feipoa wrote:

noshutdown wrote:

The old k6-2 core(with no on-die cache) is already capable of 550mhz with 250nm process, but k6-3 can only reach 450mhz despite voltage officially increased from 2.2 to 2.4, hindered by its on-die cache. k6-3+ and k6-2+ are 180nm process, therefore capable of 600mhz with on-die cache, but not much higher.

Anyone have long term success running an AMD K6-III-450AHX (or AFX) at 500 MHz? What voltage was necessary to facilitate this speed and for how long was the system run?

There was a retired guy I used to know that sold systems out of his home, and he would sell overclocked K6-3 systems at 500 mhz or a tad above. I honestly cant say what voltage he used, and the success rate varied, regardless of the fact that he'd use the same line of motherboard most of the time. I honestly never even understood why he bothered with the overclocking and tweaks to begin with, because most of the people buying them were his customers looking for basic systems to browse the net or other typical non-demanding stuff.

Here is a couple of links regarding K6-3 overclocking if you care to take a look.
http://www.cpu-central.com/Articles.asp?artic … =8&decor_int=28
http://www.firingsquad.com/hardware/kryotechk63550/page4.asp

feipoa wrote:

Tetrium wrote:

The AFX is better because it produces less heat.

But it runs at the same voltage. Does the AFX use a different technology scaling (aka micron process)?

Anyone know the safe upward limit on overclocking an AMD K6-III?

The AFX runs @ 2.2v and the AHX runs at 2.4v.

I don't know but I reckon the upper limit of the K6-III being at around 500Mhz

maddmaxstar wrote:

As true as that may be, finding the best board possible for the Socket 7 interface would be the best possible way to run tests o […]
Show full quote

sliderider wrote:

But even then you're still better off building a Pentium II/III/Celeron or Athlon system because even overclocked to 600mhz the FPU in the K6 line is still too far behind to keep up in games. There are also issues with AGP in some SS7 motherboards and even if you have a SS7 motherboard designed for overclocking with bus speeds over 100mhz, there's no guarantee that faster memory will work with them, either. SS7 was a good, cheap alternative to Intel in it's day but it's hard to justify building one now with Intel parts being cheap enough that price is no longer an issue.

As true as that may be, finding the best board possible for the Socket 7 interface would be the best possible way to run tests on as many Socket 5/7 processors as possible on the same hardware platform, seeing that many Super7 could support chips as far back as P54C all the way to K6-III+ Overclocked, which would include Cyrix/IBM/ST chips, Rise and IDT as well. It would be a great board for Retro Benchmarking.

And as for how well the K6-2/III cores did in Games, they might not have been as fast as anything Intel, but they were still fast enough for the games of their day, especially when you find all the 3DNow patches.

This conversation sounds an awful lot like "486: The Next Generation". 😜

Even still, if you're going to build a system from scratch around a K6-2+/III+ overclocked to 600mhz, you're much better off just getting the parts for PIII-600 Coppermine or Slot A Athlon Thunderbird 600 instead. Having a SS7 system around for testing is one thing, but if you're going to be using it for games that require a 600mhz processor then you're better off going with a real 600mhz chip in a stable motherboard rather than an overclocked chip mounted in a glitchy one.

There are those out there who desire an equal balance of nestalgia with gaming, whereby a Slot 1/A may be too modern to satisfy the craving.

Tetrium, good point on the voltages. I suppose I'll have to test the stability at 500 MHz for myself. chipdb also indicates there being an AMD K6-III-475.

feipoa wrote:

noshutdown wrote:

The old k6-2 core(with no on-die cache) is already capable of 550mhz with 250nm process, but k6-3 can only reach 450mhz despite voltage officially increased from 2.2 to 2.4, hindered by its on-die cache. k6-3+ and k6-2+ are 180nm process, therefore capable of 600mhz with on-die cache, but not much higher.

Anyone have long term success running an AMD K6-III-450AHX (or AFX) at 500 MHz? What voltage was necessary to facilitate this speed and for how long was the system run?

Define long term. I used mine constantly at 580Mhz at default voltage. As for how long, well - a couple of months. Till I got bored and exchanged it with K7 rig 😜

I was really, really disappointed with the performance of K6/3. But I must add that the performance and stability is somewhat better on Ali then on MVP3. Also, no AGP issues 😀 And Win98 was working beautifully!

elfuego wrote:

I was really, really disappointed with the performance of K6/3. But I must add that the performance and stability is somewhat better on Ali then on MVP3. Also, no AGP issues 😀 And Win98 was working beautifully!

i also consider ali5 boards far better than mvp3, but still had trouble in windows2000.

default win2000 acpi kernel: incorrectly detects cpu frequency, games/benchmarks go out of speed control and may freeze.
force apm kernel at install: solves the problems above, but takes 10 minutes to boot after agp driver is installed.

still i want to hear anyone else running ss7 platforms in win2000?

feipoa wrote:

Anyone have long term success running an AMD K6-III-450AHX (or AFX) at 500 MHz? What voltage was necessary to facilitate this speed and for how long was the system run?

