70ns SIMM's with P100's or higher
thei...@torolab6.torolab.ibm.com
I'm looking to upgrade my P66 system to a P100 (possibly P120, P133)
motherboard with the triton chipset (i.e. ASUS P55TP4-XE).
And and I'd really like to use my existing 72-pin 70ns SIMM's
on the new m/b. First of all, can they be used with the faster
pentiums with enough waitstates? And if so, how severe is the
performance hit? i.e, is a P120 / 70ns SIMM's combo with lots of
waitstates effectively slower than a P90 / 70ns combo with a lower
number of waitstates? ( I'm talking in terms of OS's like Linux/NT/OS2/DOS
and regular applications and DOS games which aren't overly memory
intensive )
Any info would be much appreciated.
Leonard Theivendra
IBM Software Solutions
John Horigan
The 120 MHz and 90 MHz Pentium processors both run with a 60 MHz CPU
bus so your DRAM would have the same wait states for both. In my
limited experience I have not seen a chipset memory controller that
can get better timings with 60ns 72 pin fast page mode (FPM) SIMMs than
it can with 70ns SIMMs, the 10ns difference is not enough to shave off
a clock at 60 MHz. 60ns EDO SIMMs generally have one less burst wait
state than 70ns FPM SIMMs, but they both have the same lead-off wait
states.
All in all, I would say that your SIMMs will do fine in the ASUS P55TP4-XE
motherboard. I'm not familiar with ASUS motherboards, make sure you get
one with a synchronous (or pipeline sync burst) cache.
--------------------------------------------------------------------------
John W. Horigan MHPG Mobile Modelling and Architecture M/S: SC9-15
Intel, Santa Clara (408)765-4773 jhor...@mipos3.intel.com SC9-2 pole F3
'Demand unearned rewards.' -- Survival Research Labs
VoiceCom Systems
>In article <3njpr6$6...@tornews.torolab.ibm.com>, thei...@torolab6.torolab.ibm.com writes:
>|>
>|> I'm looking to upgrade my P66 system to a P100 (possibly P120, P133)
>|> motherboard with the triton chipset (i.e. ASUS P55TP4-XE).
>|> And and I'd really like to use my existing 72-pin 70ns SIMM's
>|> on the new m/b. First of all, can they be used with the faster
>|> pentiums with enough waitstates? And if so, how severe is the
>|> performance hit? i.e, is a P120 / 70ns SIMM's combo with lots of
>|> waitstates effectively slower than a P90 / 70ns combo with a lower
>|> number of waitstates? ( I'm talking in terms of OS's like Linux/NT/OS2/DOS
>|> and regular applications and DOS games which aren't overly memory
>|> intensive )
>|>
>|> Any info would be much appreciated.
>|>
>|>
>|> Leonard Theivendra
>|> IBM Software Solutions
>|>
>The 120 MHz and 90 MHz Pentium processors both run with a 60 MHz CPU
>bus so your DRAM would have the same wait states for both. In my
>limited experience I have not seen a chipset memory controller that
>can get better timings with 60ns 72 pin fast page mode (FPM) SIMMs than
>it can with 70ns SIMMs, the 10ns difference is not enough to shave off
>a clock at 60 MHz. 60ns EDO SIMMs generally have one less burst wait
>state than 70ns FPM SIMMs, but they both have the same lead-off wait
>states.
My father just bought a Micronics M54Pi board with a 100MHz Pentium
and he got 2 8Meg 70ns SIMMs. I tried out a game (Magic Carpet) to
check this new beast out, and it crashed over and over. After shutting
down all memory shadowing (but leaving both internal and external caches
enabled), it seemed to become more stable. I have only fiddled with it
for one night so far, but I think a 100MHz machine (which means 66MHz
in the motherboard) NEEDS 60ns SIMMs. He did buy the extra 256k of cache
(total 512k), and both sets of chips (original, soldered on MB) and
the socketed are rated at 15ns. I wonder if the cache speeds could be
marginal? He is going to try to get 60ns if possible and if anyone cares
I'll post a follow up.
He did try to install windows before I disabled the shadowing, and it
hung during the install procedure. After a couple tries, it finally installed
but then when he tried to get into Windows it hung (Please, no jokes).
