• Please review our updated Terms and Rules here

Speed Issues On Games Despite Meeting System Requirements

If Your Motherboard has cache chips that plug in IC sockets then they most likely are replaceable with new ones from Ebay or Aliexpress.
If they are soldered direct onto the motherboard then The only easy fix is a new motherboard.

Could You please take a photo of the entire Motherboard with cards removed if possible?

PS. If they are plugged in then reseating them might fix it, but that is a long shot.
 
Last edited:
Sure, I can take a pic of the motherboard. Won't be until tonight though.

How do I get them out? Do they just pull straight out?
 
Don't pull anything other than cards that plug into either PCI/VLB/ISA slots for now.

Removing socketed IC's can be tricky and You can bend pins if You aren't very careful.

I replaced the AMD 133 chip in my 486 with a Intel DX2-66 for comparisons as well as old times sake.

Ran Landmark 2 (There is a version 6 as well). I have a PCI 2MB S3 Trio32 btw, not a Trident.

Doom got 31.3 FPS.

Cachechk also ran
 

Attachments

  • Dsc00667a.jpg
    Dsc00667a.jpg
    71.8 KB · Views: 1
  • Dsc00668a.jpg
    Dsc00668a.jpg
    100.4 KB · Views: 1
Last edited:
Decided to just go the new mobo route. Snagged that ASUS board for $30, so I think it's a pretty cheap option to try out. Here's hoping that cache isn't fake/damaged either!
 
Imperious said:
Don't pull anything other than cards that plug into either PCI/VLB/ISA slots for now.

Removing socketed IC's can be tricky and You can bend pins if You aren't very careful.

I replaced the AMD 133 chip in my 486 with a Intel DX2-66 for comparisons as well as old times sake.

Ran Landmark 2 (There is a version 6 as well). I have a PCI 2MB S3 Trio32 btw, not a Trident.

Doom got 31.3 FPS.

Cachechk also ran
Click to expand...

See, your landmark settings are similar to mine (except for the video card, of course). So am I going crazy over nothing here? Bahhhhh.

I should have a nice collection of older pc parts once all is said and done.
 
Guybrush3pwood said:
No idea what any of this means.

Click to expand...

The main memory speed is horrible (as was the speed to videomem), so I think that is the issue there (it should be 30+ MB/s).
That may have to do with cache and/or waitstate settings.
The 8 KB cache is performing fine, so the CPU is okay and running at ~66 MHz.
There appears to be a problem with the L2 cache, because it is reporting that 'megabyte #4 isn't being cached!' and "This machine seems to have one cache!?".

The problem might be as simple as just re-seating the cache-chips though.
 
Scali said:
The main memory speed is horrible (as was the speed to videomem), so I think that is the issue there (it should be 30+ MB/s).
That may have to do with cache and/or waitstate settings.
The 8 KB cache is performing fine, so the CPU is okay and running at ~66 MHz.
There appears to be a problem with the L2 cache, because it is reporting that 'megabyte #4 isn't being cached!' and "This machine seems to have one cache!?".

The problem might be as simple as just re-seating the cache-chips though.
Click to expand...

How do I reseat cache though? I looked online and didn't see any tutorials. I don't want to break anything.
 
Guybrush3pwood said:
How do I reseat cache though? I looked online and didn't see any tutorials. I don't want to break anything.
Click to expand...

Just carefully lift the chips out of the sockets a bit, by sticking something flat underneath them, eg a small flathead screwdriver or a pair of pliers. Be careful not to lift them too much on one side, because then you'll bend the legs.
Then push them back down again, perhaps wiggle them around a bit, so the sockets are scraped clean somewhat.
On your photos the chips don't seem to be seated that well, some stick out quite far.
 
Guybrush3pwood said:
Ok, I'll give it a shot. Should they be seated all the way down?
Click to expand...

Yup, push them down as far as they will go... they may not be entirely flush with the socket, but should be much tighter than some in your picture (the one with the red dot seems to be seated best).
 
