Up to this day my opinion was :
When I overclock my graphics card and do not see any improvements in speed, then the CPU is the limiting factor. I saw this on my PC :
A Geforce4-TI4200 overclocked to the settings of a 4600 makes no sense, as long as the Pentium3-1400 is the limiting factor.
Now I have built a new PC and experience the exact same thing :
The V5-6000 runs along with a Pentium4-2800 and I get absolutely no improvement when I clock the Voodoo to 183 Mhz. Not even a few percents.
Is it possible, that the P4-2800 still isnt strong enough to max out the 6000 ? If that is the case, what CPU actually IS strong enough ?

Any suggestions ?

Regards

Rolo01

You do remember to reboot after you've set the new clockspeed of your V5 6K, don't you?

Well, I used V-Control and clicked on APPLY after changing the clock setting, so a reboot isnt necessary. At least it wasnt with the V5-5500...

QuoteOriginally posted by Rolo01

Well, I used V-Control and clicked on APPLY after changing the clock setting, so a reboot isnt necessary. At least it wasnt with the V5-5500...

My good is experience is with AMD Athlon CPU's, they are based on HP's RISC technology.

What you could do is build aa AMD AthlonXP 2700+ of 3000+ system with the ASUS A7V333-R Rev 1.02 mainboard, flsh the bios to 1017 so you have Barton support to 3000+, okay yau can fit in a Barton 3200+ but this mainboard will not support the 400Mhz FSB, but a little Overclocking should help you there.

Why the ASUS A7V333-R, well this is known to ben one one of the best VIA KT333 mainboards for the AthlonXP CPU's to 3000+ support and also it is the best mobo for friendly Voodoo AGP support.

So I think you will with an AMD based system, get more protetial out of you V5 6K than on your P4 system. The performance of an AthlonXP 2600+ is about the same of a P4 at 2800Mhz. Remember, that the rating 2600+  actually equals the performance of an  AMD Athlon Thunderbird at 2600Mhz and NOT a Pentium4 at 2600Mhz!!! It has been reviewed many times.
Just for the AMD rating confusion.

And another reason to choose for AMD is, that the most Voodoo cards work best on Athlon systems, games like Unreal, Unreal Tournament and  Descent3,  are all AMD Athlon & 3dfx Optimized Optimized.

I hope that I gave you a nice tip here.

P.S. It was not my intention to scare you, but just to help you out and tell you the truth the way it really is. I hope you understand my feelings, Rolo and no offense please, just help´n out. That´s what we´re here for right
First I´m  gonna find some drivers for my keyboard

Hey Obi-Wan,
thanks for your input and dont worry, I dont feel offended at all.
The reason why I built another intel system is that I have no experience with AMD systems.
And I was curious whether it is possible to build a good system with a Pentium4 and a V5-6000. In case the 6000 would not work well in that mainboard I would have used the cpu in my primary rig.
Its a pity that overclocking the FSB is not an option, because the 6000 is very sensitiv as soon as the AGP clock rate rises. You get lots of pixel errors very fast.
Dont get me wrong, I am not disappointed by the performance of the 6000. In fact, I am very pleased to see what that baby can do...
I read somewhere that the scalability of the V5-5500 is at an end at lets say 1300 Mhz, so I guessed the 6000 would be at an end at about 2400 Mhz, but maybe I was wrong...

QuoteOriginally posted by Rolo01

Hey Obi-Wan,
thanks for your input and dont worry, I dont feel offended at all.
The reason why I built another intel system is that I have no experience with AMD systems.
And I was curious whether it is possible to build a good system with a Pentium4 and a V5-6000. In case the 6000 would not work well in that mainboard I would have used the cpu in my primary rig.
Its a pity that overclocking the FSB is not an option, because the 6000 is very sensitiv as soon as the AGP clock rate rises. You get lots of pixel errors very fast.
Dont get me wrong, I am not disappointed by the performance of the 6000. In fact, I am very pleased to see what that baby can do...
I read somewhere that the scalability of the V5-5500 is at an end at lets say 1300 Mhz, so I guessed the 6000 would be at an end at about 2400 Mhz, but maybe I was wrong...

