VOGONS


First post, by jasa1063

User metadata
Rank Oldbie
Rank
Oldbie

Since building a 486 Pentium Overdrive system, I have found using the Turbo Button and/or disabling the L1 cache has been more effective than I anticipated for running older DOS games that may require more than XT or 286 system could handle. I was looking at 386 build, but now I really don't need one. I was wondering if others have had similar results.

Last edited by jasa1063 on 2020-10-09, 20:56. Edited 1 time in total.

Reply 2 of 12, by jasa1063

User metadata
Rank Oldbie
Rank
Oldbie
Doornkaat wrote on 2020-10-09, 20:38:

That's what the button is meant for. 😉
Some implementations are better than others - nice to hear that yours works well for you.👍

I phrased it wrong. I know the what the Turbo Button's function was supposed to do. I was just surprised how well it worked on my particular build. In my case it cuts the CPU speed by a factor of 75%. I updated my verbiage it my original post to reflect that.

Reply 3 of 12, by Doornkaat

User metadata
Rank l33t
Rank
l33t

Ah, ok. I was kind of confused about your post but it was still nice to hear. 😅
75% deturbo sounds actually quite good! I only ever actually measured the deturbo on my Pentium 60 where the it actually halves the CPU bus frequency. The computer runs almost exacly half as fast in everything I tried it with.
What board are you using?

Reply 4 of 12, by jasa1063

User metadata
Rank Oldbie
Rank
Oldbie
Doornkaat wrote on 2020-10-09, 21:02:

Ah, ok. I was kind of confused about your post but it was still nice to hear. 😅
75% deturbo sounds actually quite good! I only ever actually measured the deturbo on my Pentium 60 where the it actually halves the CPU bus frequency. The computer runs almost exacly half as fast in everything I tried it with.
What board are you using?

I am using a Lucky Star LS486E Rev C1. My Pentium Overdrive 83Mhz runs at about 20Mhz when using the Turbo Button.

Reply 6 of 12, by jasa1063

User metadata
Rank Oldbie
Rank
Oldbie
Doornkaat wrote on 2020-10-09, 21:33:

About 20MHz actual internal frequency or equivalent speed to a 20MHz Pentium?
In case of the former I'd assume the fsb is actually lowered to ~8MHz?

From what I can measure it's the FSB that is getting lowered, not the actual CPU frequency.

Reply 7 of 12, by Doornkaat

User metadata
Rank l33t
Rank
l33t

Again, I'm a little confused: Lowering the fsb frequency results in a lower CPU core clock. The POD83MHZ would run at ~20MHz if the fsb is lowered to ~8MHz.
Maybe the deturbo function utilises a similar approach as discussed in this thread?
Do you have a 486DX-33 to try the deturbo on? I'm really curious how low that'll go. 😁

Reply 8 of 12, by clueless1

User metadata
Rank l33t
Rank
l33t

From what I understand, different motherboards de-turbo differently. Some manipulate memory timings, some FSB, some clock multiplier, etc. On mine, it will slow a 66Mhz 486 to roughly 486/20Mhz speeds. And even better, it will work in addition to L1 cache disabling. Here's an example based on tons of benchmarking and comparing to known results from other systems:
486/66 with turbo and L1 enabled = 66Mhz 486
486/66 with turbo disabled and L1 enabled = 20Mhz 486
486/66 with L1 disabled only = 25Mhz 386
486/66 with L1 and turbo disabled = 10Mhz 286

The more I learn, the more I realize how much I don't know.
OPL3 FM vs. Roland MT-32 vs. General MIDI DOS Game Comparison
Let's benchmark our systems with cache disabled
DOS PCI Graphics Card Benchmarks

Reply 9 of 12, by The Serpent Rider

User metadata
Rank l33t++
Rank
l33t++

Lucky Star LS486E Turbo button = halved FSB clock. So minimal speed you can achieve is 12 Mhz FSB and 30 Mhz on POD83. Potentially 12 Mhz on POD, if you disable cooler.

I must be some kind of standard: the anonymous gangbanger of the 21st century.

Reply 10 of 12, by jasa1063

User metadata
Rank Oldbie
Rank
Oldbie

I know Norton SI 8.0 is not the most accurate benchmark, but running at full speed the score for my LS486E with a Pentium Overdrive is 264 and shows 83Mhz. De-turboed the score drops to 66 and shows 20Mhz.

Reply 11 of 12, by Azarien

User metadata
Rank Oldbie
Rank
Oldbie
clueless1 wrote on 2020-10-10, 10:34:
From what I understand, different motherboards de-turbo differently. Some manipulate memory timings, some FSB, some clock multi […]
Show full quote

From what I understand, different motherboards de-turbo differently. Some manipulate memory timings, some FSB, some clock multiplier, etc. On mine, it will slow a 66Mhz 486 to roughly 486/20Mhz speeds. And even better, it will work in addition to L1 cache disabling. Here's an example based on tons of benchmarking and comparing to known results from other systems:
486/66 with turbo and L1 enabled = 66Mhz 486
486/66 with turbo disabled and L1 enabled = 20Mhz 486
486/66 with L1 disabled only = 25Mhz 386
486/66 with L1 and turbo disabled = 10Mhz 286

I always thought that disabling Turbo turns off the clock multiplier (i.e. sets it to 1), so that 486 DX2 would work at half the speed and 486 DX4 would work at a quarter speed.
This is the first time I read that it may do different things depending on the motherboard..

Reply 12 of 12, by The Serpent Rider

User metadata
Rank l33t++
Rank
l33t++

so that 486 DX2 would work at half the speed

DX2 multiplier can't be changed by any means. Internal multiplier was a novelty at that time.

I must be some kind of standard: the anonymous gangbanger of the 21st century.