Kamerat wrote:

melbar wrote:

@Kamerat […]
Show full quote

@Kamerat

I am little wondering about your values 3DBench & DOOM with T-Bird & GeforceFX5500. Should not be the FX5500 as strong as my GF2MX-400 or even little stronger?
When i look at the specs on wiki, the FX5500 has even with 64bit DDR a 20% higher memory bandwidth like a stock GF2MX400. Also the raw power is twice (FX5500 - 1080MOperations/s, GF2MX-400 - 400MOperations/s).
Finally the T-Bird should be little faster than a K6-2...

I don't know why I got so bad results. Changing the graphic card to ATI Radeon 9700 Pro improved 3DBench and Doom somewhat, 407.8fps vs. 332.6fps and 84.49fps vs. 80.57fps. Still my Doom result is kind of shitty compared to your K6-2 500 @ 117.81fps.

For a comparison of K7 have a look at entry 632 in phil's spreadsheet, it's my Duron "spitfire" 650MHz, Nvidia RIVA 128.

Kamerat wrote:

melbar wrote:

@Kamerat […]
Show full quote

@Kamerat

I am little wondering about your values 3DBench & DOOM with T-Bird & GeforceFX5500. Should not be the FX5500 as strong as my GF2MX-400 or even little stronger?
When i look at the specs on wiki, the FX5500 has even with 64bit DDR a 20% higher memory bandwidth like a stock GF2MX400. Also the raw power is twice (FX5500 - 1080MOperations/s, GF2MX-400 - 400MOperations/s).
Finally the T-Bird should be little faster than a K6-2...

I don't know why I got so bad results. Changing the graphic card to ATI Radeon 9700 Pro improved 3DBench and Doom somewhat, 407.8fps vs. 332.6fps and 84.49fps vs. 80.57fps. Still my Doom result is kind of shitty compared to your K6-2 500 @ 117.81fps.

I bet, though, with L1+L2 disabled, the Radeon 9700 Pro scores very similar to the FX5500.

melbar wrote:

I've found your entries in Phil's VGA database. Yes, the SiS 5591 chipset might be a strong player with the IBM 6x86MX (@266MHz) […]
Show full quote

I've found your entries in Phil's VGA database. Yes, the SiS 5591 chipset might be a strong player with the IBM 6x86MX (@266MHz).
I see three different row's as your said. One with the Mystique 220, one with the G450 and finally one with the onboard SiS 6326 AGP.

All of them seems to be really strong at DOOM, but there is this situation:
With 3DBench & PCPBench, the onboard has no chance, but with DOOM it's around 17% over the dedicated cards. Looks strange right?

My system with the SiS530 on the chart has very interesting results too when comparing the onboard AGP to other PCI cards (there is no AGP slot on the motherboard).
3dbench
TNT2 M64 PCI-293.9
SiS530 onboard AGP-255.5
V3-2000 PCI-242.1

PCPBench
TNT2 M64 PCI-113.9
V3-2000 PCI-108.1
SiS530 onboard AGP-96.7

Doom
SiS530 onboard AGP-682 (109.5 fps)
TNT2 M64 PCI-689 (108.4 fps)
V3-2000 PCI-797 (93.7 fps)

SiS 6326 loves DOOM, no doubt about it!!! I'll tell you what else is refreshing about this benchmarking, Intel is getting a good kicking! Take away Quake and watch the underdogs smile 😀 I'll have go with a IDT winchip C6 later on, however I don't expect that to look too dazzling! 🤣

Not the system I was intending, I was sure this was a C6 Winchip machine, but no it's a intel Pentium 200 MMX. Either way, it's all plugged up now motherboard is a ATX DFi 586ITXD Intel 430TX, graphics a Cirrus logic Laguna3D 5464 4MB PCi, memory 64MB SDRAM.

All caches on:
3Dbench2 100.5
PCP bench 42.3
DOOM 1236 realtics.
Speedsys 151.10

Level2 off:
3Dbench 99.5
PCP Bench 35.3
DOOM realtics 1301
Speedsys 151.09

Level 1 off level 2 on:

3D Bench2 34.7
PCP Bench 9.7
DOOM realtics 4768
speedsys 36.47

Both caches off:

3Dbench2 14.7
PCP bench 3.6
DOOM Realtics 13367
Speedsys 15.03

The poor old Cirrus logic Laguna is a bit of a dog for DOS performance, but it is fairly capable as a early 3D card. It can run final reality well & even 3D MARK 99 MAX that's something that a lot of these 1st gen 3D cards cannot do, certainly beats the Mystique etc in that department.
I'll install a IDT winchip into the K5 machine later on and give some scores. Should be getting a RiSE MP6 266 CPU from China in the next week too, so we can see how that fairs too.