Any ideas? Am I totally wrong? The Micron P90's we have here at work
are using the exact same MB and they have 60ns SIMMs, just for reference.
I only wish they made 50ns.
>--------------------------------------------------------------------------
>John W. Horigan MHPG Mobile Modelling and Architecture M/S: SC9-15
>Intel, Santa Clara (408)765-4773 jhor...@mipos3.intel.com SC9-2 pole F3
>'Demand unearned rewards.' -- Survival Research Labs
- Brian (voic...@world.std.com)
VoiceCom Systems
>VoiceCom Systems (voic...@world.std.com) wrote:
>: jhor...@gws27.sc.intel.com (John Horigan) writes:
>: My father just bought a Micronics M54Pi board with a 100MHz Pentium
>: and he got 2 8Meg 70ns SIMMs. I tried out a game (Magic Carpet) to
>: check this new beast out, and it crashed over and over. After shutting
>: down all memory shadowing (but leaving both internal and external caches
>: enabled), it seemed to become more stable. I have only fiddled with it
>: for one night so far, but I think a 100MHz machine (which means 66MHz
>: in the motherboard) NEEDS 60ns SIMMs. He did buy the extra 256k of cache
>: (total 512k), and both sets of chips (original, soldered on MB) and
>: the socketed are rated at 15ns. I wonder if the cache speeds could be
>: marginal? He is going to try to get 60ns if possible and if anyone cares
>: I'll post a follow up.
>I care. Please post a follow up.
OK, This gets sort of interesting.... Here is the follow up:
After mucho playing with settings and other things, including bringing
the machine to the place we got the motherboard, it seems that it was
the extra 256k of cache that was causing MOST of the problems. After
removing the extra 256k of cache (leaving only the standard soldered 256k)
the system ran Windows without any problems and also ran the game (Magic
Carpet). But the game will STILL hang (with or without external cache enabled)
if I hit <ESC> to go to the main menu, then click to restart the level or
rejoin the game. Every time I tried, it hung in the exact same place.
I tried the game on the Micron P90 (Micronics M54Pi motherboard) and it
never hung. But there are some differences between the Micron's motherboard
and my dad's: The Micron has a P90, my dad has a P100. The Micron has the
buggy P90, my dad has the fixed P100. The Micron has a PS/2 keyboard connector,
my dad has a standard size keyboard connector. The Micron has a PS/2 mouse
port, my dad has NO mouse port. The Micron has 2 8Meg 60ns SIMMS installed in
bank 2 (!!!), my dad has 2 8Meg 70ns SIMMS installed in bank 0 (like the
manual says).
I tried moving the Micron's 60ns memory to bank 0, and the machine didn't
even notice. It worked just fine either way (isn't that weird?).
So, I do NOT think the game has a bug that is crashing the system, because
it works just fine on a (very) similar motherboard with the same name
(M54Pi). I have not tried 60ns SIMMs yet. (The place doesn't have any, if
you believe it). I believe they have been doing a lot of testing on the
motherboard and they cannot get 512k of cache to work very well at all,
no matter the brand or speed.
The manual to the motherboard is from JUNE 1994, and is totally wrong
or completely misses key sections about system setup (like jumpers to set
when using a P100 vs P90, etc.)
I am going to pick up a Triton-based motherboard made by Micronics tonight.
It will have 256k of cache (they ordered extra cache, but it isn't in).
I hope this board doesn't hang.
Some notes:
I always thought Micronics made some of the best motherboards out there,
but after this fiasco, I really am not too sure anymore. The 100MHz Pentium
has been out for months. 512k of cache is nothing major. Why doesn't Micronics
have a new manual to go along with the M54Pi?
Micronics tech support couldn't even help the dealer when he called them
up with questions about jumper settings because the manual didn't have the
info (or was wrong, in the case of cache jumper settings).
The dealer still thinks there might be some jumper problem or something that
is causing the lockups when using 512k cache.
At least the dealer was nice enough to offer a Triton-based board for the
same price. But I bet it won't have EDO RAM or burst cache. I am praying
that the cache they had to order WAS burst cache, otherwise why buy Triton?
