Disable cache and hitting "DX2-66"? \ VOGONS

Disable cache and hitting "DX2-66"?

Topic actions

First post, by brostenen

Posted on 2015-09-14, 22:26

brostenen Offline

Rank l33t++

Rank: l33t++
Posts: 6670
Joined: 2014-05-07, 21:13
Location: Northern Jutland, Denmark.

Hi all.
There is really a lot of talk about taking one fast system, and turning it into a 386/486 class machine by disabling cache on CPU/Mobo.
Now I am wondering. Wich exact system will give a DX2-66, when disabling CPU Cache only, and leaving the Mobo Cache enabled?

Don't eat stuff off a 15 year old never cleaned cpu cooler.
Those cakes make you sick....

My blog: http://to9xct.blogspot.dk
My YouTube: https://www.youtube.com/user/brostenen

001100 010010 011110 100001 101101 110011

Reply 1 of 8, by boxpressed

Posted on 2015-09-14, 23:01

boxpressed Offline

Rank Oldbie

Rank: Oldbie
Posts: 1438
Joined: 2014-05-20, 18:29
Location: Wisconsin, USA

I have a K6-3+ on an FIC VA-503+ (VIA MVP3). I run the CPU at 5.5x and 100 MHz. When I use SETMUL to change the multiplier to 2 and to disable the L1 cache, Speedsys gives me a rating that is similar to a 66 MHz 486.

Reply 2 of 8, by brostenen

Posted on 2015-09-14, 23:34

brostenen Offline

Rank l33t++

Rank: l33t++
Posts: 6670
Joined: 2014-05-07, 21:13
Location: Northern Jutland, Denmark.

Thanks. Yet I was more looking into just disabling L1 in order to get that performance.

Don't eat stuff off a 15 year old never cleaned cpu cooler.
Those cakes make you sick....

My blog: http://to9xct.blogspot.dk
My YouTube: https://www.youtube.com/user/brostenen

001100 010010 011110 100001 101101 110011

Reply 3 of 8, by boxpressed

Posted on 2015-09-15, 00:09

boxpressed Offline

Rank Oldbie

Rank: Oldbie
Posts: 1438
Joined: 2014-05-20, 18:29
Location: Wisconsin, USA

I have a KT133A system with an Athlon XP 2400+, and when the internal cache is disabled. Speedsys also gives me a rating close to a 66MHz 486.

Reply 4 of 8, by brostenen

Posted on 2015-09-15, 21:50

brostenen Offline

Rank l33t++

Rank: l33t++
Posts: 6670
Joined: 2014-05-07, 21:13
Location: Northern Jutland, Denmark.

Tried to disable L1 on my P133. It is somewhere around 486 25mhz to 50mhz.
Some benchmark tools say one thing, others say another.

Don't eat stuff off a 15 year old never cleaned cpu cooler.
Those cakes make you sick....

My blog: http://to9xct.blogspot.dk
My YouTube: https://www.youtube.com/user/brostenen

001100 010010 011110 100001 101101 110011

Reply 5 of 8, by oerk

Posted on 2015-09-15, 22:02

oerk Offline

Rank Oldbie

Rank: Oldbie
Posts: 631
Joined: 2014-08-20, 09:59

brostenen wrote:
Thanks. Yet I was more looking into just disabling L1 in order to get that performance.

Why? As long as the end result is the same, why do you care?

Reply 6 of 8, by brostenen

Posted on 2015-09-15, 22:08

brostenen Offline

Rank l33t++

Rank: l33t++
Posts: 6670
Joined: 2014-05-07, 21:13
Location: Northern Jutland, Denmark.

oerk wrote:
brostenen wrote:
Thanks. Yet I was more looking into just disabling L1 in order to get that performance.

Why? As long as the end result is the same, why do you care?

Because it is nice to have the numbers. Yeah... It's the same thing about GFX benchmarks.
Why know what cards that you don't own can crunch? Because it is nice to know.

Actually. Knowing these CPU numbers, are something that are not covered here on Vogons.
Well... Not extensively covered, as in making a database with as many CPU/Mobo combo's as possible.
Perhaps someone should make a chart, just as we have these GFX benchmark charts. 😁

Don't eat stuff off a 15 year old never cleaned cpu cooler.
Those cakes make you sick....

My blog: http://to9xct.blogspot.dk
My YouTube: https://www.youtube.com/user/brostenen

001100 010010 011110 100001 101101 110011

Reply 7 of 8, by Agent of the BSoD

Posted on 2015-09-15, 23:43

Agent of the BSoD Offline

Rank Member

Rank: Member
Posts: 194
Joined: 2011-03-11, 01:26
Location: Oregon

On my K6-III+ with FIC PA-2013, when I ran it at 100MHz at 4x multiplier and disabled L1, I got about DX2 50 speeds. Could probably get up to 66 by increasing the multiplier further.

Reply 8 of 8, by PhilsComputerLab

Posted on 2015-09-15, 23:52

PhilsComputerLab Offline

Rank l33t++

Rank: l33t++
Posts: 6174
Joined: 2014-09-28, 03:33
Location: Western Australia

In all the cache experiments I did, and I did quite a few 😊 , I found that the DX2-66 performance is not obtainable. Even if you pick the fastest cache-less processors (MMX and Cyrix), I got up to a DX2-50 if I remember correctly. That is still pretty decent IMO.

EDIT: On that chart it looks very close. Maybe you can hit the speed, this was all years ago...

I found this older image of mine:

The black shaped area shows what speeds you cannot obtain with cache tricks on various processors.

YouTube, Facebook, Website

Go to top of page Go to top of page

Back to General Old Hardware