My K6-3 experiences:

K6-III 400 2.4V AHX: Wobbly at 450 MHz / 2.5V. This one got fried in an attempt to reach 500 MHz.

K6-III 450 2.2V AFX: Works fine up to 550 MHz, not stable at 600, even at 2.4V.

K6-3+ 400 1.6V ATZ: Works fine up to 550 MHz, not stable at 600, have used this one for years at 2.0V (because board doesn't go lower) and passive cooling (sawed-off Athlon heatsink).

never got any k6-3+...
i have 3 ss7 cpus:
k6-2+ 500 acz 2.0v, stable at 550, unstable at 600.
k6-2+ 550 acz, got it recently, didn't test cause the harddisk is now used on the 815e rig.
k6-3 450 ahx 2.4v, didn't expect any overclock ability.

and, whats the difference between acz and aczm?

noshutdown wrote:

and, whats the difference between acz and aczm?

Dunno, they both seem to run @ 2.0v

Edit:I can't find the page where they explain the part numbers right now. Could be they have different maximum temperatures or something?

Edit2: Nah, it's not the maximum temperature. Both are listed at 85 degrees

A pitty AMD didn't release the AMD K6-III (non-plus) using the 0.18 micron scaling. I suppose if the AMD K6-III-450 AFX works reliably at 500 MHz, that is good enough for me.

I wonder what is physically different, besides the scaling, between the plus and non-plus chips which causes an Asus P5A-B v1.05 motherboard to only work with the non-plus chips?

feipoa wrote:

A pitty AMD didn't release the AMD K6-III (non-plus) using the 0.18 micron scaling. I suppose if the AMD K6-III-450 AFX works reliably at 500 MHz, that is good enough for me.

I wonder what is physically different, besides the scaling, between the plus and non-plus chips which causes an Asus P5A-B v1.05 motherboard to only work with the non-plus chips?

The only other difference I know of is that the mobiles were specifically made for laptops and their multiplier behaves differently in a way compared to the desktop chips. I don´t know how that manifests itself though.
I was wondering the same thing though

It's been some time since I experienced the 0.18µ K6+ incompatibility on my P5A rev. 1.06, but as far as I remember this combo was so slow it was completely useless. For example, when moving around the cursor in the BIOS setup, the screen would react a couple of seconds after pressing the keys. I also remember having to flash an old BIOS version without K6+ support, or the board wouldn't boot at all!

Most times the later Aladdin V revisions -- those with the ability to cache more than 128 MB RAM -- are blamed for causing the trouble, however there are scarce reports of non-ASUS boards with late Aladdin V revisions not having this issue.

My Asus P5A-B v1.05 (ALi Aladdin V) motherboard only has 512 KB of L2 cache onboard. Some apparently have 1 MB of L2 cache. Anyway, I tested my board and it can cache up to 512 MB of SDRAM using just 512 KB of L2 cache. This is all according to CTCM7. When I put in 768 MB of RAM, CTCM7 indicates that the memory range above 512 MB is not cached.

If anyone has a VIA MVP3 motherboard, do you know, or can you check, how much RAM is cacheable for how much L2 cache your board has? I think most boards have 1 MB, but I am under the impression that 1 MB can only cache up to 128 MB for some reason.

For a retro rig, 512 MB is plenty, however, if I use half that (256 MB), will the system run faster? I figure it has to be less taxing on a system to cache less RAM and that way more of the cache can go to storing RAM memory (that is, it will increase the cache hit rate). If using 256 MB, is it faster to have two 128 MB pieces in 2 DIMMs, or one 256 MB piece in 1 DIMM?

I'm building a Cyrix MII-433GP rig so I won't have the L2 cache on the CPU; I will need to rely on motherboard cache.

feipoa wrote:

I'm building a Cyrix MII-433GP rig so I won't have the L2 cache on the CPU; I will need to rely on motherboard cache.

You have the actual 433GP part or are you overclocking a 400GP chip?

As for your memory question:I can't give a straight answer, but I reckon using fewer memory banks will usually be less taxing for the chipset.

The Ali bug (the one that can cache more then 128MB BUT can't run the mobiles) was I thought a chipset bug and not specifically a problem that has something to do with ASUS. Please correct me if I'm wrong.
What it means is that you either have an earlier version of the chipset and can run mobiles, BUT it won't cache more then 128MB RAM (unfortunately, all of my Ali boards have the early version of the chipset). Or you have a later version with an almost endless cacheable area BUT you can't run the mobiles.

I'm considering building exactly your rig as well (Cyrix MII @ 300Mhz...or maybe even 350?) but due to the cold here all my hardware tinkering has been put on hold. Hence I haven't given any more results in the SuperPi thread.