So, to summarize:
The Micronics M54Pi (with the standard keyboard connector) seems
to have big problems with P100 chips and/or 512k cache. It hung with
external cache turned OFF (even with only 256k), so who knows.
I guess that's about it.. This Triton-based thing better work.
- Brian
Franklin T. Kang
sympathize...
Attempt #1
SuperMicro P54CI board with 512k 15nS cache and 32 MB 70nS DRAM
The system would constantly randomly reboot. I took it back to the dealer
and he claims that the manufacturer shipped the wrong speed of cache
SRAM. He believes 100Mhz requires 12nS. He didn't have any other
motherboards in stock and offers to trade me another board.
Attempt #2
SuperMicro P55CWA board with 256k 12nS cache and 32 MB 70nS DRAM
This time the sytem would randomly lock up. This occured no matter
whether I was in DOS, Win16 or Win NT. Mysteriously Dr. Watson
refused to run under Win NT. After letting the dealer have the
computer back for a week to run tests, they couldn't find the problem.
I have to suspect the problems for both boards may be the 70nS
DRAM. The manual does say 60nS should be used.
Aftermath
By this point, 2 weeks have passed and I have spent 20+ hours on
my $4000 system unsucessful in getting it to work. I have opened
up the computer and switched parts so many times that I can
probably do it in the dark. So, I give up and I return the computer
for a refund.
I'm so sick of dealing with this that I don't even feel like buying
another computer. However, I manage to place an order with Dell.
Hopefully, when it arrives I'll be able to unpack it and start using
it.
Frank
Philip M. Burton
: evan...@cs.uidaho.edu (Jason Evans) writes:
: The manual to the motherboard is from JUNE 1994, and is totally wrong
: or completely misses key sections about system setup (like jumpers to set
: when using a P100 vs P90, etc.)
: I always thought Micronics made some of the best motherboards out there,
: but after this fiasco, I really am not too sure anymore. The 100MHz Pentium
: has been out for months. 512k of cache is nothing major. Why doesn't Micronics
: have a new manual to go along with the M54Pi?
: So, to summarize:
: The Micronics M54Pi (with the standard keyboard connector) seems
: to have big problems with P100 chips and/or 512k cache. It hung with
: external cache turned OFF (even with only 256k), so who knows.
Brian,
Aside from any other issues, your dealer should have known that the
early'94 generation of Pentium motherboards from Micronics wouldn't work
at 100 MHz. The fault may have been the vendor's. A friend used to work
at Micronics, in marketing, and he once volunteered this information to
me, late last year. Don't know when/if the problem got fixed.
Once upon a time, kiddies, when the world was young, there were (only)
three wise motherboard-maker-men. Their names were Micronics, Mylex, and
AMI.
Now, kiddies, how many of the systems that you saw in the last PC
Magazine review used one of these motherboards? Were they the best
performers of the test group? In the most economical systems? Methinks not.
The moral, if there is one: Past history is not always a reliable guide
to the presen tin this crazy industry.
--
Phil Burton p...@netcom.com
Palo Alto, CA
VoiceCom Systems
>In <D80KJ...@world.std.com> voic...@world.std.com (VoiceCom Systems) writes:
>>evan...@cs.uidaho.edu (Jason Evans) writes:
>>>VoiceCom Systems (voic...@world.std.com) wrote:
>>>: jhor...@gws27.sc.intel.com (John Horigan) writes:
>[micronics horror story deleted]
>Have you tried a different keyboard?
Hmmmm.. no I didn't think of that. The keyboard is a Northgate OMNIkey
which has been in use for about 5 years on a 386-20. I'll have to give
that a try.
I do not have the new Triton-based board yet. The dealer wants to do
more testing. I'm going to try to get it on Friday.
- Brian
VoiceCom Systems
>On Wed, 3 May 1995 17:57:22 GMT VoiceCom Systems (voic...@world.std.com) wrote:
>: evan...@cs.uidaho.edu (Jason Evans) writes:
>: The manual to the motherboard is from JUNE 1994, and is totally wrong
>: or completely misses key sections about system setup (like jumpers to set
>: when using a P100 vs P90, etc.)
