VOGONS


Reply 60 of 91, by havli

User metadata
Rank Oldbie
Rank
Oldbie
ruthan wrote:

3) I dont thing that something like card requiring DX9 exists, DirectX is backward compatible..

Yes, there is... newer cards like GeForce 6/7 for example will not work properly without high enough version of directX. Or to be more precise their drivers won't. Once I tried to run GeForce 7800 GS AGP on Athlon Thunderbird PC.... D3D didn't work because there wasn't recent enough build of dx9c installed and I couldn't install it because it required SSE (which Tbird lacks).

HW museum.cz - my collection of PC hardware

Reply 61 of 91, by BushLin

User metadata
Rank Member
Rank
Member
ruthan wrote:
Nvidia drivers version influence in Windows 98: Tested with 1 GHz PIII, so its hard to say with it measn for SS7, but.. Expendab […]
Show full quote

Nvidia drivers version influence in Windows 98:
Tested with 1 GHz PIII, so its hard to say with it measn for SS7, but..
Expendable(Direct3D-640x480x16)
- TNT2 - 3% decrease up to driver 61.76, so its virtually nothing, because some bugfixes degraded performance.
- Geforce 256(=1) /G2 - there is huge gradual performance decrease - from 95 to 52 FPS in 71.XX driver.
- Geforce 4 - there not results for 7x and 8x driver, but between 28x32 and 61.76 si decrease few percent, and
its 30% difference between G4 and older cards with 61.76 driver.
- Geforce 5800 / 5950 there is decrease i would say 15% with 71.xx driver, but in 81.98 its back and otherwise difference is < 10%, so no big deal.
Expadable(Direct3D-1024x768x32)
- TN2 is the same
- Geforce 1 /2 - its very similar again huge decrease.
- Geforce 4 - now there is big difference from 77 fps to 50 Fps.
- Geforce 5/5 - except bad 71.xx thre is i would say 5% difference.

Expendable(Direct3D-1600x1200x32) // Very similar to 1024x768.
Drakan 640x480x16 / 1024x768x32- very similar to Expendable 1024x768
Drakan 1600x1200 - simlar to Expendable 640x480
Unreal Tournament - 640x480x16 - there are same gaps for all cards except bat GF4 performance with 56.64.. its
gradual, but from oldest, fastest to newest its 15%.
Unreal Tournament - 1024x768- TN2 is slightly gradually faster, other otherwise its similar to 640x480 but over difference is even smaller - 10%.
Quake 2 640x480x16 (OpenGL) - there not big difference -/+ 10, for Geforce 4 and FX even less.
Quake 2 1024x768x32 (OpenGL) - +/- 15% - best driver is 30.82,, for Geforce 4 and FX even less.
Quake 2 1600x1200x32 (OpenGL) - +/- 15% - best driver is 30.82,, for Geforce 4 and FX minimal gap.
Quake 3 640x480x16 (OpenGL) - except TNT2 where is 10% last its in few percent +/- here we can see for which game where drivers probably optimized, or great IDsoft code, i dunno, , except early 6.31 and older drivers.
Quake 3 1024x768x32 +/- few percents, except early 6.31 and older drivers.
Quake 3 1600x1200x32 - +/- few percents, except early 8.05 and older drivers.
MDK2 (Direct3D) 640x480x16 - there is 15 decrease from 45.23 driver to newest, there is small gain from 56.64 for G4/ FX.
MDK2 1024x768x32- except 6.31 drivers and earlier, there is max 10% difference, less with G4/G5.
MDK2 1600x1200x32 - - except 8.05 drivers and earlier, there is max 10% difference, less with G4/G5.
Serious Sam - 640x480 - decrease for TNT2, other cards are ok.
Serious Sam - 1024x768 - decrease for TNT2, other cards are ok in +-5%
Serious Sam - 1600x1200 - +- 15%, except 81.92 there where FX 5800 an Geforce 4 are 25% slower, G 5900 is only few
percent slower.

