Yeah of course a local high speed low latency L2 cache is unmatched compared to the 'classic' P1 board design.

I wonder how much of the K6-2+/III/III+ advantage in 3DBench actually stems from the on-die L2 cache, considering the much higher clock rate and different architecture.

If you compare the K6-2 and K6-2+ scores and taking the K6-2+'s extra 200MHz into account, the score isn't unreasonably high and the K6-2 might theoretically reach that score were it at the same clock speed.

Maybe someone can test the K6-2 and K6-2+/III/III+ at the same speed in 3dbench in the same system.

Yea that would be quite interesting!

Once I have more time (hopefully within a week or so), I'll start to chip in 😉

I'm desperately curious to try out some of my stuff (unlocked P2, MMX (and hope my MMX 266 mobile chip works!!), Cyrix MII, Winchip, Rise, K6-2/3/+, and maybe even an Athlon XP mobile).

One more Slot 1 system bench:

Intel-Celeron-100-8-800-ASUS-P2B-Intel 440BX-128KB-AGP-3dfx Voodoo 4 4500-32MB-sprcorreia-852.5

Here's another bench:

Intel-Celeron-100-9-900-ASUS-P2B-Intel 440BX-128KB-AGP-3dfx Voodoo 4 4500-32MB-sprcorreia-912.5

Mau1wurf1977 wrote:

Make sure you read the first post regarding FASTVID to unlock full DOS VGA performance on Slot 1 machines.

Or benchmark from within W98. I have uploaded your results and can change the score any time!

I've tried the same setup and tested under DOS and Windows 98 and found that in Windows card performs much slower than DOS... Any suggestions?

Not sure what's going on. Stick to DOS is all I can think of?

Mau1wurf1977 wrote:

Not sure what's going on. Stick to DOS is all I can think of?

This would be best.
There can be many things that may influence the results. Thing is, we want results that compare hardware to hardware and using the same OS (in this case a bootable disk) minimizes influencing the results beyond what we are trying to do here (which is benchmarking hardware).

Here's one bench that will (hopefully) bridge the gap between 486 and Pentium processors:

Intel-Pentium-50-1.5-75-QDI-Titanium IB+-Intel 430TX-512KB-PCI-Matrox Mystique 220-4MB-sprcorreia-79.3

So, not wanting to go the 486 way, this seems like a choice. And paired with a slower graphic card it looks even better:

Intel-Pentium-50-1.5-75-QDI-Titanium IB+-Intel 430TX-512KB-PCI-ATI Rage II-2MB-sprcorreia-67.3

Hell, if i throw in an ISA card it looks like this:

Intel-Pentium-50-1.5-75-QDI-Titanium IB+-Intel 430TX-512KB-ISA-ISA-Oak Technology OTI 087-1MB-sprcorreia-30.4

Intel-Pentium-50-1.5-75-QDI-Titanium IB+-Intel 430TX-512KB-ISA-ISA-Oak Technology OTI 077-1MB-sprcorreia-25.8

I got a couple Socket 4 boards, with cpu's. Could bench those also.
Oh...the plans in my head...I need to STFU, get my hiney in the attic and actually start doing something. AAAHHHRRRGG!!!! 😁

AMD-K6-2-66-6-400-PCChips-M571-SiS5598-512KB-PCI-Riva 128-4MB-unmei220-394

AMD-K6-2+-66-6-400-PCChips-M571-SiS5598-512KB-PCI-Riva 128-4MB-unmei220-472.4

I don't know if underclocked results are valid. This mobo only supports up to 83MHz FSB. Just for comparison purposes, I run both the K6-2 and K6-2+ at 66MHz FSB. Here is the result of the K6-2+ if I use 83MHz FSB:

AMD-K6-2+-83-6-500-PCChips-M571-SiS5598-512KB-PCI-Riva 128-4MB-unmei220-589.6

If this mobo could reach 100MHz FSB, I would be around 700fps I guess...

I have a note to make, wouldn't it be a good idea to not only list the cpu (Pentium III) but also it's core? (Katmai, Coppermine, Tualatin, Tualatin-s). These are basically different enough to warrant separate entries instead of just Pentium III, right?
Or perhaps a column where users can give a little extra info (like type of mainboard memory used like EDO or FPM).

And unrelated. I see mention of a GeForce TNT2 M64, this should be the Riva TNT2 M64 as the TNT2 family actually predates the GeForce 1 😉

^Good point on CPU. With the exception of higher clock speeds and the addition of SSE, the Katmai core was no different from the Deschutes Pentium II core. Tualatins, on the other extreme, had significant differences. Having that extra information could be helpful to some people.

However, I don't know how big of a difference memory makes. EDO and FPM, in real use and in benchmarks, don't show significant difference. For the newer CPUs on the chart, there's very little chance that memory used is something other than bus speed SDRAM. Then again, there were a good few Pentium III motherboards that used RDRAM.

Another couple of lines from me - after changing the memory parameters in the bios I can get 76.9 and even 83.3:

AMD-5x86-33-4-133-Abit-AB PB4 Rev1.3-Ali 1487-256kb-PCI-S3 Trio 64v+-unknown-ratfink-76.9
AMD-5x86-33-4-133-Abit-AB PB4 Rev1.3-Ali 1487-256kb-PCI-S3 Virge-unknown-ratfink-83.3

The change in graphics card is not instrumental here, from what I've seen the results don't change whatever pci card I use [trio, virge, rush].

I hope the format is correct.

just a few results will post more soon

intel-pentium-60-1.5-90-microstar-ms6159-ali5-512kb-pci-ati xpert@work-8mb-recoil2525-85.7

intel-pentium-60-2-120-microstar-ms6159-ali5-512kb-pci-ati xpert@work-8mb-recoil2525-100.9

intel-pentium-66-2-133-microstar-ms6159-ali5-512kb-pci-ati xpert@work-8mb-recoil2525-113.2

intel-pentium mmx-66-3-200-microstar-ms6159-ali5-512kb-pci-ati xpert@work-8mb-recoil2525-141.5

intel-pentium mmx-66-3.5-233-microstar-ms6159-ali5-512kb-pci-ati xpert@work-8mb-recoil2525-150.2

cyrix-6x86/p150+gp-60-2-120-microstar-ms6159-ali5-512kb-pci-ati xpert@work-8mb-recoil2525-125.9

cyrix-6x86l/p200+gp-50-3-150-microstar-ms6159-ali5-512kb-pci-ati xpert@work-8mb-recoil2525-145.2

cyrix-mkll/333gp-83-3-250-microstar-ms6159-ali5-512kb-pci-ati xpert@work-8mb-recoil2525-210.6

amd-k6/2 300-100-3-300-microstar-ms6159-ali5-512kb-pci-ati xpert@work-8mb-recoil2525-229.7

recoil2525 wrote:

cyrix-mk2/333gp-83-3-333-microstar-ms6159-ali5-512kb-pci-ati xpert@work-8mb-recoil2525-210.6

With cyrix mk2, did you mean the Cyrix MII or M2?
The 333gp part is it's PR rating, it's not it's actual speed. The M2 officially ran upto 300Mhz for the M2-433.

Here's a poor 386 result, only has 1920kb ram could that be the cause?

AMD-386DX33-33-1-33-Asus-ISA386C-SiS-64kb-ISA-Speedstar-1mb-ratfink-7.2

No Ram amount shouldn't make an impact.

Check the turbo button and also run a CPU benchmark to see if the CPU speed is ok.

SIS should be fine, worst case it might be the video card.

it a cyrix mll black top one

www.cpu-world.com/CPUs/MII/Cyrix-MII-33 ... CPGA).html