Over 100 fps is an awesome score 😀

On Slot 1 boards make sure you run menu option 1 to unlock some extra VGA performance.

http://youtu.be/IWd1cQYEcIk?t=5m49s

With a MMX 233 mine does 94.31 but still ahead of the rest 😀

I tried finding out the cachable area.

The board takes 256 MB just fine. I tried 2 tools:

It got 103.31 to be exact. PCI riva TNT P233mmx. PII 450 on a BX board and a Voodoo 3, 100.52.
Yes on the P2 I was running the MTRRLFBE program too.

My board doesn't have dimm slots, but good to know it supports 256mb.
I wonder why the SIS's boards so fast with this game! 😲 thats pretty impressive to beat a PII that is on a 100mhz FSB.

On the VGA benchmark database: https://docs.google.com/spreadsheet/ccc?key=0 … drive_web#gid=0

There is an entry with a MMX 233 getting 108.09 with a GeForce2 MX on a GA-5AX. It is the first score over 100. But I agree, there is some scaling issue with Doom. Quake on the other hand is much better at showing differences. It scales to the moon 😀

Yea quake is much better, I think PCPbench does pretty good too.
3dbench put out a really high score on my Cyrix, for its low speed, so did doom actually, not as high as the P233mmx tho.

Phil: I found a SP97-XV for 20 dollars. Bought it 😀

I'll bench it when I get it

Two new boards benched:

Abit PE-5 (SiS 50X) Socket 5
PC Chips M519 (Opti Viper) Socket 7

The PC Chips board is the worst performing Socket 7 board I've ever seen

Phil: The SP97-XV was never shipped, the seller refunded me 🙁

vetz wrote:

Two new boards benched: […]
Show full quote

Two new boards benched:

Abit PE-5 (SiS 50X) Socket 5
PC Chips M519 (Opti Viper) Socket 7

The PC Chips board is the worst performing Socket 7 board I've ever seen

Phil: The SP97-XV was never shipped, the seller refunded me 🙁

how fast it that m519? and what voltage and cpu does it support?

vetz wrote:

Phil: The SP97-XV was never shipped, the seller refunded me 🙁

Oh that's a shame.

noshutdown wrote:

how fast it that m519? and what voltage and cpu does it support?

You can see the benches in the Google Docs document linked in the first post (I haven't updated the results in the thread yet).

It supports INTEL Pentium P54C 75~200 MHz, P55C w/VRM, Cyrix/IBM 6x86 P120+/P150+/P166+, AMD K5 PR75~PR166 CPUs.
Supports 3.3v/3.5v CPU's and clock speeds 1.5x/2x/2.5x/3x.
Also has a VRM socket, so 2.8V CPU's also possible.

AMD K6 might work as well, but haven't tested it.

Just repaired copper coils on ASUS TXP4. If it starts, I will do the bench.

By the way, I'm impressed with my P55T2P4 still holding the leadership. My 233MMX system with that same board is nowhere around the top in Phil's list.

RacoonRider wrote:

By the way, I'm impressed with my P55T2P4 still holding the leadership. My 233MMX system with that same board is nowhere around the top in Phil's list.

Me too, even the ATX version XP55T2P4 didnt manage to equal the performance, same with the very high end P65UP5 board (with dual 430HX CPU board).

Regarding the 233MMX, it's not clear who's the winner. Different graphics card is used, for instance its a known fact that the Voodoo3 and Nvidia cards are quicker in VGA than a Matrox Mystique.

Why does using SDRAM in the TX board not increase performance compared to EDO RAM? Do you see any performance benefit of SDRAM for faster CPU clocks? What about an 83 MHz FSB vs. 66 MHz?

feipoa wrote:

Why does using SDRAM in the TX board not increase performance compared to EDO RAM? Do you see any performance benefit of SDRAM for faster CPU clocks? What about an 83 MHz FSB vs. 66 MHz?

Haven't used the Aopen TX board much and I don't have it available for testing atm (stowed away). SDRAM was more for increased memory speeds than anything else. EDO vs SDRAM on earlier motherboards tend to yield little performance differences. You see more of a difference on the SS7 boards, but I don't think the EDO support was that great in them to begin with.

feipoa wrote:

Why does using SDRAM in the TX board not increase performance compared to EDO RAM? Do you see any performance benefit of SDRAM for faster CPU clocks? What about an 83 MHz FSB vs. 66 MHz?

There's a THG article on the matter (can't find it 🙁 ). According to it, not only does SDRAM offer little to no benefit on a TX board, it is also quite possible to run EDO at 75 and 83 MHz FSB. THG suggests that EDO back in the day was just as capable as SDRAM at being "tech of the future" and the fact that we did not see 100MHz EDO RAM was more for marketing reasons.

One should take into account that SDRAM we use today in TX is usually taken from much newer Celeron systems and therefore more mature. In other words, it might and should perform better then SDRAM used in TX days.

From http://geekscomputer.blogspot.no/2008/07/chip … tium-class.html

Most notable was the support for SDRAM, which was about 27% faster than the more popular EDO memory used at the time. Although the support for SDRAM was a nice bonus, the actual improvement in system speed derived from such memory was somewhat limited.

This was because with a normal L1/L2 cache combination, the processor read from the caches 99% of the time. A combined miss (both L1 and L2 missing) happened only about 1% of the time while reading/writing memory. Thus with SDRAM, the system would be up to 27% faster, but only about 1% of the time. Therefore, the cache performance was actually far more important than main memory performance.

I have a pcchips m590 and that has very crappy performance also (could be bad cache or something, i don't know)... is there any good pcchips socket7 board that performs at least very close to intel boards?

Added 2 more TX boards with SDRAM/EDO/FPM results.

The first one is an ASUS TXP4 I mentionned earlier. It had its CPU power curcuits repaired a while ago. With SDRAM and thightest BIOS settings it beats the current leader. The funny thing I noticed about this board that EDO gave worse results than FPM, while BIOS settings stayed constant during benchmarks.

The second one is Lucky Star 5I-TX2A. This one deserves a word. First of all, Lucky Star offered highly budget motherboards. It's got cheapo memory slots, especially SDRAM, which are so stiff it's a pain to pull out the memory without pulling out the board! It bends easily and is quite bended already after being mounted in the wrong way since 1998. It's also an ultra-compact board, almost as tiny as late 386DX-40 boards. However, it has a decent set of features from standard I/O and IDE to USB, PS/2 and IR. As we can see from the benchmarks, this one is a solid performer. The only serious downside is the fact that only one PCI slot is unbostructed and it obstructs the battery when used. Two out of three ISA slots can have full length cards.

I have never observed any benchmarked speed advantage of EDO over FPM. Maybe somebody knows a benchmark that shows some benefit?

If you ever plan on using an AMD K6-III on a 430TX board (assuming your BIOS lets you, or if there is a K6 patch available), I noticed that running the board at 83x6 or 75x6 and using 2x128 MB of SDRAM caused memory errors in MEMTEST, even with SDRAM set to CL3. The solution was to use 4x64 MB of EDO RAM. Alternately, you can run the K6-III at 66x6 with 2x128 MB SDRAM and not have any errors. EDO at 83 MHz is slightly faster than SDRAM at 66 MHz.

Two more boards added and front post updated. I could not replicate RacoonRider's result on the ASUS P/I-P55T2P4 board, though mine is a rev 3.0 board and his is a 3.1. Main difference is that I have 2x256kb cache, while his has 512kb integrated.

god, i totally forgot about this... I have a TYAN based dual socket 7 board i'd like to see run, and maybe a supermicro one if I can get it to post right.