Much like the Celerons scores. I was thinking maybe Doom liked the TNT in the P-233mmx, but this is the same 16mb NV4 GPU, but its AGP. What is with that P-233mmx doom score 🤣 .
Edit3: Just re ran the P-233MMX, and yes it got the same high doom fps score. Does doom just run really fast on the P55c? or is it something to do with the SIS chipset its using?
I just got an S3 Virge GX/2 board and tried it out, and reported my results. To my own surprise, it performed quite similar to my GeForce FX 5500. Some tests showed that it was faster, while others reported the same speed. It also responded to the VGA boost differently, often being slower than my 5500.
Pentium MMX 233 | 64MB | FIC PA-2013 | Matrox Mystique 220 | SB Pro 2 | Music Quest MPU Clone | Windows 95B
MT-32 | SC-55mkII, 88Pro, 8820 | SB16 CT2230
3DFX Voodoo 1&2 | S3 ViRGE GX2 | PowerVR PCX1&2 | Rendition Vérité V1000 | ATI 3D Rage Pro
This tests is more lemitrd by cpu, chipset and bus speed on PCI systems and more modern.
Video card and his bus is not a bottleneck from about 486 systems.
I've started my test on 486 DX2-66 with bios default.
Tested few videocards and found similar benchmarks.
Then, I've statret bios optimisation, and on absolutely same system got 10 and 20% higher bench: (two bottom lines)
So, cache and ram timings, bus speed - is highly sensivity for dos gaming system.
I'm upload full results later, after finish my tests.
So, cache and ram timings, bus speed - is highly sensitivity for dos gaming system.
This is not really something new and unusual. Tight timings and high bus speed goes hand in hand with higher performance in every way - not just games.
When doing benchmarks it is always good to use the best possible BIOS optimizations. 10-20% gain is not uncommon with good optimizations.
BTW:
Your last and best results are quite fast for 486DX/2-66.
BTW:
Your last and best results are quite fast for 486DX/2-66.
I know. Moreover, later I've got 3D Bench:
51,3 - Am486 DX2-66 (33*2)
68,6 - Am486 DX4-100 (33*3) WB L1
81,8 - Am486 DX4-120 (40*3) WB L1
71,4 - i486 DX4-100 (33*3)
85,8 - i486 DX4-100 overclock to 120 (40*3)
so looks like this is one of top configuration, even faster, than average AMD 5x-133
RacoonRider wrote:Made a long video card benchmark run with my new build-in-progress, AMD 5x86 -133 on Acer AP43 (SIS 496/497). […] Show full quote
Made a long video card benchmark run with my new build-in-progress, AMD 5x86 -133 on Acer AP43 (SIS 496/497).
Matrox Athena showed the expected poor DOS performance; however I expected better from V2200. S3 videocards show impressive scores yet I was confused when 64V+ turned out to be slower than 64, given that 64 has 60ns EDO ram and 64V+ - 45ns. S3Trio 3D/2X has the best score of all S3s, yet I expected it to slightly outperform Millenium II and ET6000. Well, it didn't, despite SDRAM and being manufactured in 2000.
The top scores are achieved with Matrox Millenium II, S3 Trio 3D/2X, ET6000, Matrox Mystique, but ET6000 did not work in DOS text mode, so Doom score is unknown. We see almost equal scores: at this point, the CPU is no longer bottlenecked by the videocard.
P.S. I'm adding Millenium II score into the list. Are the other scores worth adding? It would create confusion if I added V2200 and Athena scores there; S3 scores are more or less identical; ET6000 is inconsistent.
Rendition cards are well known for having really slow VGA performance (mode 13h or Mode-X/Y) while having good VESA performance. You can use renutil to remap mode 13h to a VESA mode, but this is not compatible with Doom (or other games using Mode-X/Y).
Regarding the difference between 64V+ and 64: Faster RAM does not automatically make a difference, the timings actually used are set by the video cards BIOS. I you are interested you can use MCLK to check the current RAM timings and set faster timings. S3 cards usually have a big margin here, I suppose manufacturers sourced different RAM chips and used conservative timings in the BIOS.
AlphaWing wrote:
RacoonRider wrote:
AlphaWing wrote:
I wonder why the P-233 MMX getting such a good fps rate in doom?, even the PII-450 did not beat it.
Also I added all these to the database.
Did you use FASTVID for PII-450?
Yea I did. That sis board still won.
I thought FASTVID doesn't do anything for Mode-X/Y, but according to http://web.archive.org/web/20020408044840/htt … orials.php?id=4 it can double performance in Doom 2. Does it make a difference in other games?
Classic Pentium (MMX) has a faster bus transfer than PPro/PII when not using FASTVID or similar tools.
easy_john wrote:I know. Moreover, later I've got 3D Bench:
51,3 - Am486 DX2-66 (33*2)
68,6 - Am486 DX4-100 (33*3) WB L1
81,8 - Am486 DX4-120 (40 […] Show full quote
kixs wrote:
BTW:
Your last and best results are quite fast for 486DX/2-66.
