Mechanical cpu overclocking

[NumLock21]

Back in them old school days, cpu were overclockes mechanically. They need to bring these back. Not only are they a lot more stable than bios oc, it also lets post mechanical cpu oc techies, get a better understanding on how cpu multipilers work, and be one with their PC. Much like a manual vs a automatic.

 

 

[Mira Yurizaki]

1 hour ago, NumLock21 said:

Back in them old school days, cpu were overclockes mechanically. They need to bring these back. Not only are they a lot more stable than bios oc, it also lets post mechanical cpu oc techies, get a better understanding on how cpu multipilers work, and be one with their PC. Much like a manual vs a automatic.

I don't think how you overclock produces something more stable than the other. If anything, a more "stable" overclock is simply due to the looser tolerances back then when the time domain was still in tens to hundreds of nanoseconds, rather than hundreds of picoseconds like today.

 

Also I don't think flipping switches would provide a better understanding of how CPU multipliers work. I mean... the concept is simple: You have a base clock, you have some circuitry in the CPU that multiplies it to get the actual running clock.

 

EDIT: Also two more problems with the DIP switch method:

• You're limited to whatever the DIP switches provided. So in your picture, you can't go beyond 12.5 or 8.5, even if the processor could actually handle it.
• The system cannot dynamically control how fast the processor operates. I don't know about you, but I kind of like my processor running slower if it's not going to do much

[NumLock21]

8 minutes ago, M.Yurizaki said:

I don't think how you overclock produces something more stable than the other. If anything, a more "stable" overclock is simply due to the looser tolerances back then when the time domain was still in tens to hundreds of nanoseconds, rather than hundreds of picoseconds like today.

 

Also I don't think flipping switches would provide a better understanding of how CPU multipliers work. I mean... the concept is simple: You have a base clock, you have some circuitry in the CPU that multiplies it to get the actual running clock.

 

EDIT: Also two more problems with the DIP switch method:

• You're limited to whatever the DIP switches provided. So in your picture, you can't go beyond 12.5 or 8.5, even if the processor could actually handle it.
• The system cannot dynamically control how fast the processor operates. I don't know about you, but I kind of like my processor running slower if it's not going to do much

Bios oc used for today has improved over the past decade but hardware oc with them dip switches is still better, no cpu overclocking fail errors. It's either work or don't work. It also lets techies get a better understanding because you can't just pop a cpu in, even running at stock setting requires the adjustment of dip switches. This drives them to learn how and where their cpu clock speed comes from.

[Mira Yurizaki]

12 minutes ago, NumLock21 said:

Bios oc used for today has improved over the past decade but hardware oc with them dip switches is still better, no cpu overclocking fail errors. It's either work or don't work.

I mean, I guess the only thing this saves is remembering what the last known working multiplier was. But that doesn't really seem like a good enough advantage over what advantages software settings provide.

12 minutes ago, NumLock21 said:

It also lets techies get a better understanding because you can't just pop a cpu in, even running at stock setting requires the adjustment of dip switches. This drives them to learn how and where their cpu clock speed comes from.

Not really. It just tells them "I have to set these switch things based on which CPU I'm using."

 

I'd rather have the conveniences of modern software switching than deal with this. Would you rather like to get rid of PnP and deal with manually setting DMA, IRQs, and hardware addresses because it should drive people to learn about how system peripheral communication works? I'm pretty sure most people from that time period couldn't care less. They just wanted to get their damn device to work.

[it_dont_work]

This was fun, remember fucking with resistors and mainboards to fine tune voltages, those were the days...well besides all the fuck chips and mobo's

 

And it was not really mechanical, just no digital voltage regulation or multiplier. Still had to physically desolder and change parts to mess with vcore in any useful way. Most chips wouldn't post with anything other than the needed settings as well since they were binned, I had a couple of 233's that wouldn't post past 266mhz. the difference realistically was not worth it. 

 

EDIT: the fact this is new to people makes me feel old. stop that 

[NumLock21]

2 hours ago, M.Yurizaki said:

I mean, I guess the only thing this saves is remembering what the last known working multiplier was. But that doesn't really seem like a good enough advantage over what advantages software settings provide.

Not really. It just tells them "I have to set these switch things based on which CPU I'm using."

 

I'd rather have the conveniences of modern software switching than deal with this. Would you rather like to get rid of PnP and deal with manually setting DMA, IRQs, and hardware addresses because it should drive people to learn about how system peripheral communication works? I'm pretty sure most people from that time period couldn't care less. They just wanted to get their damn device to work.

Well I care cause it lets me know exactly which IRQ is reserved for which device. Now everything is automatic which is good, but that connection between you and the machine is somewhat lost.

 

1 hour ago, it_dont_work said:

This was fun, remember fucking with resistors and mainboards to fine tune voltages, those were the days...well besides all the fuck chips and mobo's

 

And it was not really mechanical, just no digital voltage regulation or multiplier. Still had to physically desolder and change parts to mess with vcore in any useful way. Most chips wouldn't post with anything other than the needed settings as well since they were binned, I had a couple of 233's that wouldn't post past 266mhz. the difference realistically was not worth it. 

 

EDIT: the fact this is new to people makes me feel old. stop that 

For me it isn't new, for some yes.

[Mira Yurizaki]

2 hours ago, NumLock21 said:

Well I care cause it lets me know exactly which IRQ is reserved for which device. Now everything is automatic which is good, but that connection between you and the machine is somewhat lost.

That knowledge is only really useful if you're manually setting IRQs for other devices. Other than that, it doesn't matter.

 

I mean, if you really want to go down this road, maybe we should start suggesting everyone go back to 80s home computers with BASIC interpreters as your only "OS"