Overall Nvidia cards performance:
- This video shows that newer Geforce cards sucks even with much faster 1 GHz PIII with Win98, so its not primary problem of slow SS7 cpu.
- This also first and so far best of show, how bad performance of G4 and G5 cards are in Windows 98, regardless of drivers, because their performance is not degrade too much with drivers as for other cards.
- Even these much faster cards, are often slower or at same level as much older cards.
- There's some Direct3D games where are Geforce 5 cards in 1/3 slower than older cards, in OpenGL games are usually slightly slower, but rarely faster, some Direct 3D games is diffence smaller 15%, not 30%.

Between Geforce 5 and 6 is quite big difference in raw power, so that still could be fast, but after this analysis i believe in it a bit less, but its about measurement not believe.. Nothing is changing for Radeon 9700, as shows XP video it has lots of horse power, now we need to see, how bad are Windows 98 drivers and how much is limited by CPU.

You've clearly done much testing or research but I'm struggling to understand much of the context, you've spent so much time doing the collecting it'd be nice to have the numbers posted or just some solid conclusions with regards to the name of the thread with something to verify them.

ruthan wrote:

...
[*] Fastest tested cards - Geforce 4 TI 4200 (slowest from Tis G4) and Geforce 3 200 Ti (slowest from Tis G3), are fastest cards on paper and fastest cards in test, even they are using more modern drivers that rest of Nvidia cards. Phil excluded them from non most of Nvidia testing comparision, because his point of view - there are too modern, not period correct etc.. and they have Incoming problem. So i dont really see reason why would be more fast cards with a bit more modern drivers like Geforce 5/6 faster, so far faster card, better result and there is not such big jump in drivers version.
...

This jumped out at me, the idea that it might not be the generational difference as much as how wide the GPU is; that the SS7 platform can only keep so many graphics pipelines running without it hurting performance, whether that's caused by the CPU, FSB, chipset or just AGP bandwidth.
I guess this is where I'd like to see some numbers as that alone might answer the thread title "Fastest AGP video card for Super socket 7 platform?"

Personally, I don't care. I do care about the people searching for answers on this forum.

Screw period correct; I wanted a faster system back then. I choose no dropped frames, super fast loading, fully compatible and quiet operation.

Reply 62 of 91, by ruthan

User metadata
Rank Oldbie
Rank
Oldbie
BushLin wrote:

You've clearly done much testing or research but I'm struggling to understand much of the context, you've spent so much time doing the collecting it'd be nice to have the numbers posted or just some solid conclusions with regards to the name of the thread with something to verify them.

Its just summary from Phils video, 2,3 post above this post..

BushLin wrote:

This jumped out at me, the idea that it might not be the generational difference as much as how wide the GPU is; that the SS7 platform can only keep so many graphics pipelines running without it hurting performance, whether that's caused by the CPU, FSB, chipset or just AGP bandwidth.
I guess this is where I'd like to see some numbers as that alone might answer the thread title "Fastest AGP video card for Super socket 7 platform?"

Its not about Socket 7, because Socket 370 with 1GHz with Win98, have exactly same issues, at least with newer Nvidia cards.. any other AGP cards are too slow to be better, with only exception and its ATI 9500-9850.. its probably last chance to find something better than Geforce 3.
With same machine Win98 and Windows 2000 / XP tests we can discover if simply about drivers regardless of OS, or its about combination of drivers and OS - Nvidia drivers for Win98.

Im old goal oriented goatman, i care about facts and freedom, not about egos+prejudices. Hoarding=sickness. If you want respect, gain it by your behavior. I hate stupid SW limits, SW=virtual world, everything should be possible if you have enough raw HW.

Reply 63 of 91, by ruthan

User metadata
Rank Oldbie
Rank
Oldbie
havli wrote:

Yes, there is... newer cards like GeForce 6/7 for example will not work properly without high enough version of directX. Or to be more precise their drivers won't. Once I tried to run GeForce 7800 GS AGP on Athlon Thunderbird PC.... D3D didn't work because there wasn't recent enough build of dx9c installed and I couldn't install it because it required SSE (which Tbird lacks).

Its Win98 related or only XP? I never heards that DirectX 9.0c is requiring SSE, but its possible.