Whell I'm not sure, I still work with the 5500 AGP, and I gotta say it works alot better in my XP2700 + than my XP2100+, especially with 4x FSAA in Unreal.

So about the theory of that a 1300 Mhz cpu is about the limmit of the 5500 , that is not what agree on my friend;) the fun begins just around the 1800+, that 2000 Mhz for P4 users[:p]

With my XP2100+ system the cpu had a natrual speed of 1733Mhz and you can compare it with a Pentium4 at 2200Mhz and yeah Unreal didn't go that well with 4x FSAA at reso 1024x768x16 [Glide] So a year or so later my dad bought a PC from his work, and we swapped CPU's He took my 2100+ and I got his 2700+ for it. and now I play Unreal very playable at 1024x768x16 [Glide] with FSAA x4 at 42 Frames per sec !!!!
And that game looks very 1337 dude!!

Just wonder if how it would look at the same reso but then with 8x FSAA!! with your 6000 LOL Is the difference big between FSAA x4 on a 5000/5500 AGP and FSAA x8 on a 6000? I was just wondering that's all or could I compare it to ATi's FSAA x6 from my Radeon 9700 Pro's?

Okay here are my 3 sys specs in short and yes I have tested alot of voodoo cards with them and I am still doing that and so on:

1. FIREFOX SMP 2002

Dual AMD AthlonMP 2200+ Thoroughbred-A, 2x 1800Mhz, 266Mhz FSB, 256KB L2 Caché
ASUS A7M266-D AMD AM760MPX
Club3D Radeon 9700 Pro 128MB DDR Core/Mem 325Mhz/620Mhz [310DDR]
1x 1GB PC-2100 ECC Reg Cl 3,0 Twin-Moss

2. GRADINKO 2600

AMD AthlonXP 2600+  Barton-B , 1917Mhz, 333Mhz FSB, 512KB L2 Caché
ASUS A7N8X2.0 Deluxe nForce2 ULTRA 400
Hercules 3D Prophet 9700 Pro 128MB DDR Core/Mem 325Mhz/620Mhz [310DDR]
2x 256MB DDR 333 PC-2700

3. SVETA 2700

AMD Athlon XP 2700+ Thoroughbred-B, 2167Mhz, 333Mhz FSB, 256KB L2 Caché
ASUS A7V333-R [VIA KT333-RAID]
3dfx Voodoo5 5500 AGP 64MB Rev 2500 A cores/mem 166Mhz
1x 512MB PC-2700 CL 2.5 Samsung Original


I've tested my Voodoo5 5500 AGP also in that Dual MP 2200+ system and Unreal with FSAA x4 did go up a little just about 25/32 frames per sec, but not sufficant, so the XP2700+ still had the prize for FSAA x4 gaming and that combined with VIA's KT333 you'll be quite on the go.

The XP2600+ has a nVidia nForce2UlTRA-400 mobo and you can try an fit your 5500 in there but the AGP x8 slot didn't agree neither did my 5500 AGP LOL[:p] so that's whay I bought myself a 5500 PCI from a good msn friend later in the year.

The Voodoo4 4500 AGP did Unreal also very good  at 1024x768x16[Glide] and FSAA x2 on my Barton 2600+ , thanks to it's universal AGP x4 adapter the Voodoo4 4500 AGP didn't mind the AGP x8 at all, and yes the nForce2ULTRA 400 is indeed a great speed improvement for my Voodoo4 4500 AGP 32MB Rev 3300 A0

What are your experiences with Voodoo4 4500 AGP, Voodoo5 5000 AGP, Voodoo5 5500 AGP and Voodoo5 6000 AGP with Unreal and FSAA at max?