I don't really know about VIA, I never used such a board

My hopes of finding a marked Cyrix MII-433GP are fading. I've searched for this chip off and on for 10 years, but only seriously in the past 8 months. But who knows, my Cyrix 5x86-133/4x popped up right at the point I decided it was impossible to obtain.

I own two Cyrix MII-400GP's (2.2v) and one Cyrix MII-366GP (2.2v). One of them will surely be long-term stable at 300 MHz (2.3v) or 333 MHz (2.3-2.4v). I also have a Cyrix MII-366GP (2.9V) that I'll probably test to see how far this more common processor will overclock to.

I'm not sure if the ALi bug is ALi-specific or Asus-specific. I'm just glad I have the motherboard that caches 512 MB of RAM. Those AMD+ chips aren't retro enough for me. If I am itching for speed, I could always drop in the AMD K6-III-450AFX and hopefully run it at 500 MHz.

Tetrium wrote:

feipoa wrote:

The Ali bug (the one that can cache more then 128MB BUT can't run the mobiles) was I thought a chipset bug and not specifically a problem that has something to do with ASUS. Please correct me if I'm wrong.
What it means is that you either have an earlier version of the chipset and can run mobiles, BUT it won't cache more then 128MB RAM (unfortunately, all of my Ali boards have the early version of the chipset). Or you have a later version with an almost endless cacheable area BUT you can't run the mobiles.

i don't have enough info but your idea sounds reasonable. unfornately boards with new rev of ali5 chipset are really rare, all i know are asus p5a v1.05 and gigabyte 5ax rev5(i have rev4.1 which seems quite popular), so the best bet is to find someone with the 5ax rev5 board.

Mau1wurf1977 wrote:

Through BIOS updates I got a K6-2+ 550 working on two of my Super Socket mainboards.

Upon further investigation, I noticed that SpeedSys shows the graphics card as 8x AGP. It seems like it would be more useful if SpeedSys indicated what mode the graphics card was currently operating in, like 2x in this case.

I just uploaded two photos of my competing VIA MVP3 and ALi Aladdin V motherboards. I noticed that the TAG RAM on this HOT board is only 8 ns (-8). The comparison PDF I uploaded in the initial post seems to indicate that this mediocre cache speed (of 8 or 10 ns) may be problematic at 100 MHz.

Mau1wurf1977, may I know what the TAG RAM speed is on your super7 motherboards?

Anyone who has had issues with VIA MVP3 boards, do you happen to know what the TAG RAM speed is? For those who haven't had issues with VIA MVP3 boards, do you know what the TAG RAM speed is?

Has anyone benchmarked their Cyrix MII's on the Asus P5A-B v1.05 motherboards (ALi)? The results are increadibly sluggish compared to the HOT-591P (VIA) using a Cyrix MII 100x3 (300 MHz).

Landmark, for example, returns an ALU/FPU score of 802/2268 for the ALi board, while the VIA board gets 3628/4076 for the same CPU. 3DBench2 also showed a reduced score of 254 for the ALi board and 390 for the VIA board. Any idea what's going on?

feipoa wrote:

Has anyone benchmarked their Cyrix MII's on the Asus P5A-B v1.05 motherboards (ALi)? The results are increadibly sluggish compared to the HOT-591P (VIA) using a Cyrix MII 100x3 (300 MHz).

Landmark, for example, returns an ALU/FPU score of 802/2268 for the ALi board, while the VIA board gets 3628/4076 for the same CPU. 3DBench2 also showed a reduced score of 254 for the ALi board and 390 for the VIA board. Any idea what's going on?

I only tested my 300Mhz part (GP400 overclocked) on an Ali board. Would be interesting if this chip performs better using a VIA chipset, but strange

Did both BIOS's recognise the chip correctly? Could it be something to do with the Liniar Burst thingy?

I will compare the register settings of both the VIA and ALi boards. So far the ALi board's BIOS does enable Linear Burst. I'll see what the VIA board does. The VIA board gets a Speedsys score of 215, whereas the ALi board shows 128. BYTEMark also shows about half speed for the ALi board.

I've read this particular Asus board runs painfully slow with an AMD K6-2/3+ processor, so I wonder if some of this sickness has spread over to the Cyrix MII chips? Unfortunately I do not have another ALi board to test this theory out on. I'd rather not use VIA boards due to the small cacheable RAM area compared to ALi boards. [Note, if your BIOS does not have an option to disable CPUID, you must disable CPUID in software to run SpeedSys with an MII].

I should probably also compare a K6-2-500 in the VIA and ALi boards to ensure the benchmarks are similar. This will help narrow the problem down to an MII-only issue.

EDIT: Yes, both BIOS' identified the CPU as Cyrix MII or MX - 433.