>: I always thought Micronics made some of the best motherboards out there,
>: but after this fiasco, I really am not too sure anymore. The 100MHz Pentium
>: has been out for months. 512k of cache is nothing major. Why doesn't Micronics
>: have a new manual to go along with the M54Pi?
>: So, to summarize:
>: The Micronics M54Pi (with the standard keyboard connector) seems
>: to have big problems with P100 chips and/or 512k cache. It hung with
>: external cache turned OFF (even with only 256k), so who knows.
>Brian,
>Aside from any other issues, your dealer should have known that the
>early'94 generation of Pentium motherboards from Micronics wouldn't work
>at 100 MHz. The fault may have been the vendor's. A friend used to work
>at Micronics, in marketing, and he once volunteered this information to
>me, late last year. Don't know when/if the problem got fixed.
So, you mean the jumpers on the motherboard that switch the main system
frequency from 60MHz to 66MHz are just for show? Come on.. that is WHY
I bought Micronics, rather than some other cheap(ahem) brand. I thought
I was paying for the piece of mind knowing the board would work properly.
The manual was from June 1994, but the board itself sure wasn't. The
manual never addressed the jumpers that change the clock freq.. The dealer
had to call Micronics tech support and even then it took a long time for
tech support to figure out what jumper goes where for 100MHz operation.
>Once upon a time, kiddies, when the world was young, there were (only)
>three wise motherboard-maker-men. Their names were Micronics, Mylex, and
>AMI.
>Now, kiddies, how many of the systems that you saw in the last PC
>Magazine review used one of these motherboards? Were they the best
>performers of the test group? In the most economical systems? Methinks not.
>The moral, if there is one: Past history is not always a reliable guide
>to the presen tin this crazy industry.
>Phil Burton p...@netcom.com
>Palo Alto, CA
So, you are saying Micronics went "cheap"?
Who makes the BEST motherboard for a Pentium100 today, then?
- Brian
Robert L. Holder @ignite
>evan...@cs.uidaho.edu (Jason Evans) writes:
>>VoiceCom Systems (voic...@world.std.com) wrote:
>>: jhor...@gws27.sc.intel.com (John Horigan) writes:
Philip M. Burton
: p...@netcom.com (Philip M. Burton) writes:
: >On Wed, 3 May 1995 17:57:22 GMT VoiceCom Systems (voic...@world.std.com) wrote:
: >: evan...@cs.uidaho.edu (Jason Evans) writes:
: >: I always thought Micronics made some of the best motherboards out there,
: >: but after this fiasco, I really am not too sure anymore. The 100MHz Pentium
: >Aside from any other issues, your dealer should have known that the
: >early'94 generation of Pentium motherboards from Micronics wouldn't work
: >at 100 MHz. The fault may have been the vendor's. A friend used to work
: >at Micronics, in marketing, and he once volunteered this information to
: >me, late last year. Don't know when/if the problem got fixed.
: So, you mean the jumpers on the motherboard that switch the main system
: frequency from 60MHz to 66MHz are just for show? Come on.. that is WHY
This is Phil Burton, who wrote the part that you replied to.
No the jumpers were not for show. They were for real. But, as it so
often happens in Silicon Valley (the state of mind, not the geography),
at the very, very last minute, someone runs compatibility/qualification
tests. Then it emerges that there is a problem. In this case, according
to my ex-Micronics pal, the problem was chipset-related. He didn't explain
further. So, until the good folks at Micronics did their tests, they didn't
necessarily know that their chipset vendor had a problem.
Is this fair? No. Is it a common scenario? Yes, all too common. Do
vendors always come clean if they don't have do. Hell, NO. [How do I
pronounce Intel?]
: I bought Micronics, rather than some other cheap(ahem) brand. I thought
Well, once upon a time, it was a premium brand. And it still may be. I
dunno. But, times change, people change, and the competitive environment
changes. All of which means that the people and company that could produce
great motherboards in the 386 and 486 eras may not still exist. It could be
a "new company," because many of the players are new. Engineering,
manufacturing, marketing, finance. I don't know anything specific about
Micronics, but I have seen this situation all too often in the Valley.
