Two memory modules go bad at the same time?

So my computer's running fine, and BAM, it suddenly becomes
nonfunctional. It repeatedly insta-reboots from inside Windows.
Windows will start successfully, but try using it for a minute or two
and the insta-reboot occurs. I figured maybe it's a problem with
Windows, so I tried my Gentoo Linux LiveCD ... not working either. The
computer boots fine and even sometimes gets the kernel booted, but it
quickly freezes after that.

So I'm thinking it may be a hardware problem. I try memtest86 from
within the Gentoo Linux LiveCD. I get millions of errors. At this
point I should probably mention that my memory consists of 2X Crucial
512MB DDR 333Mhz CL2.5 and my motherboard is an Asus A7N8X-X. So I
think one of my memory modules may be bad, so I try running with just
one. I get the errors in memtest. I try running it with just the
other memory module installed. I get the same errors in memtest:
specifically, the Err-Bits seem always to be 00ff0000.

Is it possible for two memory modules to both go bad simultaneously and
in the same fashion? I think it may be *possible*, but it's not
exactly probable. So what could be the problem then? Is my
motherboard borked?

And just to try to head off a question or two, I checked my BIOS
settings and the memory timing is set to 333Mhz.

"Cyde Weys" <cydeweys@gmail.com> wrote in message news:1129410518.445581.164210@g49g2000cwa.googlegr oups.com...
> So my computer's running fine, and BAM, it suddenly becomes
> nonfunctional. It repeatedly insta-reboots from inside Windows.
> Windows will start successfully, but try using it for a minute or two
> and the insta-reboot occurs. I figured maybe it's a problem with
> Windows, so I tried my Gentoo Linux LiveCD ... not working either. The
> computer boots fine and even sometimes gets the kernel booted, but it
> quickly freezes after that.
>
> So I'm thinking it may be a hardware problem. I try memtest86 from
> within the Gentoo Linux LiveCD. I get millions of errors. At this
> point I should probably mention that my memory consists of 2X Crucial
> 512MB DDR 333Mhz CL2.5 and my motherboard is an Asus A7N8X-X. So I
> think one of my memory modules may be bad, so I try running with just
> one. I get the errors in memtest. I try running it with just the
> other memory module installed. I get the same errors in memtest:
> specifically, the Err-Bits seem always to be 00ff0000.
>
> Is it possible for two memory modules to both go bad simultaneously and
> in the same fashion? I think it may be *possible*, but it's not
> exactly probable. So what could be the problem then? Is my
> motherboard borked?

Bad slot maybe. Possibly, power supply voltage not within
tolerance.

-- Bob Day
http://bobday.vze.com

Bob Day wrote:
> "Cyde Weys" <cydeweys@gmail.com> wrote in message news:1129410518.445581.164210@g49g2000cwa.googlegr oups.com...
> > So my computer's running fine, and BAM, it suddenly becomes
> > nonfunctional. It repeatedly insta-reboots from inside Windows.
> > Windows will start successfully, but try using it for a minute or two
> > and the insta-reboot occurs. I figured maybe it's a problem with
> > Windows, so I tried my Gentoo Linux LiveCD ... not working either. The
> > computer boots fine and even sometimes gets the kernel booted, but it
> > quickly freezes after that.
> >
> > So I'm thinking it may be a hardware problem. I try memtest86 from
> > within the Gentoo Linux LiveCD. I get millions of errors. At this
> > point I should probably mention that my memory consists of 2X Crucial
> > 512MB DDR 333Mhz CL2.5 and my motherboard is an Asus A7N8X-X. So I
> > think one of my memory modules may be bad, so I try running with just
> > one. I get the errors in memtest. I try running it with just the
> > other memory module installed. I get the same errors in memtest:
> > specifically, the Err-Bits seem always to be 00ff0000.
> >
> > Is it possible for two memory modules to both go bad simultaneously and
> > in the same fashion? I think it may be *possible*, but it's not
> > exactly probable. So what could be the problem then? Is my
> > motherboard borked?
>
> Bad slot maybe. Possibly, power supply voltage not within
> tolerance.