Im old goal oriented goatman, i care about facts and freedom, not about egos+prejudices. Hoarding=sickness. If you want respect, gain it by your behavior. I hate stupid SW limits, SW=virtual world, everything should be possible if you have enough raw HW.

Reply 64 of 91, by ruthan

User metadata
Rank Oldbie
Rank
Oldbie

Someone tested Radeon 9800 AGP XT with Win98 and 1.4 GHz P3, its not SS7, but it could give some more info: // This one is AGP 4x+ but it can give us idea about performance of these cards.
ATi Radeon (Gigabyte) 9800 XT, quick test

I also looked on some this card benchmark and it seems that it excels in OpenGl in pure theory OpenGL could have lesser overhead CPU overhead than DX..

Im old goal oriented goatman, i care about facts and freedom, not about egos+prejudices. Hoarding=sickness. If you want respect, gain it by your behavior. I hate stupid SW limits, SW=virtual world, everything should be possible if you have enough raw HW.

Reply 65 of 91, by ruthan

User metadata
Rank Oldbie
Rank
Oldbie

I just tested Radeon 9000, its seems to be working with Windows 98 and this MB, its slowest of such cards, i has halt of Vertex shader units than Radeon 8500 and and even slower clocks (gpu - 200 / 275 vmem - 500/550), number of vertex shaders units is the same (4).
Performance for just slimmer (it was cheap so i bough it for platform test) - seems to be nice Q3 - no level of V3 2000 - 28 FPS with my 800x600 settings, G3 Ti 200 - with same test give me - 37 FPS, i tested also Vampire Red. demo, it has not benchmark but performance was i thing 1/4 to 1/3 slower than with G3.. and these are OpenGL games, to with Direct3D it would be probably better..
Big question now really if with faster Radeon card would performance even with slow SS7 CPU grow or not.. Just for info Radeon 9800 which has AGP 3.3V variant too, has 4 Vertex shaders and 8 pixel shaders units and clocks are gpu - 325 / mem - 620, difference is also memory bus (128 vs. 256 bit).

Im old goal oriented goatman, i care about facts and freedom, not about egos+prejudices. Hoarding=sickness. If you want respect, gain it by your behavior. I hate stupid SW limits, SW=virtual world, everything should be possible if you have enough raw HW.

Reply 66 of 91, by meljor

User metadata
Rank Oldbie
Rank
Oldbie

I will talk about Nvidia cards, you can translate to Radeon counterparts..

I think a TNT2 is a good match for a (fast) ss7 platform speedwise. The platform is not even fast enough to maximize the performance of a Geforce1!

However, I use a Geforce2 in mine (next to a v2 sli setup which is also overkill) just because of the fact that older TnL supporting games can run on this system while it was simply too slow to run it without TnL.
But basicly it is overkill so talking about geforce3/4/5 or counterparts from Ati is simply way too much for the system.

I did some tests a long time ago with ss7 and 3dmark99 and basicly a Matrox G400max was upon the quickest and also 3dfx cards ran nice. Anything after the TNT2 nvidia-wise was a tiny bit slower and gave no advantage at all. Sure, extreme high res. will make a difference but for those games you will need a faster cpu anyway.

I like some overkill but I do not want to stress the old agp port on a ss7 too much as they were not designed to give a lot of current to the cards. Ofcourse cards with external power source do not have that issue but they are usually way too new/fast imho.

For comparison: I had an Athlon 1ghz system back in the day with a Geforce2 GTS that ran all the new games perfectly. A ss7 cpu is a long long way from the performance of an Atlon 1ghz....

asus tx97-e, 233mmx, voodoo1, s3 virge ,sb16
asus p5a, k6-3+ @ 550mhz, voodoo2 12mb sli, gf2 gts, awe32
asus p3b-f, p3-700, voodoo3 3500TV agp, awe64
asus tusl2-c, p3-S 1,4ghz, voodoo5 5500, live!
asus a7n8x DL, barton cpu, 6800ultra, Voodoo3 pci, audigy1

Reply 67 of 91, by ruthan

User metadata
Rank Oldbie
Rank
Oldbie

I got Radeon 9800 SE its not fastest of this line, but i thing that it will show as if this card does make sense or not.

