Gigabyte GA-6OXET-C vs. Abit ST6

[DSTA]

Grmbl. Take a look at this, and note the date of publication. Also might explain why not only the GBT board is affected.

Intel ® 815E Chipset Platform for Use with Universal Socket 370 Design Guide Update // Doc. # 29859401 // October 2001 (initial release)

"[...] A platform supporting the 0.13 micron technology processor must implement a workaround
required for the A-1 stepping of that processor, identified by CPUID = 6B1h.
The internal control register bit responsible for operation of the THERMTRIP circuit
functionality may power up in an un-initialized state. As a result, THERMTRIP# may be
incorrectly asserted during de-assertion of RESET# at nominal operating temperatures. When
THERMTRIP# is asserted as a result of this, the processor may shut down internally and stop
execution. In addition, when the THERMTRIP# pin is asserted the processor may incorrectly
continue to execute, leading to intermittent system power-on boot failures. The occurrence and
repeatability of failures is system dependent, however all systems and processors are
susceptible to failure. [...]"

 

[oldfart]

tainice, I think the GA6OXET is not a good board if you are overclocking a CuMine. Here is something I see on mine. If I use 1.80 Vcore (stock is 1.75), it will always hang on a reset of any kind. Here is the weird part: It will do that even with NO overclock! Try it. Run your 800EB @ 800. Set the Vcore for 1.8 and see what happens.

 

[tainice]

yeah, i did get the same kind of weird reboot problem! with default and 1.8 volt.. in any case, i rma my board for a new one, hoping to get a better one. man, i hate this, there is no other board around posses 1/5 dividers other than XET/C!! anyway, 800eb must go, and have to get a celeron 1.2 as backup. cuppermin simply won't cut it!

 

[oldfart]

I dont think a new board will help. Good luck with the 1/5 PCI also. I cant get it to work. Crash City when I try 160 FSB PLL/16 30/60 PCI/AGP. The C1.2 should work much better.

 

[tainice]

man, are u telling me that those fancy PLL can acutally do nothing besides crashing? dude, that sucks! I just received my 1.26, it delieved right on schedule, but i have no board to play with it.....last week, i got the board, but no chip, and this week situation just reversed.....Anyway, so PLL does not work right? then i will exchange a st6 then. GB....piece of .

 

[oldfart]

Well, I can tell you it doesn't work on mine. I've tried everything possible. F9, F9D BIOS, different video cards. With my PIII 1000EB, I could run 160 FSB 40/80 PCI no problem. Change to 160 FSB PLL/16 30/60 and it will crash every time.

I run a 1.26 in mine and it works fine as long as you dont use the PLL feature.

 

[Flu!d]

The extra dividers only seem to work on the XET....Not the XET-C.
Tainice....Chances are your 1.26gig chip won't hit 160mhz FSB anyway and the board is still very overclockable....More overclockable than many other boards out there. Why get rid of it just because the extra dividers don't appear to work? None of the other boards have the extra dividers?! And the XET appears to be the king in terms of memory bandwidth....

Relax dude....

Quote by DSTA....


"[...] A platform supporting the 0.13 micron technology processor must implement a workaround
required for the A-1 stepping of that processor, identified by CPUID = 6B1h.
The internal control register bit responsible for operation of the THERMTRIP circuit
functionality may power up in an un-initialized state. As a result, THERMTRIP# may be
incorrectly asserted during de-assertion of RESET# at nominal operating temperatures. When
THERMTRIP# is asserted as a result of this, the processor may shut down internally and stop
execution. In addition, when the THERMTRIP# pin is asserted the processor may incorrectly
continue to execute, leading to intermittent system power-on boot failures. The occurrence and
repeatability of failures is system dependent, however all systems and processors are
susceptible to failure. [...]"

This is very handy stuff DSTA and just goes to prove that the fault 'is' in fact due to the 815EP chipset in general and no fault of the mobo....I'm gonna keep that quote....

 

[dscline]

The extra dividers only seem to work on the XET....Not the XET-C.