I get the same memory errors no matter which memory module I use or
which slot I put it in. I guess something is wrong with the memory
pathway in general? And this would be a mobo issue (and presumably the
CPU is fine)?

The power supply voltages seems to be in bounds as follows:

VCORE 1.61V
+3.3V 3.23V
+5V 4.83V
+12V 12.22V

Cyde Weys wrote:

> Bob Day wrote:
>
>>"Cyde Weys" <cydeweys@gmail.com> wrote in message news:1129410518.445581.164210@g49g2000cwa.googlegr oups.com...
>>
>>>So my computer's running fine, and BAM, it suddenly becomes
>>>nonfunctional. It repeatedly insta-reboots from inside Windows.
>>>Windows will start successfully, but try using it for a minute or two
>>>and the insta-reboot occurs. I figured maybe it's a problem with
>>>Windows, so I tried my Gentoo Linux LiveCD ... not working either. The
>>>computer boots fine and even sometimes gets the kernel booted, but it
>>>quickly freezes after that.
>>>
>>>So I'm thinking it may be a hardware problem. I try memtest86 from
>>>within the Gentoo Linux LiveCD. I get millions of errors. At this
>>>point I should probably mention that my memory consists of 2X Crucial
>>>512MB DDR 333Mhz CL2.5 and my motherboard is an Asus A7N8X-X. So I
>>>think one of my memory modules may be bad, so I try running with just
>>>one. I get the errors in memtest. I try running it with just the
>>>other memory module installed. I get the same errors in memtest:
>>>specifically, the Err-Bits seem always to be 00ff0000.
>>>
>>>Is it possible for two memory modules to both go bad simultaneously and
>>>in the same fashion? I think it may be *possible*, but it's not
>>>exactly probable. So what could be the problem then? Is my
>>>motherboard borked?
>>
>>Bad slot maybe. Possibly, power supply voltage not within
>>tolerance.
>
>
> I get the same memory errors no matter which memory module I use or
> which slot I put it in. I guess something is wrong with the memory
> pathway in general? And this would be a mobo issue (and presumably the
> CPU is fine)?
>
> The power supply voltages seems to be in bounds as follows:
>
> VCORE 1.61V
> +3.3V 3.23V
> +5V 4.83V
> +12V 12.22V
>

It could be anything, including the CPU. After all, memory can't be read
without the CPU being involved.

But with persistent global errors of that type I'd tend to guess a BIOS
setting before hardware failure. Try 'safe' defaults and, in particular,
the slowest memory speed settings.

David Maynard wrote:

> It could be anything, including the CPU. After all, memory can't be read
> without the CPU being involved.

Is it even possible for a CPU to malfunction in such a specific way
while still letting everything else work? It's not having problems
whatsoever booting and running memtest86 ... it's just when you try to
load an OS (which needs memory) that you get the problems.

> But with persistent global errors of that type I'd tend to guess a BIOS
> setting before hardware failure. Try 'safe' defaults and, in particular,
> the slowest memory speed settings.

It doesn't seem to matter. I get the same errors. I paid attention
more time and it's more like 00XX0000 the first few seconds of
memtest86 where XX are random hex, but then it settles down to Err-Bits
00ff0000 over and over and over.

On 15 Oct 2005 14:08:38 -0700, "Cyde Weys"
<cydeweys@gmail.com> wrote:

>So my computer's running fine, and BAM, it suddenly becomes
>nonfunctional. It repeatedly insta-reboots from inside Windows.
>Windows will start successfully, but try using it for a minute or two
>and the insta-reboot occurs. I figured maybe it's a problem with
>Windows, so I tried my Gentoo Linux LiveCD ... not working either. The
>computer boots fine and even sometimes gets the kernel booted, but it
>quickly freezes after that.


So "BAM" really means that it hardly ever crashed (well, it
IS windows, but...) then regular crashing started occuring?
Had anything changed or any events occurred? For example,
moving the system, being inside it, having recently switched
from AC to electric heater in home, BIOS changes, or ???

