Maybe the fan on the CPU is not making proper contact, and it is causing the CPU to fall back to 1X mode.

I tried that several times already. The contacts look fine, though. It makes no difference. When it's slow, the Quake Benchmark under DOS only is like 80 Seconds, 12 FPS....

I've seen similar behavior with bad hard drives. When a drive is full of delayed sectors (hdd regenerator shows them), the system would be normal sometimes but when a delayed sector is reached, the performance goes to hell because of the time needed to read those sectors.

I fixed this issue forcing hdd regenerator to fix those delayed sectors and that's all. No more performance issues and my systems works great.

Note: sometimes the drive is just too bad. I have a lot of hdds that I couldn't fix. They stuck forever at delayed sectors.

Well, I tried several things. I found out that when I have disabled Virtual memory, the speed is fast in GLQuake. I was able to optimize settings so, the Benchmark went to 37 Seconds and 25FPS. Too bad, at 64MB, I guess I ran out of memory rather quick. I was only able to let the demo run, if I want to play the game, opengl32.dll crashes...

If disabling Virtual Memory makes things go fast then its the HDD that is going bad, just as ramiro77 said.

Try the following: Power up, boot, run a benchmark.
Don't turn off, instead hit reset and run the bench again.

Then let me know what happens.

I have a HOT433 with a similar issue.

I have had all kinds of unpredictable problems with the HOT-433 motherboard, and I have met 6 boards face-to-face in the past many years. I was unable to determine exactly what the cause of the seemingly random behaviour was.

What about trying speedsys? You can compare your results with another system.

alexanrs: That is also my own observation. Since windows swaps as soon as the game runs at 64MB on the board. (I will try more later)

sprcorreia: The results are exactly the same, depending the configuration. However, the benchmark results are almost unchanged under DOS. The main problem is the 3DFX Performance.

feipoa: As soon as I can get my hands on the machine again, I will test the issue of Virtual memory. I tried to install Win98 on harddrive, but had all kinds of random problems. I will use a CF Card next time.

ramiro: I found out that gaming benchmarks actually make game related problems more visible than synthetic benchmarks. I think Quake is a great tool to find out if a computer has any kind of problem. 😁

However, to get specific results, I need specific methods.

My Mainboard has some weirdness I need to explain, though.:

When I insert 4 chips for cache, to total 256 KB, and correctly jumper the cache, it always says "Cache bad! Don't enable the cache!". If I switch these chips to another bank, its the same. If I fit the whole 512KB inside, the message does not appear. Hmm. Not that this matters. With the same config, cache set to 2-2-2, the impact on GLQuake is 30.2 FPS with L2 cache enables vs. 28.5 FPS with L2 cache disabled. 😀

Exin, I have run into similar problems with this motherboard and was unable to find a solution. It is only my opinion, but I think you are waisting your time with this motherboard. At one point, I thought I got my board working fine. I installed Windows 98SE, ran some benchmarks, turned it off, powered it back up, and received errors after loading Windows. I spent many days with these motherboards - their problems are not organised or predictable. I suspect that you are going to tackle it anyway, so here is some advice from my sufferings,

Use FPM memory only - not EDO.
Use an Am5x86-133, not a POD.
Do not use the onboard IDE. Use an ATA, SATA, or SCSI PCI card.
Use double-banked cache only, that means, 9 cache modules (256 KB or 512 KB).
Set L2 cache into WT mode.
Use no more than 3 PCI cards.
Use an S3 Virge.
Forget about the PS/2 port. It is not wired correctly on both versions, 1-4.
Use BIOS version 433AUS33.ROM. It is an AMI-based BIOS.

If you continue to have trouble, I have altered a Biostar MB-8433 BIOS to work with the HOT-433, but it was selective about which PCI cards went into which slots.

Also, someone else on Vogons posted the following BIOS, which may be a later revision that 433AUS33.ROM. I am also attaching it, but I have not tested it personally.

- It is only my opinion, but I think you are waisting your time with this motherboard. At one point, I thought I got my board working fine. I installed Windows 98SE, ran some benchmarks, turned it off, powered it back up, and received errors after loading Windows.