Im old goal oriented goatman, i care about facts and freedom, not about egos+prejudices. Hoarding=sickness. If you want respect, gain it by your behavior. I hate stupid SW limits, SW=virtual world, everything should be possible if you have enough raw HW.

Reply 68 of 91, by ruthan

User metadata
Rank Oldbie
Rank
Oldbie

Unfortunately my Radeon 9800 SE has classic dead memory problem, so i could do only DOS test - i could be wrong, but my guess is that Windows 98 would be much more dead memory chip sensitive, in DOS i just ignore artefacts..
https://docs.zoho.com/sheet/published.do?rid= … 49606de568c62b9

For this is not cards from performance view worse than Voodoo 3, its bettter from performance point of view even on SS7, but we care about Win98 results.

Im old goal oriented goatman, i care about facts and freedom, not about egos+prejudices. Hoarding=sickness. If you want respect, gain it by your behavior. I hate stupid SW limits, SW=virtual world, everything should be possible if you have enough raw HW.

Reply 69 of 91, by ruthan

User metadata
Rank Oldbie
Rank
Oldbie

I got 9500 Pro, from my testing is slower than Geforce 3, so advantage is that it has DVI, not only analog output.

I would like to do some proper Win98 benchmarking, but i really dont know which game have proper inbuild benchmarks a make sense to them, it seems so lots games tested by Phil lack of it, so its hard to compare results. I already made thread about it quite long time ago:
Do Expandable / Drakan / Serious Sam 1 have inbuild benchmarks?

I still have 9800 Pro on the way its fastest one with 256 bit bus, but i would expect performance similar to Geforce 3.

So what can test?
Quake1
Quake2
Quake III
Expandabl
Unreal tournament - UTbench

What next?

Im old goal oriented goatman, i care about facts and freedom, not about egos+prejudices. Hoarding=sickness. If you want respect, gain it by your behavior. I hate stupid SW limits, SW=virtual world, everything should be possible if you have enough raw HW.

Reply 70 of 91, by ruthan

User metadata
Rank Oldbie
Rank
Oldbie

I finally did Radeon 9800 Pro benchmarks..

K6-2-550 underclocked to 500 - 5x100 ALI V, i did it for stability, but it neccesary. 512 MB RAM.
- Nvidia driver 21.83 for Geforce 3.
- ATI Catalyst 4.3
- 3Dfx 1.07

Quake 3 demo, Demo001:
- I can share config if someone need it, its 800x600x 16 bit, 2/3 texture quality
R9800 Pro Quake 3 - Catalys 6.2 - 23.2
R9800 Pro Quake 3 - Catalys 4.3 - 23.7 // Tried Q3 full gog - Four demo 20 FPS
R9800 Pro Quake 3 - Catalys 4.3 640x480 - 24.2
G3- 200 - Driver 21.83 - 640x480 - 26.2
G3- 200 - Driver 21.83 - 800x600 - 25.9
V3 2000 - 640x480 - 21.5
V3 2000 - 800x600- 21.5
V3 2000 - 800x600- 20.9
V3 2000 - 1280x1024 - 19.6
V3 2000 - 1600x1200 - 16.1

Q2 demo 3.14:
its not working with all card on my machine ATI only with Catalyst 4.3 drivers,
na with 6.2. I dunno problem could be that it was set 1600x1200 as default..
mapa je Q2demo1.dm2 nebo jen demo1.dm2
R9800 Pro 320x200 - 31 // I tried to full game 3.20 - its 33
R9800 Pro - 640x480 -31 // I tried to full game 3.20 - its 33
R9800 Pro -800x600 - 31
R9800 Pro -1024x768 -31
R9800 Pro -1600x1200 - 31
G3 Ti 200 - 320x240 - 42
G3 Ti 200 - 640x480 - 42
G3 Ti 200 - 1024x768 - 42
G3 Ti 200 - 1600x1200 - 42
V3 2000 - 640x480 - 28 FPS
V3 2000 - 800x600 - 28 FPS
V3 2000 - 1024x768 - 28 FPS
V3 2000 - 1600x1200 - 26 FPS