>
>So I'm thinking it may be a hardware problem. I try memtest86 from
>within the Gentoo Linux LiveCD. I get millions of errors.

Keep in mind that while most of what memtest does is test
memory, there is also the basic function of the system that
has to be ok too, for memory to be stable.

For example, someone brought a motherboard to me that they'd
stabbed with a screwdriver while installing a heatsink.
They knocked off several surface-mount parts around the
board's northbridge. I managed to get new parts soldered
on, and board works, but is now instable above 154MHz memory
bus. I could probably keep trying to resolder or maybe seek
(in vain) specs for the subcircuit but that's beside the
point, that other things can make memory modules seem
malfunctional even if the vast majority of the system works
fine.

>At this
>point I should probably mention that my memory consists of 2X Crucial
>512MB DDR 333Mhz CL2.5 and my motherboard is an Asus A7N8X-X. So I
>think one of my memory modules may be bad, so I try running with just
>one. I get the errors in memtest. I try running it with just the
>other memory module installed. I get the same errors in memtest:
>specifically, the Err-Bits seem always to be 00ff0000.

OK, but, do these errors ALWAYS occur at same address(es)?


>
>Is it possible for two memory modules to both go bad simultaneously and
>in the same fashion?

Remotely possible, maybe, but usually not. It's more likely
the motherboard, physical parts failing or "maybe" a really
bad power supply (less likely).



>I think it may be *possible*, but it's not
>exactly probable. So what could be the problem then? Is my
>motherboard borked?

I would try underclocking to 100FSB, 100Mem, and
conservative memory timings like 2.5, 3, 3, 7. (you will
have to manually set these, the memory may not stay at
CAS2.5 when the FSB is lowered if board is set to Auto or
SPD mode). I don't if this will prove anything but it would
eliminate physically failure in the modules and if the board
is stable underclocked it might provide some aux. use
instead of being thrown away (or is it under warranty)?
Also if you were using Turbo (or whatever it's called) bios
setting for memory, use the more conservative setting.

Also try clearing CMOS, nForce2 boards can sometimes have
wierd bios glitches... if you're running an early bios you
might also try flashing a newer bios but only if you can
regain stablity in memtest86, first. Do not flash the bios
when system is in a state where memory errors are likely.


>
>And just to try to head off a question or two, I checked my BIOS
>settings and the memory timing is set to 333Mhz.

What make/model power supply?

Does CPU use 166MHz FSB as memory does?

Examine board for failed capacitors. A7N8X used "some"
fairly good caps, but I don't remember on the "-X" version,
I vaguely recall it had some component corners cut but I
can't recall for certain which parts). Some of the caps for
the memory are pretty far over towards the right-side edge
of the board, but check the rest too.

Also check the CPU/heatsink thermal interface. If you had
used regular silicone based thermal grease it may have dryed
out over time... had that happen to a system here, so now I
only use synthetic based grease on flipchips. It might be
unrelated but you made no mention of checking temps (or
voltages for that matter) yet. Also check your memory
voltage settings in the bios.

On 15 Oct 2005 19:09:54 -0700, "Cyde Weys"
<cydeweys@gmail.com> wrote:


>The power supply voltages seems to be in bounds as follows:
>
>VCORE 1.61V
>+3.3V 3.23V
>+5V 4.83V
>+12V 12.22V

Are these readings from a multimeter at the power connector
or from the bios (or windows software) report? If from bios
or windows, they are expected. If from a multimeter, they
could be a sign your PSU may not have enough 5V capacity.

It would be useful to test these memory modules in another
system if possible.

kony wrote:
> On 15 Oct 2005 14:08:38 -0700, "Cyde Weys"
> <cydeweys@gmail.com> wrote:
>
> >So my computer's running fine, and BAM, it suddenly becomes
> >nonfunctional. It repeatedly insta-reboots from inside Windows.
> >Windows will start successfully, but try using it for a minute or two
> >and the insta-reboot occurs. I figured maybe it's a problem with
> >Windows, so I tried my Gentoo Linux LiveCD ... not working either. The
> >computer boots fine and even sometimes gets the kernel booted, but it
> >quickly freezes after that.
>
>
> So "BAM" really means that it hardly ever crashed (well, it
> IS windows, but...) then regular crashing started occuring?

