Stojke wrote:

What is the performance difference between Stingray 128/3D and Diamond Monster 3D (with S3 Virge GX)?
I want to use it in an Socket 5 Pentium 133MHz computer.

I have a Jazz Mutimedia Rush card and I would say it is 5-7 fps slower than a voodoo1 on Glquake

What are Voodoo 1 results?
I am interested into percentage, since i have only two PCI slots. Freeing one up opens more room for expansions.

According to most sources, the Voodoo Rush performed about 10% slower than a Voodoo Graphics. To resolve this companies clocked the Voodoo Rush 10% higher. Don't know if they all did it though.

Stojke wrote:

What are Voodoo 1 results?
I am interested into percentage, since i have only two PCI slots. Freeing one up opens more room for expansions.

31-32 fps. Glquake 0.97; SST_GRXCLK=50; set SST_SWAP_EN_WAIT_ON_VSYNC=1

I don't think all dos glide titles are compatible with Rush either.

I only ever want to try Final Fantasy 7 and Tomb Raider.
And HL / UT99 on my voodoo 5.

NitroX infinity wrote:

According to most sources, the Voodoo Rush performed about 10% slower than a Voodoo Graphics. To resolve this companies clocked the Voodoo Rush 10% higher. Don't know if they all did it though.

AFAIK, Hercules was the only one that boosted the clocks. They also added a little more RAM over the reference design. That's why the Hercules Rush cards generally get higher prices than the rest.

sliderider wrote:

NitroX infinity wrote:

According to most sources, the Voodoo Rush performed about 10% slower than a Voodoo Graphics. To resolve this companies clocked the Voodoo Rush 10% higher. Don't know if they all did it though.

AFAIK, Hercules was the only one that boosted the clocks. They also added a little more RAM over the reference design. That's why the Hercules Rush cards generally get higher prices than the rest.

AFAIK, this comes from first Hercules Rush card being 45 MHz. Later they used 50 just like anybody else, so equal to Voodoo Graphics. I measured the Rush was 20% slower on average.

How should i benchmark?

Quake & Quake 2 (timedemos), 3DMark 2001se.

Since timedemo doesnt display FPS information here is Player Bench and Super Scape:

CPU: Intel Pentium 1 133MHz (socket 5)
RAM: 128MB EDO 70ns

Diamond Monster 3D 4MB + S3 Virge GX 2MB
Player Bench: 16.1 FPS
Super Scape: 108.6 FPS

Hercules Stingray 128/3D 8MB
Player Bench: 16.1 FPS
Super Scape: 108.8 FPS

Stingray performance is almost exactly the same as Diamond Monster 3D + S3 Virge GX. Where it has 0.2FPS more in Super Scape.

So Virge GX and AT25 are probably CPU limited.

Which means for my needs Stingray is enough.

Now that is unexpected. Everyone always says that Rush is slower than Voodoo 1 but this test shows that it is almost exactly the same with that CPU. I wonder at what point they begin to diverge.

I can test with the following: AMD 5X86 , IBM 233+ , AMD K6-II 400MHz , Pentium II 350MHz and Pentium III 700MHz.
Also, is there any kind of NT3.5 NT4 program to test the same thing? Since this computer has two Pentium 133 processors.

Here are some more results from Super Scape and Player Bench.

AMD K6-II 400MHz

Hercules Stingray 128/3D
PlayerBench 30.3 FPS
SuperScape 184.6 FPS

Diamond Monster 3D + S3 Virge GX 2MB
PlayerBench 30.4 FPS
SuperScape 184.1 FPS

Dell Optiplex GX1 Pentium 3 700MHz

Hercules Stingray 128/3D
PlayerBench / (doesn't work at all, i blame Dell)
SuperScape 360.9 FPS

Diamond Monster 3D + S3 Virge GX 2MB
PlayerBench 39.9 FPS
SuperScape 360.6 FPS

I will try Quake newer version on AMD K6 and i will try all the rest on IBM 233+ .

[E]

AMD 5x86 133MHz , 4MB FPM memory, 256kB L2

Hercules Stingray 128/3D
PlayerBench 7.1 FPS
SuperScape 74.6 FPS

Diamond Monster 3D + S3 Virge GX 2MB
PlayerBench 7.5 FPS
SuperScape 76.6 FPS

Superscape is not 3d accelerated. What was the PlayerBench setting?

VESA mode 100 (640x400 8bpp LFB)

As I suspected, you are not testing the Voodoos at all.

You know, you aren't really that helpful if you constantly miss out on giving me instructions on how to actually test voodoo.