Even worked in a few such companies myself. (Anyone remember Fortune
Systems?)
Another point. Maybe someone figured that by the time there were many
P100s availble, that they, Micronics, would have fixed the problem. Perhaps
they actually did, but you got an "old" board. Check rev numbers and date
codes and call up your dealer.
: I was paying for the piece of mind knowing the board would work properly.
No, you were paying for a good reputation based on past accomplishments.
: The manual was from June 1994, but the board itself sure wasn't. The
: manual never addressed the jumpers that change the clock freq.. The dealer
Gee out of date documentation. Never happened on *my* watch, nosiree, no
matter what the VP of Sales had to say about this month's numbers. No, no
way Jose, not ever. And, yes, on one of those watches, I saw Santa Claus
leave Virginia her presents. Saw the ol' guy hisself.
: had to call Micronics tech support and even then it took a long time for
Did they actually help? Better than the average bear, er company. In
English yet, extra point!
: tech support to figure out what jumper goes where for 100MHz operation.
Take that issue up with the VP of Sales. Hey, he made his quota for the
month, and you ain't buying more, so what does he care.
: >The moral, if there is one: Past history is not always a reliable guide
: >to the presen tin this crazy industry.
: >Phil Burton p...@netcom.com
: >Palo Alto, CA
: So, you are saying Micronics went "cheap"?
Dunno. Just that they didn't exactly sell what you expected. There is a
novel theory called "Implied Fitness of Merchantability," which kind of
says that a company that sells you soemthing has to make sure it works
for you, or you get your money back, etc. See a lawyer for details.
: Who makes the BEST motherboard for a Pentium100 today, then?
Dunno. What defines "best?"
Once you answer that question, read some reviews. Get lots of samples
and conduct your own tests. Including extended burnins to simulate end
of life. That is what the big guys do, if they buy any outside part.
You and can't exactly do that. We are at the mercy of PC Magazine and
Computer Shtupper, er Shopper. Grim, isn't it.
I think, seriously, that our only protection is the net.
Hope this was helpful.
: - Brian
--
Steve Bonds
VoiceCom Systems <voic...@world.std.com> wrote:
>I am going to pick up a Triton-based motherboard made by Micronics tonight.
>It will have 256k of cache (they ordered extra cache, but it isn't in).
>I hope this board doesn't hang.
>
>Some notes:
>
> I always thought Micronics made some of the best motherboards out there,
>but after this fiasco, I really am not too sure anymore. The 100MHz Pentium
>has been out for months. 512k of cache is nothing major. Why doesn't Micronics
>have a new manual to go along with the M54Pi?
Companies seem to be pumping out pentium board at a fantastic rate. I haven't
seen one yet that I would call "bug-free". Even the P90 boards are flaky, and
the P100s are much less stable.
>I guess that's about it.. This Triton-based thing better work.
Good luck. I've seen some problems with Triton boards. The cache is
just as twitchy (but this is a VERY common problem-- even on good 486 boards)
and the speed just isn't anything like what they've advertised.
Now that pentium prices have actually dropped a bit for retail, the only thing
holding me back is the lack of a truly stable pentium motherboard. I have my
486DX4/100 and it takes all the abuse I can dish out (strange OS's, constantly
changing parts, bizarre memory configurations, etc. etc.)
If someone has a recommendation for a REALLY stable P90 board, let me know and
I'll take a look at it. My latest disappointment was an SIS chipset board
with an Award BIOS that just plain refused to do LBA on a WD31000 or 31200
EIDE drive. (SIS85C50X shipset) Different boards/drives showed the same
problems. :(
-- Steve Bonds
--
000 000 7777 | sbo...@u.washington.edu and Steve...@hmc.edu
0 0 0 0 7 |-----------------------------------------------------------
0 0 0 0 7 | Childhood is short... [Calvin & Hobbes]
000 000 7 | ...but immaturity is forever.
Edmond Underwood
the machine to the place we got the motherboard, it seems that it was
the extra 256k of cache that was causing MOST of the problems. After
removing the extra 256k of cache (leaving only the standard soldered 256k)
the system ran Windows without any problems and also ran the game (Magic
Carpet). But the game will STILL hang (with or without external cache enabled)
if I hit <ESC> to go to the main menu, then click to restart the level or
rejoin the game. Every time I tried, it hung in the exact same place.