Exactly. Working fine with only the minor videocard problems I
recently talked about, although I even got those fixed until this.

> Had anything changed or any events occurred? For example,
> moving the system, being inside it, having recently switched
> from AC to electric heater in home, BIOS changes, or ???

Nahh, I haven't been inside the system recently or changed anything,
just BAM, not working.


> For example, someone brought a motherboard to me that they'd
> stabbed with a screwdriver while installing a heatsink.
> They knocked off several surface-mount parts around the
> board's northbridge. I managed to get new parts soldered
> on, and board works, but is now instable above 154MHz memory
> bus. I could probably keep trying to resolder or maybe seek
> (in vain) specs for the subcircuit but that's beside the
> point, that other things can make memory modules seem
> malfunctional even if the vast majority of the system works
> fine.

I haven't done anything like that to this mobo :-O

> >At this
> >point I should probably mention that my memory consists of 2X Crucial
> >512MB DDR 333Mhz CL2.5 and my motherboard is an Asus A7N8X-X. So I
> >think one of my memory modules may be bad, so I try running with just
> >one. I get the errors in memtest. I try running it with just the
> >other memory module installed. I get the same errors in memtest:
> >specifically, the Err-Bits seem always to be 00ff0000.
>
> OK, but, do these errors ALWAYS occur at same address(es)?

They occur at every address possible on the memory chips. Everywhere
from 512MB down on to 0MB.

> >I think it may be *possible*, but it's not
> >exactly probable. So what could be the problem then? Is my
> >motherboard borked?
>
> I would try underclocking to 100FSB, 100Mem, and
> conservative memory timings like 2.5, 3, 3, 7. (you will
> have to manually set these, the memory may not stay at
> CAS2.5 when the FSB is lowered if board is set to Auto or
> SPD mode). I don't if this will prove anything but it would
> eliminate physically failure in the modules and if the board
> is stable underclocked it might provide some aux. use
> instead of being thrown away (or is it under warranty)?
> Also if you were using Turbo (or whatever it's called) bios
> setting for memory, use the more conservative setting.

I already tried severe underclocking; it didn't change anything.

> Also try clearing CMOS, nForce2 boards can sometimes have
> wierd bios glitches... if you're running an early bios you
> might also try flashing a newer bios but only if you can
> regain stablity in memtest86, first. Do not flash the bios
> when system is in a state where memory errors are likely.

How do I clear CMOS? Is there some battery I need to remove and
replace?

> >And just to try to head off a question or two, I checked my BIOS
> >settings and the memory timing is set to 333Mhz.
>
> What make/model power supply?

I don't remember and I'm not seeing any obvious sticker on the PSU
without having to take the whole thing out, but it's black, has three
green glowing fans in it, it's big, it cost a lot, I remember
researching it and it was a good brand, and it's something like 550W.

> Does CPU use 166MHz FSB as memory does?

CPU is an Athlon Barton 2500+ so I think the answer is yes.

> Examine board for failed capacitors. A7N8X used "some"
> fairly good caps, but I don't remember on the "-X" version,
> I vaguely recall it had some component corners cut but I
> can't recall for certain which parts). Some of the caps for
> the memory are pretty far over towards the right-side edge
> of the board, but check the rest too.

Nothing obvious.

> Also check the CPU/heatsink thermal interface. If you had
> used regular silicone based thermal grease it may have dryed
> out over time... had that happen to a system here, so now I
> only use synthetic based grease on flipchips. It might be
> unrelated but you made no mention of checking temps (or
> voltages for that matter) yet. Also check your memory
> voltage settings in the bios.

Nothing out of the ordinary.

