Well it depends ...how about you feed it the missing .6v and see what happens, if it works then its undervolted, if it burns then it wasnt. (.2v increments shouldn't hurt it too much as long as it has a decent cooler)

Either way you have your answer.

TrashPanda wrote on 2022-03-11, 08:40:

Well it depends ...how about you feed it the missing .6v and see what happens, if it works then its undervolted, if it burns then it wasnt. (.2v increments shouldn't hurt it too much as long as it has a decent cooler)

Either way you have your answer.

Yeah, I was hoping to get my answer without blowing up my cpu 😂

Sphere478 wrote on 2022-03-11, 08:49:

TrashPanda wrote on 2022-03-11, 08:40:

Well it depends ...how about you feed it the missing .6v and see what happens, if it works then its undervolted, if it burns then it wasnt. (.2v increments shouldn't hurt it too much as long as it has a decent cooler)

Either way you have your answer.

Yeah, I was hoping to get my answer without blowing up my cpu 😂

heh its a Cyrix, if it isn't dead by now you wont be killing it.

Sphere478 wrote on 2022-03-11, 08:27:

So if a pentium mmx (0.35uf) runs at 2.8v […]
Show full quote

So if a pentium mmx (0.35uf) runs at 2.8v

But a cyrix 433gp works at 2.2v (and is also 0.35uf)

Does that mean that the cyrix is under volted?

Or is 2.8 unsafe for the cyrix?

Process technology is one factor in voltage, but so is design within that process. You can't assume all chips on same process will require or indeed survive the same voltage. A chip is specified for a certain voltage. That specification may be conservative or very challenging, but regardless, over or undervolting is defined as deviating from spec, not from whatever might be usual on a certain process node.

In general I'd say that Intel's P55C was massively overengineered, extremely robust and therefore very tolerant of strongly deviating voltages. They are known to survive for years at 3.3V (although I would NOT recommend that), can generally be massively overclocked at stock voltage and many people undervolt them by at least 10% at stock speeds. Conversely, Cyrix was always battling the limits of the more primitive chip fabs they had to work with, frequently ran very hot, and were known to have very little headroom - either for overclocking or for messing around with voltage in either direction. Also consider the P55C was a 1996 design released in early 1997, when the 433GP was 1998 design not available until mid 1999. It's really comparing apples and oranges.

Sphere478 wrote on 2022-03-11, 08:27:

So if a pentium mmx (0.35uf) runs at 2.8v […]
Show full quote

So if a pentium mmx (0.35uf) runs at 2.8v

But a cyrix 433gp works at 2.2v (and is also 0.35uf)

Does that mean that the cyrix is under volted?

Or is 2.8 unsafe for the cyrix?

Dude, there's a reason the P55C maxed ot at 233 while that Cyrix does 300.
Those 350 nm were not the whole story.
I've read somewhere that earlier Cyrix were notorious for varying quality because they got bad deals with their manufacturing contractors, namely IBM.
But the MII was build by National Semiconductor.

Also, this page
http://www.cpu-galerie.de/html/cyrixmii.html# … 20II-400GP-3x95
tells us that there was a low voltage version of the MII that was done in 250 nm - which would explain a lot.

So, you're probably not going to kill it, but probably shorten its life expectancy, I'd assume.

My K6-III+ 400ATZ is rated at 1.8 V. It'll gladly take more but if you wanted to run it at 1.8 V, now that's a socket 7 board that's hard to find!

Usually the ibm branded cpus were more tollerant to oc compared to Cyrix.

About the MII there are indeed 2 versions, v2.9 and v2.2 (there are rare batches of v2.7 too, I have one).
There are some subtle architerctural differences between the v2.2 and v2.9, at the same frequency the higher voltage is a bit faster.

Last but not least, my pentium 233 mhz is able to run at 300 without issues, so despite no official changes the last batches of pentium mmx were able to achieve higher frequencies than Cyrix Mii. Years ago there was a guy who was selling the last produced batch of pentium mmx (on cpuworld forum and then ebay) which were even more overclockable.

Process technology is not directly connected with the voltage, as dionb said. It also depends on the architecture of the cpus, otherwise with today nanomerters we should have a negative cpu voltages 😁

Nemo1985 wrote on 2022-03-11, 14:29:

Usually the ibm branded cpus were more tollerant to oc compared to Cyrix.

Yep, that's more precisely what I had read about the Cyrix fabs at IBM.
They did not outright select their yield and gave the shitty ones to Cyrix.
But, they had several production lines and their own chips were made on the more sophisticated ones.