I know. Moreover, later I've got 3D Bench:
51,3 - Am486 DX2-66 (33*2) 68,6 - Am486 DX4-100 (33*3) WB L1
81,8 - Am486 DX4-120 (40*3) WB L1 71,4 - i486 DX4-100 (33*3)
85,8 - i486 DX4-100 overclock to 120 (40*3)
so looks like this is one of top configuration, even faster, than average AMD 5x-133
Can you also post a picture of Speedsys and Cachechk with Am486 DX4-100 (33*3) WB L1 and i486 DX4-100 (33*3). What motherboard are you running this on? Thanks!
I can some Cyrix chips and also a IntelDX4 Overdrive. Nice.
IntelDX2/4 Overdrive (DX2odpr66, DX4odpr100) is absolutely same as intel DX2/4, but work on 5v mobo.
I can't find any difference in features or speed between odpr and ordinary DX2/4, so I stop tests 486 Overdrive and use only iDX4-100.
Pentium overdrive (PDPV) is more interesting, but mine is still not arrived.
Can you also post a picture of Speedsys and Cachechk with Am486 DX4-100 (33*3) WB L1 and i486 DX4-100 (33*3). What motherboard are you running this on? Thanks!
Unfortunately I don't have WB intel "&EW" right now for side by side comparison.
But I do test on Amd, and there is not to much increasing on WB cache: .
And here you can see difference between 128 and 256 kb cache (line 3 and 4) .
(this tests on different mobo - 4dps)
I can't post cachechk, since I save only speedsys and benchmark results and don't run any other utilities. But I can do it later, if it's important.
PS sorry about watermarks, you can open picture link and there pictures w/o it.
I usually run Cachechk too. I kinda think its shows a little better how fast the memory is then Speedsys.
Done for amd dx4-100 for this time:
cachechk:
1CACHECHK V7 11/23/98 Copyright (c) 1995-98 by Ray Van Tassle. (-h for help) 2 CMOS reports: conv_mem= 640K, ext_mem= 31,744K, Total RAM= 32,384K 3 "AuthenticAMD" 486 Clocked at 100.0 MHz 4 Reading from memory. 5 MegaByte#: --------- Memory Access Block sizes (KB)----- 6 1 2 4 8 16 32 64 128 256 512 1024 2048 4096 <-- KB 7 0: 11 11 11 11 11 20 20 20 20 30 -- -- -- us/KB 8 1: 11 11 11 11 11 20 20 20 20 30 30 30 30 us/KB 9 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 <--- same as above. 1019 20 21 22 23 24 25 26 27 28 29 30 31 <--- same as above. 11 12 Extra tests---- 13 Wrt 16 16 16 16 16 16 16 16 16 16 16 16 16<-Writing 14 This machine seems to have both L1 and L2 cache. [reading] 15 L1 cache is 16KB-- 103.6 MB/s 10.1 ns/byte (279%) (185%) 3.9 clks 16 L2 cache is 256KB-- 55.9 MB/s 18.8 ns/byte (150%) (100%) 7.2 clks 17 Main memory speed -- 37.1 MB/s 28.2 ns/byte (100%) [reading] 10.8 clks 18 Effective RAM access time (read ) is 112ns (a RAM bank is 4 bytes wide). 19 Effective RAM access time (write) is 60ns (a RAM bank is 4 bytes wide). 20 "AuthenticAMD" 486 Clocked at 100.0 MHz. Cache ENABLED. 21 Options: -t0
ctcm:
1XMM not installed ! 2CTCM uses Memory from address 00120000 3Int15-Memory=32505856 4 5CTCM, translated by Thomas Pabst, copyright ct-Mag. V1.5b/t2 6Processor-Timing : i486DX,i487DX,i486DX2,am486DX,am486DX2,am486DX4,am5x86 7Processor CPUID : am486DX4-WB/ Typ:00 Fam:04 Mod:09 Stp:04 8Clck : 99.8 MHz 9internal Bus : 32 Bit between CPU and primary Cache or Memory 10FPU : i487-Typ 11L1 Cache : 16 KByte,4-way associative 12L2 Cache : 256 KByte, direct mapped 13Write Strategy L1 : Write Back 14Write Strategy L2 : Write Back 15Dirty Tag L2 : ok 16 17 Through Put & Bus Performance: Main Memory from 00120000 18 19Best Time for 16K MOVSD (Cache /Page Hits) : 125 mcs - 130.9 MByte/s 20average t. for 16K MOVSD (Miss + Hit) : 461 mcs - 35.5 MByte/s 21average t. for 16K MOVSD (if clean) : 645 mcs - 25.4 MByte/s 22average t. for 16K MOVSD (if dirty) : 1012 mcs - 16.2 MByte/s 23Main Memory 16K MOVSD (Cache misses) : 1077 mcs - 15.2 MByte/s 24 25average with 256 KB L2-Cache /DOS (640K) : 519 mcs - 31.5 MByte/s 26average with 256 KB L2-Cache /WIN (4M ) : 633 mcs - 25.9 MByte/s
(Bios options may be not 100% identical, as in previous test, i've test to many boards and cpu, so may be change something)
I'm don't use this utils before. What is more important in this reports?