*Sigh* ... it looks like I'm going to have to buy a new mobo? I'll
test the memory in another system tomorrow but I don't think that's
going to be the problem.

I've been looking to do a system upgrade anyway so maybe this is just
serendipity disguised as misfortune. What I'm looking for is a Socket
939 AMD64 processor, and a mobo to go along with it. At some point in
the future I'd like to get a nice PCI Express graphics card but in the
meantime I'd like to hang onto my Radeon 9800 Pro if I can; do they
even make mobos with both PCI-X and AGP? Also, my memory (assuming
it's not borked) should transfer over to the new computer, right?

In article <1129457580.226499.204020@f14g2000cwb.googlegroups .com>,
Cyde Weys says...

> They occur at every address possible on the memory chips. Everywhere
> from 512MB down on to 0MB.
>
One thing everyone has overlooked is the cache on the CPU.


--
Conor

"You're not married, you haven't got a girlfriend and you've never seen
Star Trek? Good Lord!" - Patrick Stewart, Extras.

On 16 Oct 2005 03:13:00 -0700, "Cyde Weys"
<cydeweys@gmail.com> wrote:


>> OK, but, do these errors ALWAYS occur at same address(es)?
>
>They occur at every address possible on the memory chips. Everywhere
>from 512MB down on to 0MB.

Hmm. I would wonder if there is a bios corruption or
memtest86 bug, and would try a newer version of memtest86+,
if you hadn't already. That doesn't necessarily explain the
crashing, but if the board is headed for the trash anyway
you have free reign to make last ditch efforts without worry
of failure. Maybe try stabbing the board with a
screwdriver? ;-)



>I already tried severe underclocking; it didn't change anything.

I'd also examine the memory contacts, and the slot contacts.


>
>> Also try clearing CMOS, nForce2 boards can sometimes have
>> wierd bios glitches... if you're running an early bios you
>> might also try flashing a newer bios but only if you can
>> regain stablity in memtest86, first. Do not flash the bios
>> when system is in a state where memory errors are likely.
>
>How do I clear CMOS? Is there some battery I need to remove and
>replace?

Pull AC (or switch off) from power supply. CMOS jumper is
immediately to the right of the battery (at least on Dlx
version, assume it's same on -X version). Move jumper from
left & middle pin to midde and right pin, then return to
leftmost two pins, plug ac back in. You could pull battery
instead, wait a few minutes then put it back in. You might
check battery voltage too, just in case... but I doubt
that's a problem since the error state is reproducable and
other bios settings aren't lost.



>> What make/model power supply?
>
>I don't remember and I'm not seeing any obvious sticker on the PSU
>without having to take the whole thing out, but it's black, has three
>green glowing fans in it, it's big, it cost a lot, I remember
>researching it and it was a good brand, and it's something like 550W.

Hmm.
Then it probably is good, though these days I always take
web reviews with a grain of salt, even questionable supplies
like "Ultra" get good reviews but arent' near as good as
most name-brand units.


>
>> Does CPU use 166MHz FSB as memory does?
>
>CPU is an Athlon Barton 2500+ so I think the answer is yes.

I'd pull the 'sink and check the thermal interface just for
the heck of it.


>>Also check your memory
>> voltage settings in the bios.
>
>Nothing out of the ordinary.

You might also try upping voltage to 2.6 or 2.7, though if
there were the errors even when underclocked, AND memory bus
was locked to 100% and fixed manual timings in bios, so it
too was underclocked, it would seem unlikely to help.

>
>*Sigh* ... it looks like I'm going to have to buy a new mobo? I'll
>test the memory in another system tomorrow but I don't think that's
>going to be the problem.

Well at least they're cheaper than they used to be. You
might keep your eye open for a refurb from newegg or
chiefvalue to reuse the CPU.


>
>I've been looking to do a system upgrade anyway so maybe this is just
>serendipity disguised as misfortune. What I'm looking for is a Socket
>939 AMD64 processor, and a mobo to go along with it. At some point in
>the future I'd like to get a nice PCI Express graphics card but in the
>meantime I'd like to hang onto my Radeon 9800 Pro if I can; do they
>even make mobos with both PCI-X and AGP? Also, my memory (assuming
>it's not borked) should transfer over to the new computer, right?