Well I suppose you all rembered this:

3dfx is concentrating on maximizing revenues while we ride out our component shortage and executing on future product releases. For the near future, this means making sure the 5500 series is successful. I do not think it is any secret that giving out dates prematurely has hurt our company in the past. The V5 6000 product is in development. We will be releasing the V5 6000 this summer. When the release date solidifies, we will make sure everyone knows what it is.

BB
Senior PR Manager
3dfx, Inc.

When I saw that message for the first time, I couldn't wait for that moment to come, but it never came didn't it Yes everyone it is very sad that 3dfx is gone. The only thing we can do is collect the leftovers and keep the drivers up to date and enjoy the games we have to play with them of course. 3dfx isn't dead for me yet. Voodoo5 still ownz:D

And over here:

http://members.aon.at/grandadmiralthrawn/index.htm?http://www.hardoverclock.de/html/reviews/3dfx_voodoo5_6000_agp/v5-6000_1.htm
is a nice review of the Voodoo5 6000 Enjoy!

Yes, the Rewiev of Grand Admiral Thrawn is really good.
Another article worth reading is here :
http://www.sudhian.com/showdocs.cfm?aid=413

BTW, I downloaded the latest Unreal patch 226 and will perform some tests this evening...

Ok, here are the results of the Unreal benchmarks.
I made a mistake and benched in 32 bit color, but anyway...

Image Insert:

85.27 KB

QuoteOriginally posted by Rolo01

Ok, here are the results of the Unreal benchmarks.
I made a mistake and benched in 32 bit color, but anyway...

Image Insert:

85.27 KB

Wow that's pretty impressive there Rolo  But now the funny thing , the Voodoo4-2 4800 32MB DDR-RAM is slower than the Voodoo4 4500 AGP 32MB SD-RAM...howcome? is it that  because the 4800 is still on the buggy track or is it that the DDR memory is alot slower than the 4500's 128Bit SDRAM?

But even the Voodoo5 5000 wins it by a few frames from the 5500 without FSAA LOL:D
And yeah okay the Voodoo5 6000 is still the king indeed:) Bravo there my friend.

Finally I get an idea how the Voodoo4-2 4800 also performs, by the way I never knew you had one of those[:I]

Hi to both,

Very good WORK Rolo01! Very interesting!

Why the v5 5000 (32mb) has the seame performance vs voodoo5 5500 (64mb)? Perhaps because both have the same fill rate! Then so that we need 64mb ram...

I tought that v4 4800 had a performance more high! Perhaps why the v4 4800 Daytona works to 64 bits vs 128bits of the v4 and v5? I don't know!

Thanks
Oscar

Perhaps UT is not using the full potential of the ram? Do other games or benchmarks give you similar results? It's disturbing that the higher-end cards seem to be matched or beaten by the lower ones, when you know they shouldn't be.

Somewhere I read that the DDR-Ram on the daytona is only 64bit, I guess it was on Rashly´s page. My opinion is that a single gpu ( like the one on the daytona ) does not benefit from DDR-Ram because there is not enough data transfered to the ram. A 2 chip version ( like the V5-2, that was planned but never produced ) would make more sense.
The daytona runs without heatsink, not even passive cooling. Nevertheless I did not experience any bugs, no pixel errors or such. Games like Unreal or Jedi Knight 3 run very good, even Neverwinternights runs great with MESA. But in fact the daytona IS a low end card.
I was surprised that the 5000 equals the 5500, maybe I should try higher resolutions with FSAA, somewhere the additional ram must have an effect...

Rolo, would you be willing to test the V4 4800 and V4 4500 at 140Mhz, 150Mhz, 160Mhz and possibly 170Mhz so we can see how the V4 4800 scales up compared to the V4 4500?

I'd really like to know the theoretical potential of the V4 4800.

Yeah, I can test it but need some advice.
I use DDR266 Ram, so it might not be overclockable. Can I set the Ram to run asynchronus from FSB ?