Still not convinced that the cache is the problem here. I had one of the infamous "fake L2 cache" 486 boards in the days and actually performance was like 10% worse than a legit board, something you could definitely measure but barely "feel". It's like your CPU runs at 60 MHz instead of 66... Pretty bad, but not the reason for "slideshow" performance in games imho.


I doubt it's the video card, I mean VLB was pretty fast compared to ISA and at low resolution in System shock (320x240) it seems so unlikely that the graphics card is the bottleneck here, otherwise doom would also be that slow, keep in mind this is not a 3D card doing the rendering work, it's just coughing out frames.


I remember that back in the days there were also fake CPUs around. But after the benchmark program identified it as "Intel" and "66 MHz" that seems unlikely, right?
 
kyodai said:
Still not convinced that the cache is the problem here. I had one of the infamous "fake L2 cache" 486 boards in the days and actually performance was like 10% worse than a legit board, something you could definitely measure but barely "feel". It's like your CPU runs at 60 MHz instead of 66... Pretty bad, but not the reason for "slideshow" performance in games imho.


I doubt it's the video card, I mean VLB was pretty fast compared to ISA and at low resolution in System shock (320x240) it seems so unlikely that the graphics card is the bottleneck here, otherwise doom would also be that slow, keep in mind this is not a 3D card doing the rendering work, it's just coughing out frames.


I remember that back in the days there were also fake CPUs around. But after the benchmark program identified it as "Intel" and "66 MHz" that seems unlikely, right?
Click to expand...

Now I have a fake cpu? Ah c'mon.

Also, just replaced the 8mb of ram with 32mb. It made zero difference in the benchmarks, but that's hardly surprising, no?

I'm at a loss. The only thing I can say is that the motherboard is bottlenecking it. Either that or the HDD. I have another I could test out, but without an external floppy drive to make some 6.22 disks I have no way of installing it. Ah well.
 
Guybrush3pwood said:
Now I have a fake cpu? Ah c'mon.

Also, just replaced the 8mb of ram with 32mb. It made zero difference in the benchmarks, but that's hardly surprising, no?

I'm at a loss. The only thing I can say is that the motherboard is bottlenecking it. Either that or the HDD. I have another I could test out, but without an external floppy drive to make some 6.22 disks I have no way of installing it. Ah well.
Click to expand...

You don't have a fake cpu, forget that. Cachecheck couldn't even find that You had any external motherboard cache, that is the problem.
 
Imperious said:
I would move the cache chips around and re-run cache check. They are all the same so shift the bottom 4 up the top and vice versa.
Click to expand...

I agree with this statement. Earlier today I did this with the cache on my own 486 (I hadn't noticed any specific performance issues, but I ran a benchmark and thought the score was lower than it should be) and discovered that one of the chips actually had a bunch of bent pins. Unbent them and now my cache is working and my system is actually getting benchmark results similar to other machines with the same CPU.
 
Imperious said:
I would move the cache chips around and re-run cache check. They are all the same so shift the bottom 4 up the top and vice versa. That's if You can bothered as You
have a new mobo on the way.

You can get them here http://www.aliexpress.com/item/IS61C256AH-20N-IS61C256AH-15N-IS61C256A-20N/1784090287.html
I'd just try the new mobo first though.
Click to expand...

Interesting, I'll keep that in mind.

jmetal88 said:
I agree with this statement. Earlier today I did this with the cache on my own 486 (I hadn't noticed any specific performance issues, but I ran a benchmark and thought the score was lower than it should be) and discovered that one of the chips actually had a bunch of bent pins. Unbent them and now my cache is working and my system is actually getting benchmark results similar to other machines with the same CPU.
Click to expand...

At this point, I don't feel like taking the motherboard out and putting it back in again. I'll wait for the new motherboard and see where we stand there.
 
Guybrush3pwood said:
Either that or the HDD.
Click to expand...

The HDD won't be bottlenecking Doom with 8 MB, let alone 32 MB. The whole game will fit in memory, so it's only on start of the game that the HDD affects the performance, and whenever it needs to load new level data. But it won't affect the general framerate in-game.
 
You must log in or register to reply here.
Back
Top