PC3200 has a fair chance of being reused, unless it just
happened to be instable on the new board. Most boards don't
have both PCI Express and an AGP slot, and having both means
you're using up board real-estate to implement it so there
would be a loss of one or more regular PCI slots, but, yes
there's a ULI chipset that supports it.
http://www.anandtech.com/mb/showdoc.aspx?i=2524

My guess is that it is the motherboard. I have two Crucial 512MB DDR400
RAM modules and they've never given me any problems for 2-3 years now.

Any chance you can test your components on a separate mobo?

Cyde Weys wrote:
> So my computer's running fine, and BAM, it suddenly becomes
> nonfunctional. It repeatedly insta-reboots from inside Windows.
> Windows will start successfully, but try using it for a minute or two
> and the insta-reboot occurs. I figured maybe it's a problem with
> Windows, so I tried my Gentoo Linux LiveCD ... not working either. The
> computer boots fine and even sometimes gets the kernel booted, but it
> quickly freezes after that.
>
> So I'm thinking it may be a hardware problem. I try memtest86 from
> within the Gentoo Linux LiveCD. I get millions of errors. At this
> point I should probably mention that my memory consists of 2X Crucial
> 512MB DDR 333Mhz CL2.5 and my motherboard is an Asus A7N8X-X. So I
> think one of my memory modules may be bad, so I try running with just
> one. I get the errors in memtest. I try running it with just the
> other memory module installed. I get the same errors in memtest:
> specifically, the Err-Bits seem always to be 00ff0000.
>
> Is it possible for two memory modules to both go bad simultaneously and
> in the same fashion? I think it may be *possible*, but it's not
> exactly probable. So what could be the problem then? Is my
> motherboard borked?
>
> And just to try to head off a question or two, I checked my BIOS
> settings and the memory timing is set to 333Mhz.

None of your suggested diagnostics have helped me get the computer
working :-( At this point it looks like I'll be upgrading a bit ahead
of schedule. Here's the tentative parts list (from NewEgg):

EPoX EP-9NPAJ Socket 939 NVIDIA nForce4-4X ATX AMD Motherboard - Retail
Model #: EP-9NPAJ $88.99

MSI NX6800-TD128E Geforce 6800 128MB DDR PCI Express x16 Video Card -
Retail
Model #: NX6800-TD128E $194.00

AMD Athlon 64 3500+ Venice 1GHz FSB Socket 939 Processor Model
ADA3500BPBOX - Retail
Model #: ADA3500BPBOX $219.00

I'm just looking for suggestions, like if I can get something else with
the same performance for a lower price, or if an upgrade that doesn't
cost a lot is worth it, etc.

Also, I notice that the mobo says it uses DDR Standard DDR 400 (PC
3200). The memory I currently have is DDR 333. Will it still work in
the mobo or will I have to get new memory too?

Ugh, this is all looking very expensive.

Also, if anyone has any suggestions on overclocking options, I'm
receptive to those. Can I safely overclock the CPU and GPU? What
aftermarket coolers would you recommend? Thanks!

Cyde Weys wrote:
> David Maynard wrote:
>
>
>>It could be anything, including the CPU. After all, memory can't be read
>>without the CPU being involved.
>
>
> Is it even possible for a CPU to malfunction in such a specific way
> while still letting everything else work? It's not having problems
> whatsoever booting and running memtest86 ... it's just when you try to
> load an OS (which needs memory) that you get the problems.

I don't understand your "not having problems" running memtest description
when you then list the problems memtest reports.

Based on the description below, though, the processor wouldn't be high on
my guess list as processor problems are usually more random, if it runs, or
catastrophic preventing it from running (for very long).


