I may pull the 2 sticks out then, but haven't notice any increase or decrease in performance at all.

Why do they make board that cant cache there full RAM capacity?

Does anyone know where I might find a bios upgrade for this board? I wonder what a new bios might support?

Oldskoolmaniac wrote:

I may pull the 2 sticks out then, but haven't notice any increase or decrease in performance at all.

Why do they make board that cant cache there full RAM capacity?

Does anyone know where I might find a bios upgrade for this board? I wonder what a new bios might support?

If removing half the RAM doesn't have any disadvantage, then you might as well save that memory for another build 😀

They probably done it like this to keep manufacturing costs down while still being able to sell a board that supports this humongous amount of RAM! 😁 (but it sure can't cache all of it, but we'll only mention this in the fine print shht).

I'd forgotten, but the board (the 850 or FR500) is actually made by BCM, model number IN5598 and this board actually does have a modded BIOS. It won't help you with the cacheable area just by flashing it, but the flashing can enable your board to be upgraded with one of the AMD ss7 mobiles (K6-2+ or K6-III+) which will make the cacheable area problem go away.

The modded BIOS file is on this page http://web.inter.nl.net/hcc/J.Steunebrink/k6plus.htm
Search for "IN5598" and you should find it right away.

It's probably a good idea to first save the old BIOS file before flashing it with this one, just in case something goes wrong for whatever reason.

Tetrium you are awsome, thanks for all your help on this I got a ton of k6 processors laying around and this bios will help with supporting my k6-2 500mhz and a 128GB HDD.

Question.... what is the tool for flashing this bios? Ive flashed a ton of different bios through DOS, but they've all had their own tool packaged with them.

**Edit** I used the Award flasher for my bios and it keep getting errors saying boot block mismatch so I downloaded uniflasher and that was a lot easier to use. I created a backup first then tried installing the beta bios and again boot block mismatch then i restored my backup, but maybe this bios wont work for a packard bell board.

It wont even flash bios 1.06 witch is from packard bell for this board.

Oldskoolmaniac wrote:

Tetrium you are awsome, thanks for all your help on this I got a ton of k6 processors laying around and this bios will help with […]
Show full quote

Tetrium you are awsome, thanks for all your help on this I got a ton of k6 processors laying around and this bios will help with supporting my k6-2 500mhz and a 128GB HDD.

Question.... what is the tool for flashing this bios? Ive flashed a ton of different bios through DOS, but they've all had their own tool packaged with them.

**Edit** I used the Award flasher for my bios and it keep getting errors saying boot block mismatch so I downloaded uniflasher and that was a lot easier to use. I created a backup first then tried installing the beta bios and again boot block mismatch then i restored my backup, but maybe this bios wont work for a packard bell board.

It wont even flash bios 1.06 witch is from packard bell for this board.

I think I should've mentioned that, even though we apparently have the same boards and even though I actually used this board for a build in the past, I never actually flashed the board I have as I always was worried about it ruining the board and it worked fine with what I intended to do with it.
I did do my homework though 😜

Frankly, I don't know for sure which flasher will work on this board and finding definitive answers with hardware this old and parts that weren't particularly popular, can be quite a task in itself, often it ends up having to do some (brave) detectivework by yourself (which is part of the fun 😁)

I did find this page http://www.bcmcom.com/tech/in5598/IN5598.htm and it contains the official BIOS update for the IN5598 board.
Both are compressed archives and both contain an AWARD flasher for the BIOS file, perhaps this one will work?
But keep in mind that one should never assume something will work, so I'd advice to keep staying a bit cautious.

The page also includes jumper settings and such, but keep in mind that a board made by a manufacturer for a large OEM like PM may have slight, but significant differences when compared to the retail board, having a different BIOS is often one of them.

And about the CPU voltages, iirc this board actually supports CPU voltage in increments of 0.1v and iirc all the jumpers did, was basically adding either 0.8v, 0.4v, 0.2v or 0.1v, something along those lines.

I'll edit this reply once I figured them out again.

And yw 😀

edit: Apparently what I said above about the CPU voltages only holds true for the largest part.
JP8 seems to add 0.1v
JP9 seems to add 0.2v
JP10 seems to add 0.4v
JP11 seems to add 1.0v and not 0.8v as I suggested was the case above.