I tried the game on the Micron P90 (Micronics M54Pi motherboard) and it
never hung. But there are some differences between the Micron's motherboard
and my dad's: The Micron has a P90, my dad has a P100. The Micron has the
buggy P90, my dad has the fixed P100. The Micron has a PS/2 keyboard connector,
my dad has a standard size keyboard connector. The Micron has a PS/2 mouse
port, my dad has NO mouse port. The Micron has 2 8Meg 60ns SIMMS installed in
bank 2 (!!!), my dad has 2 8Meg 70ns SIMMS installed in bank 0 (like the
manual says).
I tried moving the Micron's 60ns memory to bank 0, and the machine didn't
even notice. It worked just fine either way (isn't that weird?).
, I do NOT think the game has a bug that is crashing the system, because
- Brian
--------------------------------------------------------------------
This is all very interesting. I have a Micronics M54Pi with 512k cache and mixed
60ns and 70ns SIMMS. As far as where the SIMMS are supposed to be, you start
from bank 0 and move your way up (although Micron puts them at bank 2, hmmm...).
I overclocked it to 100 MHz by changing jumpers 1 and 36-38. No problems.
I would think that if the extra cache were causing problems, it would have
to do with the BIOS interacting with the cache and letting the system
know that the cache is active or non active (not to mention JP2 being closed).
I am convinced that the problem is doubtfully not the Micronics motherboard
here, (I know many people using that same Motherboard w/ a 100 MHz Pentium
who aren't having problems) but that it may either be a defective board, an
old BIOS (lates BIOS is Phoenix 4.04 N 15 for the serial mouse), or incorrect
settings on the motherboard or the BIOS. What are your jumpers set to on the
motherboard, particularly JP's 1,2,36-38? That may tell us a little more.
Best of Luck!
---------------------------------------------------------------------------------
Edmond Underwood
Systems Management Group
Computing & Network Services (University of Colorado)
E-mail: unde...@Colorado.Edu
Philip M. Burton
: In article <D80KJ...@world.std.com>,
: VoiceCom Systems <voic...@world.std.com> wrote:
: Companies seem to be pumping out pentium board at a fantastic rate. I haven't
: seen one yet that I would call "bug-free". Even the P90 boards are flaky, and
: the P100s are much less stable.
Many of the boards are "supposed" to support both P90 and P100 with a jumper
change. Obviously the Micronics board we have been discussing is an
egregious example of a board that doesn't. Others? Specifically boards by
Intel or Asus? Which chipset?
: Good luck. I've seen some problems with Triton boards. The cache is
: just as twitchy (but this is a VERY common problem-- even on good 486 boards)
What is "twitchy?"
: and the speed just isn't anything like what they've advertised.
Do you mean that if you run a benchmark under the same conditions as a
reviewer, you get lower scores?
: 486DX4/100 and it takes all the abuse I can dish out (strange OS's, constantly
: changing parts, bizarre memory configurations, etc. etc.)
NEVER, NEVER had any of that on the no-name 386/40 I'm using now.
: I'll take a look at it. My latest disappointment was an SIS chipset board
: with an Award BIOS that just plain refused to do LBA on a WD31000 or 31200
: EIDE drive. (SIS85C50X shipset) Different boards/drives showed the same
: problems. :(
ASUS? Other? If ASUS, have you upgraded to the latest BIOS?
As a general comment, I think that flash BIOSes are a GOOD THING, but only
if you've bought from a reputable vendor who will actually do updates. The
flash approach is an implicit admission that no one can get it right the
first time. I'm glad for that admission. It's healthy and good for users.
Ben Moore
>Once you answer that question, read some reviews. Get lots of samples
>and conduct your own tests. Including extended burnins to simulate end
>of life. That is what the big guys do, if they buy any outside part.
>You and can't exactly do that. We are at the mercy of PC Magazine and
>Computer Shtupper, er Shopper. Grim, isn't it.
And, what a cooincidence! They're both published by the same company!
