430TX Chipset RAM Limitation
Mitch4949
the Intel website to the effect that this chipset is not capable of paging
more than 64MB of RAM, and if you install more than that amount, you will
experience a "performance hit" when the overage is reached. This is very
disturbing, because my PC vendor stated that my PC would handle up to 256MB.
Of course, he must've been referring to the motherboard, but wasn't that a
little dishonest? Is what I read at Intel the truth, and is there anything I
can do about it, short of getting a new CPU?
Mitch
Trif
yes the TX can only cache 64 meg. It can address more.
I have seen many arguments about this, but it was my observation that
in real world performance, the benefit from going beyond 64 meg far
outweighed, the slight, 1-2 percent hit.
In ohter words, say you are running Windows? It will be much happier
with 128 meg of ram, and perform better than if you stuck with 64.
My opinion...
Trent Worthington
wrote:
>I have the Intel Triton 430TX rev.1 chipset. I recently found a statement on
>the Intel website to the effect that this chipset is not capable of paging
>more than 64MB of RAM,
ITYM caching. Big difference.
> and if you install more than that amount, you will
>experience a "performance hit" when the overage is reached.
True.
>This is very
>disturbing, because my PC vendor stated that my PC would handle up to 256MB.
>Of course, he must've been referring to the motherboard, but wasn't that a
>little dishonest? Is what I read at Intel the truth, and is there anything I
>can do about it, short of getting a new CPU?
You can still use 256MB, it just won't be cacheable. The motherboard
is the limitation. Consider an upgrade for 150 bucks or so to a
Celeron.
>
>Mitch
--
xaodox <xao...@freeuk.co.uk> Slurps me big time in
Message-ID: <T1Ff5.78$_8.8...@nnrp3.clara.net>
Trent Worthington, is a God a warrior and a gent,
we are not fit to suck his feet or dine on his excrement.
At night we crune by the light of the moon of his courage and his creed,
long live the Worthington seed.
Trent Trent, top of the hero class,
Trent, Trent we live to kiss his ass.
Martin Atkinson-Barr
Trent Worthington wrote:
> On Thu, 03 Aug 2000 16:35:35 GMT, "Mitch4949" <mkafte...@att.net>
> wrote:
>
> >I have the Intel Triton 430TX rev.1 chipset. I recently found a statement on
> >the Intel website to the effect that this chipset is not capable of paging
> >more than 64MB of RAM,
>
> ITYM caching. Big difference.
>
> > and if you install more than that amount, you will
> >experience a "performance hit" when the overage is reached.
>
> True.
>
Install a K6-III, with 256Kb on-chip cache that will cache ALL of the memory. I
too have a TX chipset and I use 128MB memory in this way. In fact it may be a
great solution as the motherboard cache of the lower 64MB corresponds to the
Windows location of device drivers. The lower 64 gets both on-die cache and
motherboard cache, the upper memory (64MB in my case) is served by the K6-III
cache.
I saw in Frys that K6-III 333MHx were offered at $29.95. I use a 450MHz (cost you
about $150). Far superior to the Celeron, smoother response and more reliable
IMHO. There's also the K6-2+ and the K6-III+ if you can find them, they are new.
Make sure you get a 2.2v version of the K6-III. Some people will say the current
draw on the K6-III is large, it used to be but not anymore.
Trent Worthington
The question is, would an AMD part work on his board. He did not state
the manufacturer of the motherboard, so they may not work.
Martin Atkinson-Barr
Trent Worthington wrote:
>
>
> The question is, would an AMD part work on his board. He did not state
> the manufacturer of the motherboard, so they may not work.
> --
>
Well I've never found a Socket 7 motherboard that didn't work with a K6 series - I've
used about 8 K6-IIIs in various motherboards and a lot more K6-2s. Sometimes you need
a BIOS update (remember its the BIOS writer that does CPU detection, not the
motherboard mftr i.e Award, AMI etc).
If anyone finds a socket7 that cannot make use of a K6 -2 or K6-III lets hear from
them. I think AMD also has a list of "approved motherboards" on their website -
however omission from the list does not mean the K6 won't work.
Best regards
Trent Worthington
Ahem, try an Intel made motherboard ;-)
Martin Atkinson-Barr
Trent Worthington wrote:
> On Mon, 07 Aug 2000 22:12:18 GMT, Martin Atkinson-Barr
> <mc...@worldnet.att.net> wrote:
>
> >
> >
> >Trent Worthington wrote:
> >
> >>
> >>
> >> The question is, would an AMD part work on his board. He did not state
> >> the manufacturer of the motherboard, so they may not work.
> >> --
> >>
> >
> >Well I've never found a Socket 7 motherboard that didn't work with a K6 series - I've
> >used about 8 K6-IIIs in various motherboards and a lot more K6-2s. Sometimes you need
> >a BIOS update (remember its the BIOS writer that does CPU detection, not the
> >motherboard mftr i.e Award, AMI etc).
> >
> >If anyone finds a socket7 that cannot make use of a K6 -2 or K6-III lets hear from
> >them. I think AMD also has a list of "approved motherboards" on their website -
> >however omission from the list does not mean the K6 won't work.
> >
>
> Ahem, try an Intel made motherboard ;-)
>
> --
>
I have, they work fine.
mel...@teleport.com
Martin Atkinson-Barr wrote:
That would be a pretty neat trick, at least if the bios was original Intel. As an Intel bios
from the TX era checks the cpuid just about the first thing up. And if it isn't "Genuine
Intel" then it just stops.
The Shadow Knows....