Here are the official and guessed CPU voltages of this board, 0 means no jumper and 1 means the jumper is installed

PB FR500/850 official

­JP - 8 9 10 11
1.8 - 0 0 0 0
2.2 - 0 0 1 0
2.8 - 0 0 0 1
2.9 - 1 0 0 1
3.2 - 0 0 1 1
3.3 - 1 0 1 1
3.5 - 1 1 1 1

PB FR500/850 including educated guesses

JP - 8 9 10 11
V - 0.1 0.2 0.4 1.0

1.8 - 0 0 0 0
1.9 - 1 0 0 0
2.0 - 0 1 0 0
2.1 - 1 1 0 0
2.2 - 0 0 1 0
2.3 - 1 0 1 0
2.4 - 0 1 1 0
2.5 - 1 1 1 0
2.8 - 0 0 0 1
2.9 - 1 0 0 1
3.0 - 0 1 0 1
3.1 - 1 1 0 1
3.2 - 0 0 1 1
3.3 - 1 0 1 1
3.4 - 0 1 1 1
3.5 - 1 1 1 1

I tried to make some kind of table that's somewhat easy to read, but it didn't come out as I had intended 😵

The possible lack of CPU voltages of 2.6v and 2.7v is of no concern, as s7 chips are rarely ran on this voltage (usually run on >3.3v, 2.8v/2.9v and then a big gap to 2.2v/2.4v and the bottom for this board being 1.8v (virtually any of the K6+'s that have an official 1.6v CPU voltage can run perfectly fine on 1.8v).

Oldskoolmaniac wrote:

I may pull the 2 sticks out then, but haven't notice any increase or decrease in performance at all.

Well, if you're lucky, you'll only notice decrease in performance for the part of memory that is uncached... Which hopefully is the last memory the machine will use. So as long as you don't go over 128 MB of memory usage, the performance is the same (you could argue that once you go over 128 MB, it's better to have uncached memory than not to have memory at all, because it will induce disk swapping... however, modern OSes and various applications are designed to make use of whatever memory is free in your system, so in some edge cases it could be that having more than 128 MB causes your system to use more memory, and therefore pushing applications into the uncached area, where they would otherwise be cached).
If you're unlucky (I have a Pentium II with 440BX chipset like this), the cache drops out for all memory, and the system slows down to a crawl when you put in too much memory.

Oldskoolmaniac wrote:

Why do they make board that cant cache there full RAM capacity?

It's not the board, it's the chipset. Its cache controller is simply not designed to cache more than 128 MB.
The reason is probably because designing a cache for a larger array of memory makes it more complex, and would lower overall performance.
So they probably figured it's better to design it for best performance up to 128 MB, because that's the most common use-case.
Remember, SIS is a budget brand. They wanted to keep cost down.

Oldskoolmaniac wrote:

Does anyone know where I might find a bios upgrade for this board? I wonder what a new bios might support?

The BIOS can't fix this, it's a physical limitation in the chipset.

SIS boards are starting to really put a bad taste in my mouth.

So got my 333mhz k6-2 put in jumper are set for 66.6mhz fsb 2.2v got it to boot into windows after the third time and now it goes back to constant loop of rebooting before booting to windows. If I try any other jumper for the volts I get blank screen and no post, Ive try just about everything.

Scali wrote:

Well, if you're lucky, you'll only notice decrease in performance for the part of memory that is uncached... Which hopefully is the last memory the machine will use. So as long as you don't go over 128 MB of memory usage, the performance is the same

This is incorrect because Windows uses memory from the top down, so it will start by using memory from the uncached area, so unless you do load a lot of programs at once all the memory usage will go to the slow part. DOS only uses the first 64MB anyway, and in that case only the cached area will be used, so it will never use the uncached portion.

That said, with Pentiums the difference in performance is less than 10% so you won't be able to tell with the naked eye. Even with K6s where there is a greater impact you probably also won't be able to tell the difference without running some benchmarks.

Scali wrote:

If you're unlucky (I have a Pentium II with 440BX chipset like this), the cache drops out for all memory, and the system slows down to a crawl when you put in too much memory.

