Here are a couple more speed sensitive games to try:

Wing Commander 1 (fast 386/very slow 486 is optimal)
Trantor the Last Stormtrooper (XT class needed)
Stormlord (286)

Outside of those ranges the above are either unplayable or controls don't respond correctly.

With the aid of setmul for L1 cache and software multiplier settings and bios disabling of L2 cache I can get those playable on 120-133mhz 486/586 class machines.

They are examples of some earlier games where just one 'old' machine (say a 486) just won't cut it, and where a much more modern machine will struggle by not being able to be dropped low enough.

Ultima 7 is one of the excellent but sadly speed dependent games, was made for something around 486 33MHz and works bad on slower/faster systems.

Interesting news. Glad to hear people are having success with this. I have two TB kits but Adafruit is having delays as they move. So I'm still waiting on the two PICOs I ordered.

V

Okay I managed to get it set up in my case , I tried wing commander like megatron requested , slowed it down and the intro ran at that right speed , also for A10 tank killer , which has an issue with the title credits scrolling too fast now was running correctly.

Other than my mounting looking a bit ugly and one of the screws not reaching which is mostly cosmetic anyways , I have the bigger problem of the display going only up to 200 and not showing the correct cpu speed for some reason.

AppleSauce wrote on 2024-08-05, 17:28:

And we have success , I got the pc down to 4.77 MHZ and later tried out 33mhz and played some doom , I had to shrink the window down to get it to run faster which seems about right for a 486.

Interesting that it's reporting 4.77 MHz, but the CPU score is 15.4. I would be wrong, but is Speedsys is calibrated so that XT speed is 1.0?

I have found with ACPI/ODCM throttling that benchmarks and performance don't tend to correlate, so I imagine this is the same case with the Throttleblaster.

Btw, what are the specs of the PC you're using this on? Is this on a Pentium 233?

Shponglefan wrote on 2024-08-07, 20:22:

Interesting that it's reporting 4.77 MHz, but the CPU score is 15.4. I would be wrong, but is Speedsys is calibrated so that XT […]
Show full quote

AppleSauce wrote on 2024-08-05, 17:28:

And we have success , I got the pc down to 4.77 MHZ and later tried out 33mhz and played some doom , I had to shrink the window down to get it to run faster which seems about right for a 486.

Interesting that it's reporting 4.77 MHz, but the CPU score is 15.4. I would be wrong, but is Speedsys is calibrated so that XT speed is 1.0?

I have found with ACPI/ODCM throttling that benchmarks and performance don't tend to correlate, so I imagine this is the same case with the Throttleblaster.

Btw, what are the specs of the PC you're using this on? Is this on a Pentium 233?

I mean its possible but I hope its working , it could be that it wont dial into that slow of a speed , however at least it seems to be able to slow things down to run the games at the right pace.
Yeah its the same 233 I've been min maxing for ages.

The IPC of newer CPU is better, so where an 8088 took 20 cycles on some instructions the newer arch maybe only takes 4. That's why a V20 will speed up a classic 8088 4.77Mhz machine also.

Okay... I'll stick the meat in my sandwich, here's some details, https://www2.math.uni-wuppertal.de/~fpf/Uebun … 2/opcode_i.html check out the divide and multiply operations, whew.

edit: heh if you ever want to prove a 486 is twice as fast as a pentium write a benchmark that consists entirely of 4bit left and right shifts 🤣

Okay I've got it dialed in , turns out you need to wire up a button then long press to adjust the frequency with the pot then short press to exit , you can also set the PWM Period from 1-256.
Unfortunately my win95 ssd which was my secondary drive died so I can't test ecstatica 2 in windows like I wanted to, I tried to hook it up to my win 10 pc to access it but it hangs my pc after like being connected for 10 seconds and when I copy files , so ill need to get a replacement.

At least for now I can use the CF card with DOS 6.22 to play DOS stuff.
Also I've noticed when you dial in frequencies between 233 and like 60 odd mhz the computer produces a coil whine high pitched noise , I'm not sure if that's normal or not.

BitWrangler wrote on 2024-08-08, 18:00:

The IPC of newer CPU is better, so where an 8088 took 20 cycles on some instructions the newer arch maybe only takes 4. That's why a V20 will speed up a classic 8088 4.77Mhz machine also.

Okay... I'll stick the meat in my sandwich, here's some details, https://www2.math.uni-wuppertal.de/~fpf/Uebun … 2/opcode_i.html check out the divide and multiply operations, whew.

edit: heh if you ever want to prove a 486 is twice as fast as a pentium write a benchmark that consists entirely of 4bit left and right shifts 🤣

Right so what you're saying is :because Pentium cpus can execute instructions more quickly it will always inherently have a better score thus wont be exactly like running a 486 even at a lower clock speed?

Yes, in general a pentium needs to be running around 30% slower to approximate speed of a 486, thus a Pentium 60 is about equivalent to a DX4/100. However, instruction mix varies, and the speed critical parts of a game may have different characteristics game to game, thus trying to target a DX2/66 might have you running a Pentium at anywhere from 30-50Mhz effective on a per game basis.

It gets a bit weirder past 486 though because of more highly pipelined CPU cores, and the role L1 has in feeding those, which is why L1 shutoff can have such a large effect, as it might mean that the pipeline does not fill, making a proportion of clock cycles "empty", because not all spots in the line are taken.

AppleSauce wrote on 2024-08-10, 05:49:

Also I've noticed when you dial in frequencies between 233 and like 60 odd mhz the computer produces a coil whine high pitched noise , I'm not sure if that's normal or not.

That's probably normal. When I was testing ODCM / ACPI throttling with Pentium 4 processors, I noticed with Prescott processors I would get coil whine with certain settings.