>>But with persistent global errors of that type I'd tend to guess a BIOS
>>setting before hardware failure. Try 'safe' defaults and, in particular,
>>the slowest memory speed settings.
>
>
> It doesn't seem to matter. I get the same errors. I paid attention
> more time and it's more like 00XX0000 the first few seconds of
> memtest86 where XX are random hex, but then it settles down to Err-Bits
> 00ff0000 over and over and over.

That sounds like a bus problem, such as weak driver, bad bus terminator,
dirty/bent contacts, warped memory connector, etc., on those specific bits
or some kind of contamination on the motherboard affecting the traces.
Could even be cracked traces from over-flexing (although I don't know what
would have caused the flexing). Flexing can also break the solder joints
under the BGAs as can overheating, like say the northbridge.

I'd concentrate first on a thorough cleaning of the motherboard, the memory
connectors and the memory module contacts mainly because it's doable and
there's not much you can do about cracked traces or the BGA solder.

On 16 Oct 2005 20:18:12 -0700, "Cyde Weys"
<cydeweys@gmail.com> wrote:

>None of your suggested diagnostics have helped me get the computer
>working :-( At this point it looks like I'll be upgrading a bit ahead
>of schedule. Here's the tentative parts list (from NewEgg):
>
>EPoX EP-9NPAJ Socket 939 NVIDIA nForce4-4X ATX AMD Motherboard - Retail
>Model #: EP-9NPAJ $88.99
>
>MSI NX6800-TD128E Geforce 6800 128MB DDR PCI Express x16 Video Card -
>Retail
>Model #: NX6800-TD128E $194.00
>
>AMD Athlon 64 3500+ Venice 1GHz FSB Socket 939 Processor Model
>ADA3500BPBOX - Retail
>Model #: ADA3500BPBOX $219.00
>
>I'm just looking for suggestions, like if I can get something else with
>the same performance for a lower price, or if an upgrade that doesn't
>cost a lot is worth it, etc.

The price above on a regular GF 6800, 128MB, looks a bit
high but maybe I'm only contrasting to sale prices.
Regardless, I think you can get near same performance for
much less $ in a 6600GT, or if you look out for deals you
might be able to get a 256MB version for around $200, though
I don't know if it would o'c as well.

Never been a fan of Epox boards, but never tried that one
either. Harder to know if it's reasonable except in
retrospect.


>
>Also, I notice that the mobo says it uses DDR Standard DDR 400 (PC
>3200). The memory I currently have is DDR 333. Will it still work in
>the mobo or will I have to get new memory too?

For best performance, yes you will need new memory. Most
boards support async, slower memory bus but buying all new
parts just to underclock the memory bus for older memory
seems a bit limiting.

>
>Ugh, this is all looking very expensive.

Yep, you could always get another skt A board now and buy
the other parts a piece at a time, preferribly the
motherboard last so you had all other parts available to be
sure it isn't problematic- and motherboard is usually most
likely defective or unworkable in some way, IF anything
were.


>
>Also, if anyone has any suggestions on overclocking options, I'm
>receptive to those. Can I safely overclock the CPU and GPU? What
>aftermarket coolers would you recommend? Thanks!

"Safely"? Hard to classify really, but yes if you test
stability and monitor temps, including those of the
motherboard, it can be safe. No guarantee though, it's all
on you to make sure it stays within acceptible parameters.

Zalman and Thermalright make some nice 'sinks, get one that
takes a super-large quiet fan.

kony wrote:
> On 16 Oct 2005 20:18:12 -0700, "Cyde Weys" wrote:
> >MSI NX6800-TD128E Geforce 6800 128MB DDR PCI Express x16 Video Card -
> >Retail
> >Model #: NX6800-TD128E $194.00

> The price above on a regular GF 6800, 128MB, looks a bit
> high but maybe I'm only contrasting to sale prices.

Yeah, I forgot to mention that there is a $30 off rebate.

> Regardless, I think you can get near same performance for
> much less $ in a 6600GT, or if you look out for deals you
> might be able to get a 256MB version for around $200, though
> I don't know if it would o'c as well.