Cuttoon wrote on 2022-03-11, 14:42:

Yep, that's more precisely what I had read about the Cyrix fabs at IBM. They did not outright select their yield and gave the sh […]
Show full quote

Nemo1985 wrote on 2022-03-11, 14:29:

Usually the ibm branded cpus were more tollerant to oc compared to Cyrix.

Yep, that's more precisely what I had read about the Cyrix fabs at IBM.
They did not outright select their yield and gave the shitty ones to Cyrix.
But, they had several production lines and their own chips were made on the more sophisticated ones.

This was only true after the National takeover , neither Cyrix nor IBM much cared for national and in fact, I don’t believe any IBM made chips bore the Cyrix logo until 1997,
Cyrix had to contract out to others until that point, once the deal with IBM was fractured by National that’s when you saw strange pricing and announcements from IBM

One such announcement was the IBM will produce k6-2-333 and Cyrix MII-333 and IBMs enhancements show the Cyrix 333 will be faster than the k6-2 and will charge a premium on the Cyrix chip.

IBM actually designed and forced the “new” 333 chip due to Nationals incompetence

So was 333 the fastest ibm cyrix chip?

Btw, I made a edit to OP

Cuttoon wrote on 2022-03-11, 11:53:

Dude, there's a reason the P55C maxed ot at 233 while that Cyrix does 300. Those 350 nm were not the whole story. I've read som […]
Show full quote

Sphere478 wrote on 2022-03-11, 08:27:

So if a pentium mmx (0.35uf) runs at 2.8v […]
Show full quote

So if a pentium mmx (0.35uf) runs at 2.8v

But a cyrix 433gp works at 2.2v (and is also 0.35uf)

Does that mean that the cyrix is under volted?

Or is 2.8 unsafe for the cyrix?

Dude, there's a reason the P55C maxed ot at 233 while that Cyrix does 300.
Those 350 nm were not the whole story.
I've read somewhere that earlier Cyrix were notorious for varying quality because they got bad deals with their manufacturing contractors, namely IBM.
But the MII was build by National Semiconductor.

Also, this page
http://www.cpu-galerie.de/html/cyrixmii.html# … 20II-400GP-3x95
tells us that there was a low voltage version of the MII that was done in 250 nm - which would explain a lot.

So, you're probably not going to kill it, but probably shorten its life expectancy, I'd assume.

My K6-III+ 400ATZ is rated at 1.8 V. It'll gladly take more but if you wanted to run it at 1.8 V, now that's a socket 7 board that's hard to find!

I have several pentium mmx chips that will do 300

Neat find on the link!

Looking at mfg dates and clocks and voltages I think the low voltage chips are all using the same core as the 433gp also interesting to see that there is a 433gp version of lv (at the same voltage 🤣)

Nemo1985 wrote on 2022-03-11, 14:29:

Usually the ibm branded cpus were more tollerant to oc compared to Cyrix. […]
Show full quote

Usually the ibm branded cpus were more tollerant to oc compared to Cyrix.

About the MII there are indeed 2 versions, v2.9 and v2.2 (there are rare batches of v2.7 too, I have one).
There are some subtle architerctural differences between the v2.2 and v2.9, at the same frequency the higher voltage is a bit faster.

Last but not least, my pentium 233 mhz is able to run at 300 without issues, so despite no official changes the last batches of pentium mmx were able to achieve higher frequencies than Cyrix Mii. Years ago there was a guy who was selling the last produced batch of pentium mmx (on cpuworld forum and then ebay) which were even more overclockable.

Process technology is not directly connected with the voltage, as dionb said. It also depends on the architecture of the cpus, otherwise with today nanomerters we should have a negative cpu voltages 😁

Any links to more info on this last batch?

These charts have absolute maximums listed for many CPUs, https://www.pchardwarelinks.com/elec_pentium.htm or VNE as it can be called, Voltage Not to be Exceeded. However, these are "it can tolerate it for a few seconds while the board detects CPU and puts it right" rather than can be run at that for long, so you want to be a couple or three notches down from maximum for even extreme overclocks.

Sphere478 wrote on 2022-03-11, 18:49:

Any links to more info on this last batch?

It was something discussed a couple of years ago on cpuworld forum, I can't find any link actually, sorry.

Nemo1985 wrote on 2022-03-11, 22:52:

Sphere478 wrote on 2022-03-11, 18:49:

Any links to more info on this last batch?

It was something discussed a couple of years ago on cpuworld forum, I can't find any link actually, sorry.

No prob, thanks 😀