33% higher is hardly a 'small' increase. Yes, that would be significantly faster. Overall performance doesn't quite scale with FSB clock, but it comes close, particularly at high multipliers, where RAM access really is a big bottleneck. The P3-1000EB was generally faster than the 1100E...

But you might have trouble finding the 1133. It was a paper release by Intel to one-up AMD that failed, it wasn't stable and Intel got caught out by various review sites. It got quietly withdrawn. Later the speed was re-released as a Coppermine-T, but that is an FC-PGA2 CPU that may not be compatible with your motherboard.

Much easier option: go for a 1000EB. Very easy to find. Save the 1100E for systems that can't do 133MHz FSB.

How would I identify a 1000EB? I have a Socket 370 1GHz chip (this is a S370 i815 board) but I don't know what stepping it is or whatever.

See the 1000/256/133.

The alternative, 1000E, would be 1000/256/100.

IMHO if it won't shorten the life of the product or do other damage, why not?
Don't cost nothin'...
Seriously though every little bit can help. Overall performance is a sum of all tweaks.

dionb wrote on 2020-04-19, 22:43:

But you might have trouble finding the 1133. It was a paper release by Intel to one-up AMD that failed, it wasn't stable and Intel got caught out by various review sites. It got quietly withdrawn. Later the speed was re-released as a Coppermine-T, but that is an FC-PGA2 CPU that may not be compatible with your motherboard.

Am I wrong if I think the FC-PGA2 is just because of the integrated heat spreader? And that the 1133 was made in FC-PGA2 as both Coppermine as well as Coppermine-T?

A 33% higher FSB won't cause any big step in performance if the rest of the system doesn't change. Well, you need to change RAM, which should be obvious. But first mistake is thinking that PC133 RAM is generally 33% faster than PC100 and makes the system fly. Cheap RAM isn't, it just has the timings adjusted to work with 133 MHz. Derived busses stay the same (PCI, AGP, ISA, etc.) regardless of the FSB, so you gain nothing there.

Unless you put in a CPU that is not only using 133 MHz FSB but also has a noticeably higher clock rate, you gain less than 10% - if at all.

H3nrik V! wrote on 2020-04-20, 08:32:

dionb wrote on 2020-04-19, 22:43:

But you might have trouble finding the 1133. It was a paper release by Intel to one-up AMD that failed, it wasn't stable and Intel got caught out by various review sites. It got quietly withdrawn. Later the speed was re-released as a Coppermine-T, but that is an FC-PGA2 CPU that may not be compatible with your motherboard.

Am I wrong if I think the FC-PGA2 is just because of the integrated heat spreader? And that the 1133 was made in FC-PGA2 as both Coppermine as well as Coppermine-T?

That is not so easy to answer because there is a certain amount of ambiguous info on the web re these matters.
There was one 1.133Mhz cpu (mentioned above) from the cC0 stepping, sSpec SL4HH, released in july 2000 and called back in august the same year but it was a slot 1 version.
Starting june 1st 2001 Intel released three 1.133Mhz cpu's from the cD0 stepping, SL5B2, SL4YV, SL5QK and these came in FC-PGA2 package with heat spreader.
According to the Intel spreadsheet (see below) the SL4YV (which occurs 2x in the spec sheet) was also released in a (FC-PGAxvI) package which some internet sources seem to have 'translated' into a regular socket 370 package (but who knows?).
There is also mention that some of the so called Coppermine-T cpu's differed from regualar Coppermines in so far that they could run on 1.25 volt,
http://www.cpu-world.com/sspec/SL/SL5QK.html
these were meant for servers but would probably? run also normally on regular socket 370 boards.

My machine can't overclock so going down to a lower clocked but higher FSB CPU (I checked my 1GHz chip and it's indeed the 1000EB) probably wouldn't make much difference...I was running PC133 RAM already but I have no idea of the timings and such of the RAM sticks.

PARKE wrote on 2020-04-20, 11:27:

That is not so easy to answer because there is a certain amount of ambiguous info on the web re these matters. There was one 1.1 […]
Show full quote

H3nrik V! wrote on 2020-04-20, 08:32:

dionb wrote on 2020-04-19, 22:43:

But you might have trouble finding the 1133. It was a paper release by Intel to one-up AMD that failed, it wasn't stable and Intel got caught out by various review sites. It got quietly withdrawn. Later the speed was re-released as a Coppermine-T, but that is an FC-PGA2 CPU that may not be compatible with your motherboard.

Am I wrong if I think the FC-PGA2 is just because of the integrated heat spreader? And that the 1133 was made in FC-PGA2 as both Coppermine as well as Coppermine-T?

That is not so easy to answer because there is a certain amount of ambiguous info on the web re these matters.
There was one 1.133Mhz cpu (mentioned above) from the cC0 stepping, sSpec SL4HH, released in july 2000 and called back in august the same year but it was a slot 1 version.
Starting june 1st 2001 Intel released three 1.133Mhz cpu's from the cD0 stepping, SL5B2, SL4YV, SL5QK and these came in FC-PGA2 package with heat spreader.
According to the Intel spreadsheet (see below) the SL4YV (which occurs 2x in the spec sheet) was also released in a (FC-PGAxvI) package which some internet sources seem to have 'translated' into a regular socket 370 package (but who knows?).
There is also mention that some of the so called Coppermine-T cpu's differed from regualar Coppermines in so far that they could run on 1.25 volt,
http://www.cpu-world.com/sspec/SL/SL5QK.html
these were meant for servers but would probably? run also normally on regular socket 370 boards.1133a.JPG

As to what I can see, the Coppermine is part no 80526, Coppermine-T is 80533 and Tualatin is 80530. As I understand it, the Coppermine-T is a Coppermine core in all respects, but using the AGTL bus-interface as the Tualatin does.