Ziff-Davis! In fact, the majority of computer magazines come from
Ziff-Davis... that might explain a few things...
Venkat
It is Well known that the Micronics M54Pi does *NOT* work reliably with
P100, untill the latest revison, and correct jumper settings( not well
documented). (if you need it, send email).
MICRON has changed over to the MICRONICS M54Si MB for their entire line of
"PowerStation" systems. They work very *reliably* with P100 cpus.
In any case 60 ns simms are recommended for P100 and above.
Venkat Thirumalai
(email : ven...@infinet.com) *** email preferred ***
(Fax # 1.800.600.7105)
Ben Moore
>
>On 5 May 1995 19:09:05 GMT Steve Bonds (sbo...@u.washington.edu) wrote:
>: In article <D80KJ...@world.std.com>,
>: VoiceCom Systems <voic...@world.std.com> wrote:
>
>: Companies seem to be pumping out pentium board at a fantastic rate. I
haven't
>: seen one yet that I would call "bug-free". Even the P90 boards are
flaky, and
>: the P100s are much less stable.
>
>Many of the boards are "supposed" to support both P90 and P100 with a
jumper
>change. Obviously the Micronics board we have been discussing is an
>egregious example of a board that doesn't. Others? Specifically boards by
>Intel or Asus? Which chipset?
According to it's faxback, the Intel Premiere PCI II with the Neptune
chipset will support the 75/90Mhz or the 100Mhz, but not the 100 if it uses
the 75/90, and vice-versa. Of course, they don't bother to explain why...
it just says, "Due to design differences..." "What design differences?" you
might ask. Well, it seems they've taken Apple's approach to things: "Don't
ask! You don't need to know." "Bullshit!" I say...
GUY NUSSBAUM
computer booted up and everything was going as planed, but, the boot up
went maybe half way before the system rebooted itself up again
Why?
not that I was going to keep it overclocked , I do'nt have a fan yet,
but when I get the fan what should I do for the system not to reboot,
set the wait states what?
My system is a
33dx
256 cache 3-1-1-1 (setting if I'm not mistaken)
4meg 30 pin 70ns simms
4 meg 72 pin simms 70 ns
bus logic caching controller 2 meg onboard
vl-bus speed has 2 options under <33 over >33
and a jumper to set wait state on/off
So any help (if any) is going to be apriciated
PLEASE!!!!!PLEASE e-mail responses
Thank you
GUY NUSSBAUM
Jason Elledge
a year with no problems. The only thing I can think of is to change the
bios wait states to a slower setting (for goodness sake not 0) On my
motherboard there is a jumper the cache wait state (in addition to the
one in the bios--go figure) but if I set the jumper to 0 wait the system
runs at incredible speed and dies after a minute. But set at wait 1,
the system runs just fine at 40MHz. Of course you could have one of
those rare chips that just won't overclock well.
On 8 May 1995, GUY NUSSBAUM
wrote:
> Date: 8 May 1995 06:41:48 GMT
> From: GUY NUSSBAUM <bro...@ix.netcom.com>
> Newgroups: comp.sys.ibm.pc.hardware.chips, comp.sys.intel,
> comp.sys.ibm.pc.hardware.misc
> Subject: Overclocking 33dx, did it and....
Nick Stefanisko
: I have to suspect the problems for both boards may be the 70nS
: DRAM. The manual does say 60nS should be used.
That would be my guess too. If the manual says you have to use 60ns or
faster then you probably have to use 60ns or faster. 70's just won't
cut it.
VoiceCom Systems
As I stated on my previous posting, I got the new Micronics motherboard
with the Triton chipset last Friday (5/5/95). The model number is M54Hi
and so far, it has worked very well. Only 2 programs are having trouble
running: PC Geoworks (remember that?) and the MKS toolkit (UNIX tools for
DOS). I think Geoworks is locking quite hard (it could be some configuration
issue with the program), and MKS is locking softly (ie: hitting the keyboard
16 times will generate a beep (keyboard buffer full)). These are really no
big deal, and perhaps they can be remedied.