Expandable demo:
- 800x600 not checbox enabled, command: go -timedemo, its endless 9 minutes.``
R9800 PRO after long time 25 FPS, after few minutes 22.
G3 Ti 200 - 23 FPS, but other timedemo was started, or there is some bug in script and demo..
V3 2000 - 15.8

Unreal Tournament gog -4.36 - UT bench, Direct 3D, High quality for all, 16bit
- timedemo 1; next command - demoplay utbench, 640x480 i gettim freeze when i try change resolution :
R9800 Pro - 14 fps
G3 Ti 200- 15 FPS
Voodoo 3 - 15 FPS

I would be nice to have Matrox G400 Max for testing.

Some comments:
- If you can overclock or get better CPU it really matters, because all reasons are CPU caped.
- Radeon 9800 Pro its not faster than Geforce 3, but except Q2 it has almost same performance.
- Geforce 3 and Radeon 9800 are faster than V3 2000 PCI signifacantly.
- What really maters is less drivers overhead, but even 3Dfx is not so great in this area..
- Resolutions really doesnt matter too much, because even with 320x200 its cpu limiting factor, so can get good framerate even at such resolution in tested games regardless of details, have logic and OS simply consume all CPU power (that bad thing), good thing is that framerate is usually almost the same with in high resolutions like 1024x768,1280x1024, 1600x1200 (except V3 where is too much for this card).
- Geforce 3 is fasters, because can give you more performance in Q2 than Radeon 9800 Pro.
- Voodoo 3 maybe make sense for someone compatibility and glide
- Radeon 9800 Pro could make sense, because of digital display output. If compatibility would good - it would need more testing.
- if you care about price - Geforce 3 and Radeon 9800 are cheaper than 3dfx..
- i tested Geforce 3 Ti 200, Geforce 3 Ti 500 could be a bit faster.

Im old goal oriented goatman, i care about facts and freedom, not about egos+prejudices. Hoarding=sickness. If you want respect, gain it by your behavior. I hate stupid SW limits, SW=virtual world, everything should be possible if you have enough raw HW.

Reply 72 of 91, by ruthan

User metadata
Rank Oldbie
Rank
Oldbie

Its PCI version so no AGP drivers problem.. i doubt that chipset drivers we do much for PCI if something at all.
Windows 98 its SE vanilla, i dunno if exist some update which can give some significant performance boost?

So i would be probably chipset HW thing, if someone can post VIA test results i would be nice.

Im old goal oriented goatman, i care about facts and freedom, not about egos+prejudices. Hoarding=sickness. If you want respect, gain it by your behavior. I hate stupid SW limits, SW=virtual world, everything should be possible if you have enough raw HW.

Reply 73 of 91, by The Serpent Rider

User metadata
Rank l33t++
Rank
l33t++

Its PCI version so no AGP drivers problem..

Does not matter.

I must be some kind of standard: the anonymous gangbanger of the 21st century.

Reply 74 of 91, by appiah4

User metadata
Rank l33t++
Rank
l33t++
The Serpent Rider wrote:

Its PCI version so no AGP drivers problem..

Does not matter.

Yeah, 3dfx cards don't use any AGP specific technologies.

Retronautics: A digital gallery of my retro computers, hardware and projects.

Reply 75 of 91, by lost77

User metadata
Rank Member
Rank
Member

Looking here (https://www.anandtech.com/show/160/10) a K6-II at 333Mhz gets 59.4 FPS in Quake II with a single Voodoo 2. It seems to become CPU limited at ~76 FPS. Your scores seem WAY off.

Maybe you are running without L2 cache?

But then again even a Celeron 266 (no L2) can deliver up to 70 FPS in Quake II with a Voodoo3 2000.

Reply 76 of 91, by ruthan

User metadata
Rank Oldbie
Rank
Oldbie

Hmm,
i can check bios for L2 cache, but doubt that problem would be here. Are somewhere some CPU benchmarks Win98 test to compare, if cpu performance is ok?

Update: I did some CPU Dos testing in my pure Dos benchmarks and pathb - giving me 27s i would say that is as expected.. Bios seems to be fine?