Also not very accurate. The 440BX doesn't handle the cache at all. It's all controlled by the CPU and only some Pentium II CPUs are limited to 512MB max cacheable area. This limit doesn't exist on all 350MHz and faster models, so if you upgrade to one of those you can install as much RAM as the board allows and won't lose any performance.

So if anyone has a bios for the PB850 that would be great even a dump of you're bios would work.

Oldskoolmaniac wrote:

SIS boards are starting to really put a bad taste in my mouth.

So got my 333mhz k6-2 put in jumper are set for 66.6mhz fsb 2.2v got it to boot into windows after the third time and now it goes back to constant loop of rebooting before booting to windows. If I try any other jumper for the volts I get blank screen and no post, Ive try just about everything.

So what happened?

Constant rebooting sounds like a more serious error, perhaps it's a good idea to start basic troubleshooting.

I know this thread is a bit old, perhaps you already binned this project?

OP, I happen to still have files for this board. Attached is the (patched, I think?) v1.06 BIOS. Hope it helps.

RetroBoogie wrote:

BCM FR500.zip

OP, I happen to still have files for this board. Attached is the (patched, I think?) v1.06 BIOS. Hope it helps.

Cheers, I downloaded it also since I also have this board 😀.

I did part it out and put it away in bins, but now i really want to get it out and try that bios update. ill get around to it sometime.

devius wrote:

This is incorrect because Windows uses memory from the top down, so it will start by using memory from the uncached area, so unless you do load a lot of programs at once all the memory usage will go to the slow part. DOS only uses the first 64MB anyway, and in that case only the cached area will be used, so it will never use the uncached portion.

How is that incorrect when it depends on which OS you use?

devius wrote:

Also not very accurate. The 440BX doesn't handle the cache at all. It's all controlled by the CPU and only some Pentium II CPUs are limited to 512MB max cacheable area. This limit doesn't exist on all 350MHz and faster models, so if you upgrade to one of those you can install as much RAM as the board allows and won't lose any performance.

I think you're not being very accurate yourself.
My Pentium II system already runs into trouble when you put 512 MB in there, not *more than* 512 MB (I downgraded it to 320 MB to get performance back up).
And that is caused by the chipset, not the CPU, because as you say, even worst-case the Pentium II should handle 512MB of cacheable area (and I'd have to check, but I think I have a later model PII, so it wouldn't even have the 512MB limit in the first place). It probably has to do with how the chipset maps things like the AGP aperture and such.

Scali wrote:

How is that incorrect when it depends on which OS you use? […]
Show full quote

devius wrote:

This is incorrect because Windows uses memory from the top down, so it will start by using memory from the uncached area, so unless you do load a lot of programs at once all the memory usage will go to the slow part. DOS only uses the first 64MB anyway, and in that case only the cached area will be used, so it will never use the uncached portion.

How is that incorrect when it depends on which OS you use?

devius wrote:

Also not very accurate. The 440BX doesn't handle the cache at all. It's all controlled by the CPU and only some Pentium II CPUs are limited to 512MB max cacheable area. This limit doesn't exist on all 350MHz and faster models, so if you upgrade to one of those you can install as much RAM as the board allows and won't lose any performance.

I think you're not being very accurate yourself.
My Pentium II system already runs into trouble when you put 512 MB in there, not *more than* 512 MB (I downgraded it to 320 MB to get performance back up).
And that is caused by the chipset, not the CPU, because as you say, even worst-case the Pentium II should handle 512MB of cacheable area (and I'd have to check, but I think I have a later model PII, so it wouldn't even have the 512MB limit in the first place). It probably has to do with how the chipset maps things like the AGP aperture and such.

You really have a hard time admitting you are wrong from time to time. Things like this make someone's character look much more ugly, just a friendly hint here.

You were wrong for several reasons, first of all simply because the advice you gave was inaccurate. First of all Oldskoolmaniac has some pics of his OS in his very first message here in this thread, so luck about yes or no cacheable area has nothing to do with it as it's pretty obvious what OS he intended to use. And AGP aperture size had nothing to do with why Devius mentioned there wouldn't be a way to slow the system down using faster Pentium 2 CPUs, you're only bringing this up to cover up your own flaws.