BSA Starfire wrote:

Not the system I was intending, I was sure this was a C6 Winchip machine, but no it's a intel Pentium 200 MMX. Either way, it's […]
Show full quote

Not the system I was intending, I was sure this was a C6 Winchip machine, but no it's a intel Pentium 200 MMX. Either way, it's all plugged up now motherboard is a ATX DFi 586ITXD Intel 430TX, graphics a Cirrus logic Laguna3D 5464 4MB PCi, memory 64MB SDRAM.

All caches on:
3Dbench2 100.5
PCP bench 42.3
DOOM 1236 realtics.
Speedsys 151.10

Level2 off:
3Dbench 99.5
PCP Bench 35.3
DOOM realtics 1301
Speedsys 151.09

Level 1 off level 2 on:

3D Bench2 34.7
PCP Bench 9.7
DOOM realtics 4768
speedsys 36.47

Both caches off:

3Dbench2 14.7
PCP bench 3.6
DOOM Realtics 13367
Speedsys 15.03

The poor old Cirrus logic Laguna is a bit of a dog for DOS performance, but it is fairly capable as a early 3D card. It can run final reality well, that's something that a lot of these 1st gen 3D cards cannot do, certainly beats the Mystique etc in that department.
I'll install a IDT winchip into the K5 machine later on and give some scores. Should be getting a RiSE MP6 266 CPU from China in the next week too, so we can see how that fairs too.

Cool! Does your motherboard have a separate L2 that you are disabling? My Pentium system can only disable L1 (either Setmul or BIOS). There is no L2 to disable. And your results with L1+L2 disabled match close to my POD 200 with L1 disabled.

quite interesting actually, we see that the P5 MMX is much more sensitive to level 2 cache than Cyrix or P6's tested so far. So not quite the waste of time I was expecting.

Yes, both my intel TX boards have 512K level 2 cache that is controllable from the BIOS.

Added results from a Pentium M Dothan @800MHz. 😀

Thanks, Kamerat! Pretty dramatic difference between cached and cacheless 😀 It actually falls very closely into place with the "fast 386" goal that some of us have.

Does anyone have a 486 they'd be willing to submit results for? gerwin says Setmul will work with 486 and higher.

Kamerat, thanks for the entry. I resorted and also froze the first row and column. I think it will make browsing the sheet easier.

Forgot how painful benchmarking Doom without caches is 🤣

I've added entries for a machine, a P133, that will feature in a video soon: The 3 in 1 DOS Retro Gaming PC 😀

Added my benchmarks with my K6-III+ at 6.0x66. (maybe I should do other multipliers and FSB settings, like 4.0x100 to see what happens)

Interesting to note is that having L1 disabled vs L1+L2 disabled makes no noticeable impact except in 3DBench, where it takes a good hit. Disabling L3 only also makes little impact except in PCPBench, but even then it's small.

Also seems it's impossible to disable only L2 on a K6-III+, making the benchmarks of L2 disabled and L2+L3 disabled impossible. Setmul just can't do that. Oh well.

Having three caches sure makes this interesting, though. Have to do a few more tests than other systems, 6 in total. (none, L1, L3, L1+L2, L1+L3, L1+L2+L3)

Thank you, guys. Much appreciated. 😀 Re-sorted!

Phil- when I tested my Pentium, strangely, with L1 disabled the 60Mhz FSB speeds were faster than 66Mhz. So P120 and P90 were faster than P133 and P100. I'm curious if you have an opportunity to test at P120, if the pattern will continue. If you're not able, no worries.

Agent- those are interesting results. Pretty weird about the 3dbench being so much faster at L1+L3 vs L1+L2+L3! But either of those look like they put you in the sweet spot for Wing Commander 1 (fast 386).

I have to check, but I think that board only goes down to 66.

PhilsComputerLab wrote:

I've added entries for a machine, a P133, that will feature in a video soon: The 3 in 1 DOS Retro Gaming PC 😀

Phil, your pentium (P54C(S), not a P5 right 😉 ) shouldn't be a 2x66 🤣

melbar wrote:

PhilsComputerLab wrote:

I've added entries for a machine, a P133, that will feature in a video soon: The 3 in 1 DOS Retro Gaming PC 😀

Phil, your pentium (P54C(S), not a P5 right 😉 ) shouldn't be a 2x66 🤣

Not sure what you're trying to say?

What i mean is: what is about your 2nd entry in the excel spreadsheet 😉

Ah I see it now 😀 I changed it.