2019-08-28 17.42.31.jpg
Filename
2019-08-28 17.42.31.jpg
File size
228.62 KiB
Views
1243 views
File license
Fair use/fair dealing exception

I dont have Win98 number in my memory, so i dont really know what to except, its 20 year when i had such machine and it wasnt fast either. I dunno if CPU or MB could be such way defective, well theoretically all is possible.. I know that RAM caching has some influence too, but i didnt see any related bios settings.
It could be some problem with 256 MB stick, but memtest where for hours ok.

Im old goal oriented goatman, i care about facts and freedom, not about egos+prejudices. Hoarding=sickness. If you want respect, gain it by your behavior. I hate stupid SW limits, SW=virtual world, everything should be possible if you have enough raw HW.

Reply 77 of 91, by Viper87

User metadata
Rank Newbie
Rank
Newbie

im also looking for fast GPU [only for some tests in PCmark04 for HWbot] i tested 9200SE, FX5200/5500/5700, 440MX, ti4200, [all of them are 8x agp compatilile] and none of them didnt work properly 🙁 now im looking for gf 4200 ti but only on agp 4x, anyone tried if on SS7 ?

Reply 78 of 91, by ruthan

User metadata
Rank Oldbie
Rank
Oldbie

Im still confused by this ALI V SS7 K6-2+ 500 + machine.. I installed Windows 2000 and XP besides of Windows 98, to do more testing.

With Windows 2000, im gettings same 3D apps problems, depends on installed Nvidia drivers lots of games - not starting with various errors or i got just reboots, some are working but see some big performance boosts agains Win98 results. Something seems to be faster, but there not games with inbuild benchmarking, something seems to be slightly more fluent (Diablo II Demo). Except 3D machine is stable and when game starts, its are stable too.

With Windows XP SP3, its the same, for basic 2D work, i was quite surprised that its not slow.. after boot i have used 100 MB of 512 MB ram, what is good.

I tried Everest 2.20 Home memory tests and there are as should be, so probably nothing like crippled cpu performance here.

I checked Phils video with new newer revision of this MB, it has ATA66 instead of ATA33 and he is able to reach 60 MB/s in Win98.. I with Windows XP and Attobench getting better performance up to 9 MB/s instead of 6 MB/s in Windows 98. I dunno if problem is only IDE to Sata adapter i have only 1 type. I tried RoadKill in Windows 2000, but its reporting only 2MB as DMA would be disabled, but there is not classic Windows 98 DMA checkbox per device.

I have still i bit problem find right CPU benchmark - i tried Winbench 99 - but its reporting some zillion MHz cpu speed, benchmark is completed, but it refusing to give results, because of strange CPU frequency.

Update SIS soft sandra 99 in XP - results is again good as expected, same as other PII 450 and K6-III systems..

Other strange things:
Windows 2000 - its there is strange bug that Half Life demo and Unreal tournament running too fast, character movement speed and shooting is too fast as some timer wouldnt wrong as expected.
Windows XP - Quake III demo its starting but its slide show < 1 PFS.

Im old goal oriented goatman, i care about facts and freedom, not about egos+prejudices. Hoarding=sickness. If you want respect, gain it by your behavior. I hate stupid SW limits, SW=virtual world, everything should be possible if you have enough raw HW.

Reply 79 of 91, by ruthan

User metadata
Rank Oldbie
Rank
Oldbie

I have found on this K6plus page this message:
Note that if you want to run a K6plus on a board with the Intel 430TX, 430VX, or ALi Aladdin IV(+) chipset, you may have to set "System BIOS Cacheable" in the CMOS Setup to "Disable" to get it working.

Im not sure if - Aladdin IV(+) means also version V, i have only disassembled ALI V machine - i didnt wanted have other case in my room (so i have ready to run MB and CPU ready on the shelf).. so i can test it, could this be source of that very bad Windows GPU benchmark results?

Know i trying to make VIA MVP3 machine working to make comparision.

Im old goal oriented goatman, i care about facts and freedom, not about egos+prejudices. Hoarding=sickness. If you want respect, gain it by your behavior. I hate stupid SW limits, SW=virtual world, everything should be possible if you have enough raw HW.