The Voltage for the PCI cards needs to be created oncard from the 5V line. That produces a lot of wasted heat inside the case. A V3 AGP card expects to be fed 3.3V through the AGP port. If you are lucky, the board delivers those 3.3V directly from the power supply. If not, an onboard regulator does the job, and will likely fail when it meets a V3 3000. Well known problem of the time.

If you want a low power PC, just take something like an ASRock K7S41gx with an AMD Geode, fall back on 200 MHz FSB and undervolt to minimum settings. You'll draw about the same you can get with a P2, but have a lot of computing power available.

ux-3 wrote:

.....
If you want a low power PC, just take something like an ASRock K7S41gx with an AMD Geode, fall back on 200 MHz FSB and undervolt to minimum settings. You'll draw about the same you can get with a P2, but have a lot of computing power available.

Yes I know there are better options, someone else suggested going super 7 for it's good flexibility. I'm just trying to stay within the slot 1 area since thats what this topic was originally about 😉

Tetrium wrote:

someone else suggested going super 7 for it's good flexibility.

I think that I made that suggestion, but perhaps someone else did so before.

retro games 100 wrote:

ux-3 wrote:

I have not figured out yet, how I can use my Barton XP-M on a KT133 board, which offers 1xISA.

Interesting. Are we talking QDI Kinetiz 7E-A mobo here? If yes, there's 3 things to consider. 1) What's the mobo's PCB revision number? You can see this in the bottom right hand corner, I think. (It's in one of the corners.) Your best bet is the PCB that has "2.0(2.3)" written in the corner, and also, the caps are the silver coloured ones, not the brown coloured ones. 2) What is your BIOS ROM version? You're best bet is the very last 4.1 version. 3) What XP-M CPU are you using? I know for sure that the 2400+ rated CPU (1800mhz real clock speed) works fine in this particular board. I vaguely recall that other XP-Ms did not work, for example the 2200+ rated CPU.

We are talking Kinetiz 7E-A. Its the 2.0(S2.2), brown colored caps. Bios is 4.1 LCP. I am using a Barton based XP-M 2600+. Should operate at 2000 MHz. Which, by the way, the standard XP-2400 with 256 MB cache should do too.
In fact, I have only tried the 1800 MHz Duron, which primes, but does crash in the Win98 startup. I am not quite sure, as as to why this happens. I haven't stuck my Barton into this board yet. I heard rumours that others fried theirs that way 😢

ux-3 wrote:

.....I haven't stuck my Barton into this board yet. I heard rumours that others fried theirs that way 😢

Well, you only live once! ...or maybe your Barton only lives once...ah nevermind!! 😜

ux-3 wrote:

I haven't stuck my Barton into this board [Kinetiz 7E-A / 2.0(S2.2)] yet. I heard rumours that others fried theirs that way 😢

Yes. In this post -

ISA PCI AGP mobos showdown: Intel 440 BX versus AMD 751/756 versus VIA KT133A :-) (Page 3)

I said - "I accidentally killed the XP-M CPU. I removed it from the Abit board, and put it in the QDI Kinetiz board. I've tested this CPU out many times in this board, without any problem."

I vaguely recall that the heatsink wasn't positioned on the CPU properly. The XP-M worked just fine in that board, many times. But it's also true that this board did kill it, in the end. One thing I recall is that it puts about 1.55-1.6V through the CPU. The XP-M I used only needed 1.45V. But then again, I've increased the voltate on that XP-M to 1.68V, and no problems occured.

I guess it's a bit of a risk, but then it does add to the excitement! 😉

I've killed a XP-M before myself. Cracked the core a little too much. I really don't miss the naked core days at all.

I killed a K6-III+ that way too. Removed the heatspreader because I wanted to cool it better (completely pointless with those chips btw). They are VERY easy to crack. I had even set up foam pads on the corners but that didn't protect it adequately.

It must be a gut wrenching feeling to accidentally destroy new hardware, back in the day. The significantly reduced prices of this stuff on ebay now takes quite a lot of the pressure off! 😀

Indeed. I once fried a Palomino at work while trying to get some motherboard to post. I had forgotten to add thermal grease, the CPU made a nice little impression of the written data on the die onto the heatsink, in mirror. Could actually read it just fine! 😁

Edit: Just recently I had received my 9 K6 mobiles and began to mass test them. Like I always do, I first set the fsb, multiplier and voltage to the lowest the system will allow.
Of the 9 times I had powered the system 2 times without a heatsink on the CPU and both times I disconnected the power within a second. Both mobiles are still fine! 😁

Look guys, I have no intention to kill my unlocked (factual) XP-3200+. And the fact that it was sold as XP-M 2600+ doesn't make the loss any more charming. I am currently in the process to try a retro 800 MHz all in one machine. If I succeed, the Athlon will get the job as Lan party client.

