If you didn't change the crystal oscillator or the clock gen when swapping CPUs, then yes, you were overclocking the CPU.

Perhaps I've missed something, but is this the same board?

Need help with identifying this 486 board

http://museum.ttrk.ee/th99/m/C-D/33087.htm

thanks a lot for the link!

Yes is exactly this mainboard

with the same cmos problem 😒

I want to try again the Cyrix DX2 / 50 with switch configuration and I want to try to expand the video ram, I have some similar chip at home

mav85ale wrote:

with the same cmos problem 😒

I have not had to do this yet, but it seems straightforward. I like badmojo's in situ approach:
It's 286 time!

Yes it's a good solution,

Although it is not a really problem config hard disk parameters every boot

mav85ale wrote:

thanks a lot for the link! […]
Show full quote

thanks a lot for the link!

Yes is exactly this mainboard

with the same cmos problem 😒

I want to try again the Cyrix DX2 / 50 with switch configuration and I want to try to expand the video ram, I have some similar chip at home

The Cyrix 486 DX2 50MHz works fine now and seems faster than the Intel 486DX 33MHz

Doom 1 and 2 are not very playable, but the problem is the ram

The Cyrix 486 chips were said to be 5-10% slower than the intel parts at the same clockspeed.

The DX/2-50 will be a little faster than the DX-33 at CPU intensive things, but fall behind at I/O intensive stuff. I remember reading the reviews in PC World and PC Magazine in early 1993, and those two CPUs were roughly equal.

mav85ale wrote:

Doom 1 and 2 are not very playable, but the problem is the ram

Using 'doom -nosound -timedemo demo3', my 486dx33 only does 10.4 fps with 32mb of 4mb SIMMs.

My box is a little slow though. There's a big comparison chart linked from 'Phil's Ultimate VGA Benchmark' thread, if you haven't seen that already.
Phil's Ultimate VGA Benchmark Database Project

=EDIT=
Did a fresh test just now, and I get 10.4 FPS with only 4MB installed, and with 32MB.

I noticed an improvement going from a Opti 931 soundcard to a Creative SB16/Vibra CT4180... it's possible?

I will try tonight the timedemo

The timedemo goes slowly.... the results are 1091 gametics in 7382 realtics

with ultimate doom 3863 gametics in 18690 realtics

35 x 1091 / 7382 = 5.17 FPS... Yeah, that's pretty slow. I'm surprised.

Is your Turbo button on/off ? (depending on how it's wired, and what 'turbo' does for that board.) Try a test in the other position.

With ultimate doom

35 x 3863 / 18690 = 7,23 FPS

Yes the turbo button is on, when I switch off the difference it's clear

I should have been specific that my FPS was from using Doom shareware 1.9 with 1 level of green border. For comparison, I've just run Phil's Ulitmate VGA pack, and the Doom result there is 8.5 FPS. (2.2 FPS with the Turbo switch. Zow.)

Maybe the poor performance are also guilty of integrated Paradise/WD 90C30 or mainly of ram?

mav85ale wrote:

Maybe the poor performance are also guilty of integrated Paradise/WD 90C30 or mainly of ram?

I'm not sure... I had a WD Paradise ISA card in my Commodore 386SX-16 (I think it was a 90C30 as well, not sure).
When I built my own 486DX2-66 a few years later, I re-used that card until I could afford a good VLB card.
It actually ran Doom quite well, only slightly slower than the VLB card I had later.
RAM is probably not important... cache could be?

mav85ale wrote:

Maybe the poor performance are also guilty of integrated Paradise/WD 90C30 or mainly of ram?

For interest
Re: 1992 top hardware
The Perfect IBM PC AT
https://books.google.ca/books?id=v9TVJ_G_sk8C … egrated&f=false

I'm using a modest TVGA 8900C. I tried 4MB ram, mentioned above; there was no difference. My cache is the full 256kb.

Perhaps also check autoexec.bat and config.sys?

All cache slot are filled, and switch cache are set to 64KB

every 8 cache chipset are "ISSI IS61C64A-25N 9227"

in the comparative of the pc mag. article, the best notebook mounts just the wd 90C30!.. : 😲

and in other thread the WD 90C30 it's compared to ET4000 😢

problem solved! In bios was the "internal cache" disable!

thanks to both! 😀

now doom time demo

35x1091/2035 = 18,764! 😁

EDIT:

I'm found another voice "external cache=absent set to present" now the system recognize 64KB Cache but doom decrease at 16,918 fps

Are you sure you are calculating your FPS properly?

I think so.. No?

1091 gametics in 2035 realtics with cache external disabled

1091 gametics in 2252 realtics with 64KB of cache external enabled

it's correct?