I tried to install Win98 on a "large" WD Drive(6GB). Things went haywire, you don't want to hear that....I think there's a ghost inside this board. 😀
I guess the only way will be finding the perfect drive for this board and go with it. See below why.
At the beginning, for like 5 days, software reset did not work. I always had to reset manually. Then, out of the sudden, reset works again....

I think there is still hope. My only other option are two identical boards with via chipset, since the tomato board suddenly disappeared....

-Use FPM memory only - not EDO.

I found out that that does not change anything with that board.

-Use an Am5x86-133, not a POD.

Too bad two of the legs just fell off of mine. I guess that 133Mhz CPU is a goner.

-Do not use the onboard IDE. Use an ATA, SATA, or SCSI PCI card.

Well, after trying out many different cards and harddrives, I came up with the following conclusion:
Adaptec Controllers simply won't work. ISA or PCI, but PCI Controllers don't work at all. Windows identifies them, installs the driver, yet the drivers always say the card is not present. I tried several driver/card combos.
AMD Based SCSI Controllers may work "not installed in windows", in BIOS Emulation mode. But then, randomly it happens that the data on the HDD gets corrupted. If I install the Windows driver, as soon as I access the drive, blue screen. Older Win95 driver rendered the HDD Useless, the HDD is reported as 5GB instead of 18GB.
SIL 3114 Sata controller. Windows detects it, however claims that the driver is incorrect. Tries the drivers from different sources. Not a chance.

Too bad I don't have many options in non-raid SCSI/SATA controllers....

Other cards:
SIS Graphics Card gets way too hot for its own good.
PowerVR gets detected, driver however claims the card is not present after installation.
Other Raid controllers: Mainboard will not even boot from internal IDE Drives
ISA Based Graphics cards: Tried many, no picture.
ISA Sound cards (SB Pro) however work.

-Use double-banked cache only, that means, 9 cache modules (256 KB or 512 KB).

....oh.

-Set L2 cache into WT mode.

I found that does not make any difference. Even with Cache turned off, the board behaves this way.

-Use no more than 3 PCI cards.

I guess you could say as well not use any cards at all. The board is even picky about ISA Graphics cards...

-Use an S3 Virge.

I guess I will end up doing so. The board does not function with S3 Vision cards, it seems, tried 2 of them. No picture.

-Forget about the PS/2 port. It is not wired correctly on both versions, 1-4.

Never intended to.

-Use BIOS version 433AUS33.ROM. It is an AMI-based BIOS.

Yes, that is what I'm doing.

-Also, someone else on Vogons posted the following BIOS, which may be a later revision that 433AUS33.ROM. I am also attaching it, but I have not tested it personally.

Hmmm.... I hope the keyboard still works after flashing. 😁

Feiopa: Tried your bios. The board is now dead, thanks. 😜 It does boot, but not reacts to keyboard (mostly). I can reach the bios, the board somewhat scans for the keys at some point for a short while. 😜

Exin, I didn't post my BIOS. I mentioned that the BIOS I posted was posted by someone else and that I have not tested it.

I have noticed that some BIOSes don't like to reflash well using certain flashing software brands/versions. It is always best to flash unfamiliar BIOSes with an external writer.

....which only a few people have. Now i can sell it as garbage. Not that it was a good board anyways.

You may be able to hot swap.

Actually, i thought about that. Just need to find a board that flashes the same way with the same size bios

Exin wrote:

Actually, i thought about that. Just need to find a board that flashes the same way with the same size bios

You can use "uniflash", it's less fussy about ROM sizes and such, if it can identify the flash chip and the ROM fits in it it will probably be able to flash it.

You might need to jumper the board to 5V or 12V flash programming voltage to be able to flash, read the chips documentation to see what programming voltage it wants.

For the meantime, I use a MS4144.

For everyone who wants to know how to jumper the board for Pentium Overdrive. I found a page in the last Manual revision that is not written down anywhere else:

JC1&JC3: Set
JC2&JC4:Not set
JC5:3&4;5&6
JC6:3&4
JC8&JC7:Not set

Source: http://www.elhvb.com/mobokive/Archive/Microst … nual/index.html

You have an MS-4144? That is a superior board. Why even bother with the HOT-433 when you have an MS-4144? The MS-4144 can be modified to accept a native-mode PS/2 mouse.