Trent Worthington
>
>
>Martin Atkinson-Barr wrote:
>
>> Trent Worthington wrote:
>>
>> > On Mon, 07 Aug 2000 22:12:18 GMT, Martin Atkinson-Barr
>> > <mc...@worldnet.att.net> wrote:
>> >
>> > >
>> > >
>> > >Trent Worthington wrote:
>> > >
>> > >>
>> > >>
>> > >> The question is, would an AMD part work on his board. He did not state
>> > >> the manufacturer of the motherboard, so they may not work.
>> > >> --
>> > >>
>> > >
>> > >Well I've never found a Socket 7 motherboard that didn't work with a K6 series - I've
>> > >used about 8 K6-IIIs in various motherboards and a lot more K6-2s. Sometimes you need
>> > >a BIOS update (remember its the BIOS writer that does CPU detection, not the
>> > >motherboard mftr i.e Award, AMI etc).
>> > >
>> > >If anyone finds a socket7 that cannot make use of a K6 -2 or K6-III lets hear from
>> > >them. I think AMD also has a list of "approved motherboards" on their website -
>> > >however omission from the list does not mean the K6 won't work.
>> > >
>> >
>> > Ahem, try an Intel made motherboard ;-)
>> >
>> > --
>> >
>>
>> I have, they work fine.
>
>That would be a pretty neat trick,
Indeed.
> at least if the bios was original Intel. As an Intel bios
>from the TX era checks the cpuid just about the first thing up. And if it isn't "Genuine
>Intel" then it just stops.
>
>The Shadow Knows....
--
Martin Atkinson-Barr
Trent Worthington wrote:
Interesting, then I stand corrected - I have never found a system that wouldn't upgrade to a
K6. I guess the Intel motherboards I tried didn't have Intel BIOS in them.
Trent Worthington
<mc...@worldnet.att.net> wrote:
Are you confusing an Intel *made* motherboard versus a motherboard
made by another manufacturer with an Intel chipset? Also be aware that
Intel never made a socket 7 chipset with a 100Mhz bus, which the
faster K6-2/3s would make use of.
Martin Atkinson-Barr
Trent Worthington wrote:
No, some of the motherboards I upgraded that had Intel motherboards had the Yamaha sound chipset,
they had an AMI Bios I think (I can check this weekend if you wish).
I remember I downloaded new sound drivers for those particular systems from the Intel website -
that's why I am sure they were real Intel motherboards, apart from the mftr ID on the board itself
which led me to the Intel site..
If Intel disabled non "GenuineIntel" processors in the BIOS this simply means that the startup
splash screen would identify a non-K6 compatible system by virtue of the Intel BIOS. Most of the
systems I have worked on are either AMI, Award or OEM - like HP and Compaq. Recent AMI and Award
BIOS "know" about the K6-III and identify it correctly at startup even though there is no official
mftr support (for example IWILL states their motherboard only goes to 300MHz K6 on their website).
All of the K6-2s and K6-IIIs are spec'd up to a 100MHz bus but work fine at 66MHz (I'm using a
K6-III on a TX based IWILL P55XUB with the bus set at 68.4MHz to write this, that's one of the bus
speeds supported - at 75MHz and 83MHz this system has PCI stability problems with the boards I
use). By running the chips at 6xbus this system runs at 411MHz (on a 450MHz spec K6-III). I
normally set up systems for other people at 66MHz x 6 = 400MHz to ensure that they remain stable
whatever additional boards are added.
I know what you mean when you say a 3rd party motherboard mftr and an Intel chipset. My point is
that there are no fundamental problems with the hardware on the motherboard and the K6s. If the
BIOS writer puts in an Intel processor-specific piece of code then that's not a motherboard
incompatibility is it? On the IWILL P55XUB I did have to update the BIOS for the K6-2s. I think
perhaps the more recent systems (last 4 years or so) using Intel motherboards may have avoided
Intel BIOS for exactly the reason of the K6 - to give the OEM some flexibility. Even if an Intel
BIOS were present there are still options - doesn't MrBIOS (and perhaps others) offer replacement
code?
Regards.
Jason Whorton
there an Intel made BIOS? I know of the chipsets (VX, TX, etc., but not of
an Intel made BIOS (Award, AMI, Phoenix). Also, early socket 7 boards will
not take one of these later Super Socket 7 CPU's. Also, if you are running
the bus at 66MHz, you are wasting your time and delivering a crippled
product for no reason.
Thank you,
Jason Whorton
www.microxl.com/oldcomputers/main.html
www.microxl.com/oldcomputers/Model60.html
Martin Atkinson-Barr <mc...@worldnet.att.net> wrote in message
news:3992F553...@worldnet.att.net...
Martin Atkinson-Barr
Jason Whorton wrote:
> Hi. I noticed your thread and wanted to make a few points/questions. Is
> there an Intel made BIOS? I know of the chipsets (VX, TX, etc., but not of
> an Intel made BIOS (Award, AMI, Phoenix). Also, early socket 7 boards will
> not take one of these later Super Socket 7 CPU's. Also, if you are running
> the bus at 66MHz, you are wasting your time and delivering a crippled
> product for no reason.
>
> Thank you,
> Jason Whorton
> www.microxl.com/oldcomputers/main.html
> www.microxl.com/oldcomputers/Model60.html
>
Thanks for your comment. Socket 7 boards accept K6s in all flavors provided
there is a correct voltage provided. If there is no correct voltage option a
power converter socket is made by Powerleap. Running a K6-2 at 6x66Mhz = 400MHz
works very well, not at all crippled. Better still use a K6-III where the on-die
cache reduces memory accesses (just like the difference between the old Celeron
300 and the 300A).