Help me decide between K6-III+ 400Mhz or K6-II+ 550Mhz

Guest · Post by **Guest** » Thu Sep 22, 2005 4:05 pm

Here are the specs on the K6-III 400Mhz...

K6 III+ / 400 ACR
2V CORE / 3.3V I/O
A 0051APAW

Don't know about the K6-II+ specs at this time.

It's going on a FIC PAG-2130 SS7 board w. 512k onboard.

KachiWachi · Post by **KachiWachi** » Thu Sep 22, 2005 4:17 pm

A III+ will outperform a 2+ because of the larger on-die CPU L2 cache (256K vs. 128K). Otherwise, they are the same. Both will require about 10A on VCore...so make sure your board can support that.

Guest · Post by **Guest** » Thu Sep 22, 2005 4:21 pm

How do I find out about the amp, Kachi?

KachiWachi · Post by **KachiWachi** » Thu Sep 22, 2005 4:48 pm

FIC PAG-2130 -> http://www.fic.com.tw/product/motherboa ... del_id=104

Since the board already supports a K6-III, you should have no trouble running a + series CPU in there.

Guest · Post by **Guest** » Thu Sep 22, 2005 6:32 pm

"A III+ will outperform a 2+ because of the larger on-die CPU L2 cache"

Kachi, even at 150Mhz difference? Or are you taking overclocking into consideration?

KachiWachi · Post by **KachiWachi** » Thu Sep 22, 2005 8:06 pm

You will most likely be able to run them both at the same speed.

Are these new buys or...??

Jim · Post by **Jim** » Thu Sep 22, 2005 11:50 pm

The K6-2+550 is pretty well a guaranteed bet to do 600MHz+. The K6-3+ 400 ACR is somewhat less certain to do 600MHz+, but in all likelyhood will do at least 550 MHz. As to which will perform best, that is hard to say. Generally I would give the nod to a K6-3+; though that particular kind of K6-3+ is in my opinion is near, or at, the bottom of the line. As far as performance comparison results between them are concerned, all I have to offer is to be found in the topic "MVP vs ALi chipset 3 way shootout." which has test results for both K6-2+ and K6-3+ processors, though they were on different types of boards running different amounts of ram.

sylvan2626 · Post by **sylvan2626** » Fri Sep 23, 2005 7:22 am

If I am remembering a post in another BBS from Jan (BIOSMAN), his comment was that I would need ~100MHz more in a K6-2+ CPU speed to match a K6-III+ CPU's performance because of the CPU cache size difference. (CRAP!! That sounds confusing even to me - and I wrote it! What I mean is that a K6-2+ 600MHz CPU would about equal a K6-III+ 500MHz CPU, all other things being equal - like chipset, mobo, RAM, etc.).

I have not ever tested that out - but his reputation was enough to convince me.

The K6-2+ 550MHz CPU may not get correctly recognized by the mobo BIOS. I know that my EPoX boards' BIOSes will officially support up to 500MHz CPUs, and I have not yet tried one of my K6-2+ 550MHz CPUs in it. EPoX tech support stuck to their party line about only 500MHz CPU support. I half-expect the POST process to stop when the PC tries to figure out just what CPU is in it. If that happens, then I would have to jumper down the CPU for POST.

The 400ATZ is a nice little CPU; the one I am currently playing around with will run stably from a core voltage of 1.3V to 2.4V - although I personally don't like the 2.4v setting. Should be comparable to your 400ACR. Only been able to get to 575MHz with this CPU, but if Jan is right it is still better performance than I can expect from my K6-2+ CPU.

Between those two CPUs, I would go with the 400ACR.

Stedman5040 · Post by **Stedman5040** » Fri Sep 23, 2005 8:48 am

I have a K6-2+/550 that certainly runs at 600MHz and a colleague of mine runs a K6-2+/550 on a PAG2130 board with the appropriate BIOS. From what I remember though is that this board does not support 2.0V so you have to run the CPU at 2.1V. I think a read a review that gave the performance difference between a 2+ and a III+ as being about 50MHz. IE a K6-2+/550 is about the same as a K6-III+/500. This of course was according to benchmarking and not in real life performance terms. I think it may have been an article on Lost circuits,Tomshardware or Anandtech website.