And telling someone he is wrong based on nothing more then merely your own anecdotal evidence is not just not very accurate. It's actually very poor. especially considering it was easy for you to avoid making a point in trying to desperately be the one who was correct at the cost of someone else.

Like I said, it doesn't make you look very good and you tend to do this quite a lot (even in private, except you were worse).

What's the most poor is that these errors of yours were actually quite minor, why try to cover those up that way in the first place? Not even worth worth losing face for, right?

If I am wrong about something (it happens to anyone), then I think I should simply have known better and showing that the value of someone's opinion depends on the value of the opinion itself instead of the value of the person stating said opinion is really the more civilized way and the better way forward. It's what true social skills are all about 😀

Have a nice day 😀

Tetrium wrote:

You really have a hard time admitting you are wrong from time to time.

I was not wrong. I did not provide enough details for some people's liking. Doesn't make the information I did provide wrong, or even inaccurate.

Tetrium wrote:

You were wrong for several reasons, first of all simply because the advice you gave was inaccurate. First of all Oldskoolmaniac has some pics of his OS in his very first message here in this thread, so luck about yes or no cacheable area has nothing to do with it as it's pretty obvious what OS he intended to use.

So? I was making a general observation, not aimed at Oldskoolmaniac specifically or what OSes and software he would intend to use or not.
Of course, if you have some kind of agenda, I can see how you want to try to shoehorn my words this way.

Tetrium wrote:

And AGP aperture size had nothing to do with why Devius mentioned there wouldn't be a way to slow the system down using faster Pentium 2 CPUs, you're only bringing this up to cover up your own flaws.

Devius responded to me, and he wrongly assumed that my problem was because I had *more* than 512 MB in the Pentium II system. Which I did not.
I am not covering up anything. I mentioned the 440BX chipset specifically before Devius made his comment. Anyone is free to put together a 440BX system with a Pentium II like I have, and conclude that the problem occurs with 512 MB too (probably with somewhat less, but 320 MB was the next-best configuration I could make... using 1x256 and 1x64 DIMMs instead of 2x256, which caused the slowdowns. I wouldn't be surprised if 1x256 and 1x128 would also be a problem with a 256MB AGP aperture or such).

Tetrium wrote:

And telling someone he is wrong based on nothing more then merely your own anecdotal evidence is not just not very accurate.

You mean that's what Devius and you are doing.
I merely pointed out Devius' inaccuracy in the fact that the slowdown does not necessarily occur with *more than* 512 MB of physical memory in a Pentium II system (which is what people might think if they just read it on the web, and don't have any hands-on experience). That's not anecdotal evidence, that is a fact.

Not sure why you feel like you have to drag this out like this again, with your personal attacks.

Because this thread got reported I'd like to say something.

When discussing a topic do keep in mind the context set by the original poster. Taking something out of context can, like in this case, lead to unnecessary confusion and arguments.

PhilsComputerLab wrote:

When discussing a topic do keep in mind the context set by the original poster. Taking something out of context can, like in this case, lead to unnecessary confusion and arguments.

If you are talking about 'DOS vs Windows'... that's not what this was about. If you don't understand, I suggest you talk to the parties involved.

Please, everyone just chill out so Phil doesn't have to break out the threadlocker and so OldSchoolManiac can fill us in on the latest news of his build. If someone wants to discuss the shortcomings of Klamath core Pentium IIs, perhaps it should be in another thread.

Scali wrote:

PhilsComputerLab wrote:

When discussing a topic do keep in mind the context set by the original poster. Taking something out of context can, like in this case, lead to unnecessary confusion and arguments.

If you are talking about 'DOS vs Windows'... that's not what this was about. If you don't understand, I suggest you talk to the parties involved.

If I was talking about DOS vs Windows, I would say so. I'm giving general advice to help minimise unnecessary arguments and drama going forward.

PhilsComputerLab wrote:

If I was talking about DOS vs Windows, I would say so. I'm giving general advice to help minimise unnecessary arguments and drama going forward.

I am just saying that the argument isn't about this thread. It's a personal thing, and it's been going on for a long time. Tetrium just picks out random posts in random threads to attack me personally. Hence your advice will not fix this problem. He will find other posts in other threads to pull out of context and start new unneccessary arguments and drama. That's just how he is.