Coppermine: http://www.cpu-world.com/CPUs/Pentium-III/Int … 26PZ006256.html
Coppermine-T: http://www.cpu-world.com/CPUs/Pentium-III/Int … 33PZ006256.html
Tualatin: http://www.cpu-world.com/CPUs/Pentium-I ... 3256).html
Tualatin-S: http://www.cpu-world.com/CPUs/Pentium-I ... 3512).html

Edit: So in theory, a "real" Coppermine is made in 1133, but yes, it wasn't on the market for long, before it was withdrawn for stability issues, so probably not an easy find ..

H3nrik V! wrote on 2020-04-20, 13:53:

Coppermine: http://www.cpu-world.com/CPUs/Pentium-III/Int … 26PZ006256.html Coppermine-T: http://www.cpu-world.com/CPUs/Pentium- […]
Show full quote

Coppermine: http://www.cpu-world.com/CPUs/Pentium-III/Int … 26PZ006256.html
Coppermine-T: http://www.cpu-world.com/CPUs/Pentium-III/Int … 33PZ006256.html
Tualatin: http://www.cpu-world.com/CPUs/Pentium-I ... 3256).html
Tualatin-S: http://www.cpu-world.com/CPUs/Pentium-I ... 3512).html

Edit: So in theory, a "real" Coppermine is made in 1133, but yes, it wasn't on the market for long, before it was withdrawn for stability issues, so probably not an easy find ..

Are the Coppermine-T chips the ones with an IHS? I've found those occasionally but I don't know if they can directly replace a normal Coppermine chip.

H3nrik V! wrote on 2020-04-20, 13:53:

9><CUT
Edit: So in theory, a "real" Coppermine is made in 1133, but yes, it wasn't on the market for long, before it was withdrawn for stability issues, so probably not an easy find ..

In theory everything is possible. It is even possible that Intel produced them in the form of special orders for OEM parties.
But I have not yet come across a photo of a non-T 1.133Mhz Coppermine so 'theory' is fine by me 😀.
Do you have a source for the 'withdrawn for stability issues' report for s370 releases ?

Re backwards compatibility: according to the Wiki page they run on 'older' systems.
https://en.wikipedia.org/wiki/Pentium_III

EvieSigma wrote on 2020-04-19, 20:55:

I have a Pentium III system with a 1100MHz Coppermine chip (11x100FSB). There was also a Coppermine 1133, which uses a 133FSB. Would I even notice a difference in performance going from 100 to 133FSB or is it not worth the effort of trying to find a chip?

That's a real gem, extremely rare.
So if you want to switch to FSB 133 then just buy a random Coppermine 1000. Yes, I know there are 100MHz types, but very rare, too, when you buy one on ebay you will get 99% one with FSB 133MHz.

CoffeeOne wrote on 2020-04-20, 20:36:

EvieSigma wrote on 2020-04-19, 20:55:

I have a Pentium III system with a 1100MHz Coppermine chip (11x100FSB). There was also a Coppermine 1133, which uses a 133FSB. Would I even notice a difference in performance going from 100 to 133FSB or is it not worth the effort of trying to find a chip?

That's a real gem, extremely rare.
So if you want to switch to FSB 133 then just buy a random Coppermine 1000. Yes, I know there are 100MHz types, but very rare, too, when you buy one on ebay you will get 99% one with FSB 133MHz.

I actually already have one, but I'm still not clear if the lower clock speed but higher FSB and (potentially) higher RAM performance is worth swapping.

It is worth it. Keep the 1100/100 in a vault. Put the 1000/133 in and enjoy!

EvieSigma wrote on 2020-04-20, 13:39:

My machine can't overclock so going down to a lower clocked but higher FSB CPU (I checked my 1GHz chip and it's indeed the 1000EB) probably wouldn't make much difference...

Out of interest - what machine/which motherboard?

I was running PC133 RAM already but I have no idea of the timings and such of the RAM sticks.

As for capabilities, check the chips on the PC133 DIMMs. The last digit on the chip code indicates latency. SDR-SDRAM can run at CAS2 or CAS3 timings (actually a whole lot more, but the rest are neither bottleneck for clocking nor for performance). 7.5ns latency supports CAS3, CAS2 needs 7ns or lower.

Of course, if your chips support it you still need to activate it. Possibly your motherboard does so automatically via SPD, but I'd always double-check in BIOS to be sure.

It's a Gateway OEM version of the Intel i815EEA motherboard, swapped into an older Slot 1 P3 Gateway case.

EvieSigma wrote on 2020-04-21, 14:15:

It's a Gateway OEM version of the Intel i815EEA motherboard, swapped into an older Slot 1 P3 Gateway case.

I hope that you will benchmark both the 1100/100 and the 1000/133 and report back here - I, for one, am curious after the difference.

PARKE wrote on 2020-04-21, 14:21:

EvieSigma wrote on 2020-04-21, 14:15:

It's a Gateway OEM version of the Intel i815EEA motherboard, swapped into an older Slot 1 P3 Gateway case.

I hope that you will benchmark both the 1100/100 and the 1000/133 and report back here - I, for one, am curious after the difference.

Sure, I'll do some runs. My video card is a Diamond Viper V770 32MB so that might be a tiny bit of a bottleneck compared to a GeForce 256 or a Radeon but I should still be able to get some good results.