When I switched to Cedar Mill processors, the coil whine completely disappeared.

AppleSauce wrote on 2024-08-08, 06:40:

I mean its possible but I hope its working , it could be that it wont dial into that slow of a speed , however at least it seems to be able to slow things down to run the games at the right pace.
Yeah its the same 233 I've been min maxing for ages.

If possible could you test Ultima 7?

I found with ODCM throttling it would cause significant slow downs during the screen shaking effect (earthquake) at the beginning of the game. It resulted in slow down of both the music and screen redraw times.

I'd be curious to see if the same issue manifests with the Throttle Blaster.

Shponglefan wrote on 2024-08-10, 15:19:

If possible could you test Ultima 7? […]
Show full quote

AppleSauce wrote on 2024-08-08, 06:40:

I mean its possible but I hope its working , it could be that it wont dial into that slow of a speed , however at least it seems to be able to slow things down to run the games at the right pace.
Yeah its the same 233 I've been min maxing for ages.

If possible could you test Ultima 7?

I found with ODCM throttling it would cause significant slow downs during the screen shaking effect (earthquake) at the beginning of the game. It resulted in slow down of both the music and screen redraw times.

I'd be curious to see if the same issue manifests with the Throttle Blaster.

Sure I can give it a whack , so I'd need the Forge of Virtue expansion installed as well since that apparently has a screen shaking bit at the beggining?

Shponglefan wrote on 2024-08-10, 15:17:

AppleSauce wrote on 2024-08-10, 05:49:

Also I've noticed when you dial in frequencies between 233 and like 60 odd mhz the computer produces a coil whine high pitched noise , I'm not sure if that's normal or not.

That's probably normal. When I was testing ODCM / ACPI throttling with Pentium 4 processors, I noticed with Prescott processors I would get coil whine with certain settings.

When I switched to Cedar Mill processors, the coil whine completely disappeared.

Ahk cool , good to know the cpu wasn't frying itself or something 😆

megatron-uk wrote on 2024-08-05, 18:29:

Here are a couple more speed sensitive games to try: […]
Show full quote

Here are a couple more speed sensitive games to try:

Wing Commander 1 (fast 386/very slow 486 is optimal)
Trantor the Last Stormtrooper (XT class needed)
Stormlord (286)

Outside of those ranges the above are either unplayable or controls don't respond correctly.

With the aid of setmul for L1 cache and software multiplier settings and bios disabling of L2 cache I can get those playable on 120-133mhz 486/586 class machines.

They are examples of some earlier games where just one 'old' machine (say a 486) just won't cut it, and where a much more modern machine will struggle by not being able to be dropped low enough.

Yeah So I've already tested Wing commander and slowed it down that worked
Tried Trantor and it went from basically being uplayable cause you run out of time in like 2 seconds to being totally playable and I was able to go around shooting enemies.
And stormlord was also playable after throttling down , but for some reason it mapped my right key to the numpad instead of the right arrow.

Shponglefan wrote on 2024-08-10, 15:19:

If possible could you test Ultima 7? […]
Show full quote

AppleSauce wrote on 2024-08-08, 06:40:

I mean its possible but I hope its working , it could be that it wont dial into that slow of a speed , however at least it seems to be able to slow things down to run the games at the right pace.
Yeah its the same 233 I've been min maxing for ages.

If possible could you test Ultima 7?

I found with ODCM throttling it would cause significant slow downs during the screen shaking effect (earthquake) at the beginning of the game. It resulted in slow down of both the music and screen redraw times.

I'd be curious to see if the same issue manifests with the Throttle Blaster.

Erm well the game mostly runs okay at 40 mhz , it might have had the screen lag a bit on the edges when drawing the screen like you said during the screen shake , I think the music was okay , but im using a roland CM32L with an MPU 401 so i dont know if the music has less issues with midi vs FM.

AppleSauce wrote on 2024-08-11, 13:56:

Erm well the game mostly runs okay at 40 mhz , it might have had the screen lag a bit on the edges when drawing the screen like you said during the screen shake , I think the music was okay , but im using a roland CM32L with an MPU 401 so i dont know if the music has less issues with midi vs FM.

The attachment 20240811_190103.jpg is no longer available

The attachment 20240811_192830.jpg is no longer available

I'm trying to remember if I was testing it with MPU-401 or Adlib... I think maybe Adlib? I'd have to try it again to be sure.

At any rate, thanks for testing, appreciate it.

Shponglefan wrote on 2024-08-13, 01:04:

AppleSauce wrote on 2024-08-11, 13:56:

Erm well the game mostly runs okay at 40 mhz , it might have had the screen lag a bit on the edges when drawing the screen like you said during the screen shake , I think the music was okay , but im using a roland CM32L with an MPU 401 so i dont know if the music has less issues with midi vs FM.

The attachment 20240811_190103.jpg is no longer available

The attachment 20240811_192830.jpg is no longer available

I'm trying to remember if I was testing it with MPU-401 or Adlib... I think maybe Adlib? I'd have to try it again to be sure.

At any rate, thanks for testing, appreciate it.

Happy to help.

The Serpent Rider wrote on 2024-04-09, 02:08:

and early Socket A (both with ISA) platforms were already considered as best solution for your buck as a universal retro system.

Can XP-M / Athlon 4 mobile go below 33mhz 386 speeds (with cache disabled, CPUSPD etc) without resorting to such “hardware” slowdown method?

I am confused. What is this STPCLK pin that you are connecting to on the Pentium 233? I can find reference to that pin on Pentium 4. But not on Pentium 1.... does it sometimes go by another name? I have a book that lists the pinout of the Pentium, and it doesn't mention STPCLK.