Are you sure nothing's connected to the header that could trigger the green pc feature ?

Deksor wrote:

Are you sure nothing's connected to the header that could trigger the green pc feature ?

For no particular reason I had connected the turbo switch to it, which then had a solid light at all times so when I started diagnosing this speed issue I disconnected it from the MB header. It currently doesn't have anything attached. Would that slow the CPU while allowing normal use of the computer?

On a similar note, is it possible the computer would run slow due to lack of voltage or would it just flake out? Your video card is twice as fast as mine so I think we're dealing with something on the MB.

The green pc might be enabled for some reasons, maybe that's the culprit

For the video card, it's normal : mine has 4MB of SDRAM while yours have 2MB of EDO. And that model of rage is really really fast, an S3 will run at a speed closer to 20MBps

To my knowledge I can't disable it from the BIOS... it's supposed to be open circuit for OFF right?

Gahhhrrrlic wrote:

To my knowledge I can't disable it from the BIOS... it's supposed to be open circuit for OFF right?

From what I can find, the Green PC button or LED connector is for the Sleep function... so that should have nothing to do with the slowness.

Are you sure you have the CPU jumpers set properly?

I'm wondering if you have it set for the Cyrix "133Mhz" (110Mhz - 55x2) instead of the Intel 133Mhz. Not sure how the CPU speed is actually reported on that board.
JP36 should be on 3&4 and 5&6
http://stason.org/TULARC/pc/motherboards/S/SH … um-HOT-557.html

Can you post a picture of your board so we can see exactly how all the jumpers are set?

The CPU speed is reported at POST.

As for the power reduction, you can enable/disable this within the bios in the power management setup.

Deksor wrote:

The CPU speed is reported at POST.

As for the power reduction, you can enable/disable this within the bios in the power management setup.

Yeah the speed is reporting 133 everywhere I check. I can look at the jumpers 1 more time anyway.

All that power reduction stuff I disabled in the BIOS but... it's weird - from a comp-sci / physics point of view, I can't even understand how the speed could be reported correctly, yet the throughput is low. it's like one of the pipelines is busted or something but then you wouldn't expect the system to work properly if there was any component damage... at least that would seem like an incredible fluke.

Gahhhrrrlic wrote:

Just for thoroughness, I set the entire BIOS to defaults and no difference in score.

what defaults? did you test both optimal and safe defaults if present?

The mobo provides only "BIOS default" or "SETUP default", what I used was "BIOS default" and I had no performance issue with that.

But who knows, maybe the CMOS holds some hidden feature that a "BIOS" or "SETUP" default won't reset, but the mobo uses a DALLAS accumulator, so clearing that is going to be harder than usual, or should-I say, will take a long time

I can't read the entire thread now, but have you actually tried benchmarking the system? Phil's VGA benchmark should be a pretty good tool to find out if you've got a speed deficiency, try some of the later 3D tests such as Quake and PC Player Benchmark.

A CPU can't be half-dead. It either works or it doesn't and they usually die hard, unless you have one with an exposed die (think socket 370 Coppermines and socket A Athlons). Motherboards also have components that die, but that won't affect speed, the component will just... die. For example, the southbridge which contains the IDE controller can die.

Don't give up on this, but expect to find that perhaps this is how the system performs. I'll see if I can install my Pentium 133 on a similar system and see how it performs on an S3 Virge, but I don't have a Rage II unfortunately. I'll try Tomb Raider 2 and maybe we can compare that way.

Deksor wrote:

The mobo provides only "BIOS default" or "SETUP default", what I used was "BIOS default" and I had no performance issue with that.

But who knows, maybe the CMOS holds some hidden feature that a "BIOS" or "SETUP" default won't reset, but the mobo uses a DALLAS accumulator, so clearing that is going to be harder than usual, or should-I say, will take a long time

I just had that rebuilt by one of our members, Jed, so there's a battery on it now.

F2bnp wrote:

I can't read the entire thread now, but have you actually tried benchmarking the system? Phil's VGA benchmark should be a pretty good tool to find out if you've got a speed deficiency, try some of the later 3D tests such as Quake and PC Player Benchmark.

A CPU can't be half-dead. It either works or it doesn't and they usually die hard, unless you have one with an exposed die (think socket 370 Coppermines and socket A Athlons). Motherboards also have components that die, but that won't affect speed, the component will just... die. For example, the southbridge which contains the IDE controller can die.

Don't give up on this, but expect to find that perhaps this is how the system performs. I'll see if I can install my Pentium 133 on a similar system and see how it performs on an S3 Virge, but I don't have a Rage II unfortunately. I'll try Tomb Raider 2 and maybe we can compare that way.

Deksor may have been right in that the video card was performing at spec all this time. If my CPU is scoring lower than it should at 133, then the games may be CPU-bottlenecked instead of what I originally thought (that it was GPU bottlenecked).

Video speed shouldn't slow down speedsys though

Sorry to resurrect an old thread but I would like to buy a new (identical) CPU (P54CQS) and have been unable to find one for sale on normal outlets. I figured, if there is internal damage in the CPU such as a busted pipeline, the CPU speed may report correctly but the instructions that could be executed would be curtailed sufficiently to reduce the bench scores.

Deksor, didn't you say you carry all CPU models? I don't suppose you'd be willing to sell one...

If a new CPU doesn't fix the problem, then the motherboard becomes the most likely candidate. I found precisely 1 ebay ad for the same motherboard. Unfortunately the guy wants some 400 bucks for it because he personally refurbished it with fancy heatsinks and capacitors and stuff.

I doubt the CPU would cause the issue because CPUs rarely break, and I don't see how a broken CPU could compute things normally, but slower.

I don't remember what have been tried so excuse me if this question have already been answered : have you tried to flash the latest BIOS and did you reset the bios to optimal defaults ?

Did you ever consider the horrible ISA sound card DMA overhead?
This can slow down computers by ~30%.
Remove the ISA sound card and do the tests again, there might be a *big* difference.

OK, so I broke down and bought myself a CPU. It was cheaper and easier than buying a replacement motherboard at 400 bucks (yes that's what the guy on ebay was charging).

Guess what? It WAS the cpu. It was damaged.

Why did it still work? I dunno but my naive theory at the time was that maybe a pipeline wasn't working due to internal circuit damage but the other one was. I have experienced hardware failures before that didn't completely disable the function of the device so I think this sort of thing IS possible but probably unlikely. With the replacement CPU (I bought the exact same model code) I now get a score of 96 which is perfectly in line with what it should be for a P133.

The video card speed did NOT change. Nor did the hard drive speed or anything else. They are what they are. However they seem to be enough because all the games that were running like crap before are running reasonably now. CPU-bottleneck.

Thank GOD it wasn't the motherboard though. That would have been devastating.

In other news... I have a 120 megahe.... I MEAN 75 megahertz CPU for sale 😀