As an aside I have just gotten hold of a GA-5SMM (v1.3) MB. I can post my K6-2+/550 at 133fsb with a 4.5X multiplier to 600MHz. Going to further test into Windows etc with regards to stability. The CPU runs LINUX @ 450MHz fsb100 at 1.6V no problems. Again I am going to test this board out at 133MHz to the full. Quite useful that this board has a 4X divider.

Guest · Post by **Guest** » Fri Sep 23, 2005 10:06 am

Two things:

The PAG-2130 has a minimum voltage of 2.0, at least. Hidden settings may have something else, but not sure.

Secondly, is there any way to add a 4x divider to a board? Most of them ,of course, only have a 3x.

Guest · Post by **Guest** » Fri Sep 23, 2005 10:18 am

Stedman,

What do you think of the perfomance of your GA-5SMM compared to a MVP3 chipset board? At the same FSB, do you think one is better, or about the same?

Just wondering if 133Mhz FSB is really a big deal or not, because I noticed that on some sites, and on a Sandra benchmark the SiS 530 chipset was abominable. It made me think that maybe the Sis 530 chipset wouldn't be any or much faster than a good MVP3 board at 100Mhz FSB.

Geekasaurus · Post by **Geekasaurus** » Fri Sep 23, 2005 11:06 am

The big thing that sucks about the SiS530 boards is the small L2 cache on most boards, 512K usually, which will only cache 64MB of system memory. You start adding additional memory beyond 64MB and it gets slooow. This is a major reason why the MVP3 rules at 100FSB, 1MB being the most popular amount of L2 normally found which will cache 256MB of system RAM.
I run a K6-3+ on an Asus P5S-VM/SiS530 at the 133 FSB,600MHZ with the boards L2 disabled and the performance is roughly equal to an MVP3 board at the 100FSB with its L2 cache enabled.
Nice features for SiS530 boards are ATA66 and most boards will handle up to 768MB of RAM.

sylvan2626 · Post by **sylvan2626** » Fri Sep 23, 2005 1:58 pm

My first SS7 mobo was based on the SiS530 chipset. I really liked the PnP CPU BIOS, but hated the lack of FSB higher than 100MHz and the memory bandwidth. The 450ACZ K6-III+ ran 600MHz on that board and I think it could have gone higher if there had been a higher selection of FSB.

EPoX boards are good boards.
The EPoX MVP3 group of boards will also accept up to a max of 768MB of RAM. The MVP3G2 that I am running with a 400ATZ CPU is operating at a FSB of 112MHz with the board's L2 cache enabled, although I am currently only using 1 module of 256MB PC133 RAM. Memtest86+ ran with no errors for multiple passes.

Stedman5040 · Post by **Stedman5040** » Mon Sep 26, 2005 7:49 am

With regards to the GA-5SMM it manages to load win98se at the following settings with a K6-2+/550ACZ

450MHz@1.3V/100fsb
550MHz@1.6V/100fsb
533MHz@1.6V/133fsb
600MHz@1.8V/100 and 133fsb

Tried 667MHz@2.0V/133fsb. It posted but wouldn't load into Windows. CPU-z and Sandra report 1.9V for all voltage settings so they are obviously picking up the wrong info from the sensors. It's a little unfortunate that this board does not report any temperatures. I may try and use 2.1V and a bigger cooler.

Does anyone know if there are any hidden frequency settings for this board. The manual shows settings for 66,75,83,90,95,97,100,112,124,and 133. There are six unused settings as far as I can make out. Gigabyte website shows 105 as a setting on the specs but does not show it in the manual.

KachiWachi · Post by **KachiWachi** » Mon Sep 26, 2005 10:58 am

What clock chip is on the board?

Does EVEREST post incorrectly as well?