They seem to work on mine.

 

[tainice]

They seem to work on mine.

Actually, they also worked on my board b4, but crashing everytime, not sure what'z the cause. Anyway, i will replace it with a new XET-C....why? i still like the expectional memory performance of the board, and those PLL functions... yeah, i start another new post to discusee which board, XET/C, ST6, and Tsul2-c, is the fastest..go there and post!

 

[OC2K]

There is no doubt in my mind that the dividers on the GA-60XET-C work...that is how I got up to 1608mhz.....

 

[oldfart]

There is no doubt in my mind that the dividers on the GA-60XET-C work...that is how I got up to 1608mhz

Please explain. 1608 is just 12 x 134. Wouldn't you be using a standard 1/4 PCI like any other board has or am I missing something?

 

[dscline]

Well actually, since we are using 100Mhz FSB chips, it wants to run them at 1/3. I have to use the pll dividers to get to 1/4.

 

[tainice]

guys, i was using F9, what are u guys using? especially for those luck guys who pull their PLL off, please tell what'z ur bios verision, so i know i am getting the best out of this board.

 

[oldfart]

Well actually, since we are using 100Mhz FSB chips, it wants to run them at 1/3. I have to use the pll dividers to get to 1/4

C1.2 @ 1.6 is a nice oveclock. However, the PLL had nothing to do with it. Any board will have a 1/4 PCI @ 133+ FSB. The unique featue is the 1/5 and 1/6 PCI dividers that a PLL allows. This is the one I can not get to work. 160 FSB, PLL/16 30/60 (1/5 PCI) = crash city.

 

[dscline]

That's the strange thing about it.... I don't really understand exactly how the pll thing works, but with a chip with a 100Mhz defualt FSB, pll16 = 1/4 multiplier.

 

[oldfart]

Different board makes have different ways of doing the dividers.

ASUS TUSL2: Below 133 is automatically 1/3, 133 and above is 1/4.

ABIT ST6: 1/4 and 1/3 are manually selectable.

Giga GA6OXET uses PLL values to select 1/3, 1/4, 1/5, 1/6 dividers.

All of them can do 1/4 @ 133+ FSB. Just a different way to select the option.

 

[Celemine1Gig]

C1.2 @ 1.6 is a nice oveclock. However, the PLL had nothing to do with it. Any board will have a 1/4 PCI @ 133+ FSB. The unique featue is the 1/5 and 1/6 PCI dividers that a PLL allows. This is the one I can not get to work. 160 FSB, PLL/16 30/60 (1/5 PCI) = crash city.

It's your RAM oldfart!!! Use the divider at a FSB speed of for example 140 or 150 MHz and it'll work!! I explained it in another thread that the dividers stress your RAM!

 

[dscline]

Hmm, I wonder why your results show that simply enabling the pll settings stresses the ram more? Are there any vid-pin tricks to make my system think my p3 700 is supposed to run on 133? I know how to tape a pin a slot1 to get 66 -> 100, but no idea what I need to get 100 -> 133 on a socket. If your results hold true, then I should be able to get mine faster (like it was on my BE6-2), since the 1/4 multiplier would be automatic, and I wouldn't have to use the pll multipliers to get there.

 

[TASOS]

For tainice

In my little search ...(looking for the "golden m\b")...i found out
that the new series of Soltek both 694T & 815B chipsets have
1/5 dividers...and they also have the "Red Storm" utility for o\c.

So XET is not the only one arround !!!

 

[oldfart]

Celemine1Gig, thanks for the reply. First, I wont buy new ram. I have 512 meg Crucial PC133 CL2 (2 x 256). Next ram I buy will be DDR or RDRAM. This ram will run 150+ FSB 3-2-2 5-7, so its not all that bad. I can run it 160 FSB 3-3-3 perfectly as long as I dont use the PLL. Why is my ram good enough 160 FSB, but not good enough 160 FSB PLL? I saw in you other thread you say it is stressed more. Could you elaborate on that?

Thanks again!