Unlike the previous board (M54Pi), this one had a very recent manual
(April 1995) and the QC date was 4/28/95 (I think), so this board was made
about 2 weeks ago. It has the Triton chipset, supports Fastpage mode or EDO
SIMMs (4 slots, 128Meg max). It also supports 3 types of cache (and you must
jumper it according to what cache you have.). Right now it has 70ns SIMMs
and 256k 15ns cache (unknown type of cache, most likely the cheapest: Async?).
It will be upgraded to 60ns SIMMs as soon as the dealer gets some in.
It will also be upgraded to 512k cache (probably not burst/pipelined).
So, it will not be the fastest P100 you can buy, but it should perform quite
well.
I am quite happy with this board and would recommend it to anyone looking
for a good Triton board with a SOCKET5. The dealer has done a fair amount of
testing using various operating systems: SCO UNIX/ODT, X86 Solaris, DOS/
Windows, OS/2, Win95, Novell. Not sure about Win/NT.
Many thanks to all that posted/emailed replies about my problem.
- Brian
9200...@cpccux0.cityu.edu.hk
: So, you mean the jumpers on the motherboard that switch the main system
: frequency from 60MHz to 66MHz are just for show? Come on.. that is WHY
: I bought Micronics, rather than some other cheap(ahem) brand. I thought
: I was paying for the piece of mind knowing the board would work properly.
: The manual was from June 1994, but the board itself sure wasn't. The
: manual never addressed the jumpers that change the clock freq.. The dealer
: had to call Micronics tech support and even then it took a long time for
: tech support to figure out what jumper goes where for 100MHz operation.
I'm also using M54Pi/512K cache with P5/90 processor, it seems to be the first
generation 3.3v P5 MB from Micronics Lab. So far it working well after getting
BIOS upgrade except I've to update BIOS everytime I change the slot of card -
it's interesting: isn't their Phoenix BIOS plug-n-play BIOS?
Anyway may I know which jumper used for changing the clock rate?
: So, you are saying Micronics went "cheap"?
I think Micronic's MB is quite good and update indeed. The day I bought my MB
back in JUNE 1994, M54Pi is the first one using the bug-fixed 'GT' version
of SMC I/O chip. The design and expandability of their MB is very good also.
For their 6 RAM slot, I can easily add another 32MB RAM to make my system
total 96MB without changing the old 16M/bar SIMMs. For their RAM slot
arrangement, I have flexible access to RAM slots without get the MB out of my
chassis in order to change SIMM arrangement. OTOH, for other common 4 RAM slots
old fashioned MB, there is a big trouble to do these.
I also like their simplified BIOS setting - I really tired to deal with those
'feature rich' Award BIOS...
Until now Micronics still makes solid MB I think.
: Who makes the BEST motherboard for a Pentium100 today, then?
Maybe still Micronics: Millineia... :>
Rgds
Chan Hong
Andrys D Basten
>ball/10464 <3odrt5$t2 comp.sys.ibm.pc.hardware.chips:28305 alt.comp.periphs.mainboard.asus:2496 comp.sys.intel:41676 comp.sys.ibm.pc.hardware.misc:27436
>In any case 60 ns simms are recommended for P100 and above.
Actually, they're better for borderline conditions with the P90 as well.
When I was having autocad memory halts with EISA 486's which went away with
70ns chips (years ago), it caused me to figure that 60ns might be better
when floating point calcs were 5-6 times faster with the PCI Pentium 90 on
autocad and modeling programs.
That's what I've been ordering.
However, a couple of months ago, one of these stalled with a 3Com
Etherlink III and the dealer's tech person swapped everything out, including
the motherboard, the Pentium chip, memory, cards, and the 3Com. Nothing
worked.
I suddenly noticed, at the end of the day, that the chips were 70ns rather
than the 60ns I ordered (the store was out of these and 'forgot' I ordered
60). I put in a slower network card and it all worked. Later, they brought
us the 60ns and then the speedy 3Com card no longer caused halting.
- Andrys
cc: in email as requested
--
,------------------------------------------------------------------------,
| Andrys Basten <and...@netcom.com> | CNE, Basten Micro Consulting |
| Have Early Music, Will Travel | Berkeley-San Francisco, CA |
| Harpsichord, recorders | 510-235-3861 |
+------------------------------------------------------------------------+