Ok 🤣, sorry about the frying part. Got a lil carried away 😜

If it helps you any, theres always the possibility to buy a lower rating (cheaper 😉 ) XP-M and use that for your retro rocket.
It will mean you have a slightly lower upper limit but, as those mobiles are a lot more common, you'll have all the flexibility without the risk of damaging your really good XP-M.

Myself I got a thoroughbred 2000+ mobile I pulled from a dead laptop which I use when I want to play with the unlocked clock frequency. It can go as low as the most expensive mobiles but can reach 1.8Ghz without even breaking a sweat. I'm fairly sure it can overclock much higher but I never actually tried with this particular chip.

Harvested an SL2HC, obviously terrible for power consumption but will be interesting to see if it's unlocked.

How come we don't have any Katmai fans here? I feel like I was the only guy with a Katmai P3 back in the day. There were even a few steppings that overclocked super well, such as SL35D which was a 450/100 that would happily do 600/133 very frequently cuz it had the faster cache chips. A nice upgrade from my SL2W8 PII 300@450. Never did the Celeron thing. 😉

swaaye wrote:

How come we don't have any Katmai fans here? I feel like I was the only guy with a Katmai P3 back in the day. There were even a few steppings that overclocked super well, such as SL35D which was a 450/100 that would happily do 600/133 very frequently cuz it had the faster cache chips. A nice upgrade from my SL2W8 PII 300@450. Never did the Celeron thing. 😉

I've definitely got some Katmai P3s, eg 450/100, and look forward to retesting them all as soon as I can. I recall that they all have curiously light weight fanless heatsinks. Which mobo did you use to get yours to go to 133fsb?

Some of my "lower speed" P3s are 450/100, 500/100, 550/100, 600/100. Regarding the 600, I think I've got a few different versions, one of them being a 370 socket chip mounted on a cheap sloket adapter.

BTW, I've got various P2s including 66fsb and 100fsb versions. The interesting thing is that their fanless heatsinks are much heavier/chunkier than any of the faster P3 chips.

My very first computer was a P2-350 and a couple years later I got a Katmai 550 for it. Performance was noticably better 😀

I also have an Intel server motherboard laying around, dual slot, which came with a single P2. I've upgraded it with 2 Katmai 600's. Haven't put the system back together again though, and at boot it reports the 2 cpu's as 533Mhz 133 fsb??
The mobo is an Intel Nightshade, very large motherboard!

swaaye wrote:

How come we don't have any Katmai fans here?

The katmai is very much similar to the pentium II deschutes, with the added benefit of SSE instrutions and stressed for more MHz. But the downside is the locked multiplier and the increased power usage. At this time I don't see any reason to choose a Katmai over either a Coppermine (1,65V, fast L2, more Mhz) or one of the multiplier limited Deschutes (less MHz, but flexible, lower power usage). Unless good memories are a reason of course.

Oh of course there are no reasons to choose it over a newer chip. But before Coppermine was more than a rumor, it was an exciting chip. It had a rather short reign thanks to AMD's relentless Athlon releases though! Coppermine was probably brought out faster than planned.

@Tetrium or anyone who has a multiplier limited Deschutes P-II running. Any news on the behaviour of the L2 cache at 133MHz?

gerwin wrote:

@Tetrium or anyone who has a multiplier limited Deschutes P-II running. Any news on the behaviour of the L2 cache at 133MHz?

Ow, I haven't thought of doing the test, I thought someone else would do it.

And I suppose the next post will be something like: "I didn't test it, I thought Tetrium was gonna do it" 😵

Will do the test but atm I'm kinda strapped on time right now.

I just couldn't be bothered: IF a game calls for 133MHz instead of 200 MHz (where L2 fully functions), then I guess it will not be of any harm to have L2 off too. It is almost a feature "auto L2off for old game mode".

What is the performance hit anyway?

To help the search:
My P2 333 with limited multiplier is
SL2QF
98300342-0472 MALAY