya, everything else being the same, ForceWare 77.72 is faster than 81.98.  It is just as fast as the old rig with the P4, GF6800 and FW81.98.  But it is still half as fast as the GF7800+FW94.24 on XP.  I will have to try with the 82.69 drivers and the GF6800, I forgot to try that combo.  Interesting....

Speed is your ultimate goal, but what about your problem of max AGP4x? What is shown by the 77.72 Geforce Display property tab?

[RainyShadow]

Maybe try checking with RivaTuner?