Out of all 27 6600GT cards on NewEgg, only one has 256MB RAM, and it's
not even in stock (and even if it were, it's $200). Is 256MB RAM a lot
better than 128MB? Would it be worth spending the extra money to get a
256MB version of a 6800? Are the 6600GT and 6800 vanilla comparable in
terms of performance?

> Never been a fan of Epox boards, but never tried that one
> either. Harder to know if it's reasonable except in
> retrospect.

Alright, I'll look for someone different. Any brands you tend to
prefer?

> >
> >Also, I notice that the mobo says it uses DDR Standard DDR 400 (PC
> >3200). The memory I currently have is DDR 333. Will it still work in
> >the mobo or will I have to get new memory too?
>
> For best performance, yes you will need new memory. Most
> boards support async, slower memory bus but buying all new
> parts just to underclock the memory bus for older memory
> seems a bit limiting.

In terms of memory I figure I want 2GB. I tend to not stress the CPU
so much in every day to day usage but I stress the CRAP out of memory.
You know, having like twenty things open. Corsair 512MB DDR 400 memory
is only $44 so I could get four or those and stick 'em in a board with
4 slots. Or would it be better to get 2 sticks of 1GB, which is
slightly more expensive?

> >Ugh, this is all looking very expensive.
>
> Yep, you could always get another skt A board now and buy
> the other parts a piece at a time, preferribly the
> motherboard last so you had all other parts available to be
> sure it isn't problematic- and motherboard is usually most
> likely defective or unworkable in some way, IF anything
> were.

But I'd have to get another Socket A board that supports my current AGP
video card. To be able to upgrade either the CPU or GPU would require
getting another mobo anyway. About the only thing I could upgrade with
another Socket A mobo is the memory (up to DDR 400). This is, unless,
I choose one of those dual AGP/PCI-X boards, but those look kind of
iffy.

I really can't decide whether to spend a lot of money to upgrade my
overall system or just replace the mobo and go on for awhile with what
I have.

Cyde Weys wrote:
> kony wrote:
> > Cyde Weys wrote:
> > >Ugh, this is all looking very expensive.
> >
> > Yep, you could always get another skt A board now and buy
> > the other parts a piece at a time, preferribly the
> > motherboard last so you had all other parts available to be
> > sure it isn't problematic- and motherboard is usually most
> > likely defective or unworkable in some way, IF anything
> > were.
>
> But I'd have to get another Socket A board that supports my current AGP
> video card. To be able to upgrade either the CPU or GPU would require
> getting another mobo anyway. About the only thing I could upgrade with
> another Socket A mobo is the memory (up to DDR 400). This is, unless,
> I choose one of those dual AGP/PCI-X boards, but those look kind of
> iffy.
>
> I really can't decide whether to spend a lot of money to upgrade my
> overall system or just replace the mobo and go on for awhile with what
> I have.

I'm now looking at a differential in price between upgrading and
replacing that is greater than one order of magnitude. It just doesn't
make sense to upgrade. I found the following mobo at $36:

Foxconn 600A01-6LRS Socket A (Socket 462) VIA KT600 ATX AMD Motherboard
- Retail

* Model #: 600A01-6LRS
* Item #: N82E16813186025

http://www.newegg.com/Product/Produc...82E16813186025

It has everything I need ... 5 PCI slots, DDR 400/333 FSB, and AGP 8X
slot. Unless there's some other mobo that has better features than
this one for not much more of a price.

Thanks for the help, everyone.

On 17 Oct 2005 02:07:32 -0700, "Cyde Weys"
<cydeweys@gmail.com> wrote:


>Out of all 27 6600GT cards on NewEgg, only one has 256MB RAM, and it's
>not even in stock (and even if it were, it's $200). Is 256MB RAM a lot
>better than 128MB? Would it be worth spending the extra money to get a
>256MB version of a 6800? Are the 6600GT and 6800 vanilla comparable in
>terms of performance?

Supposedly the 6800 will be marginally faster, but mainly
having it at the lower stock GPU speed means *on average* it
should have more o'c potential.

Most games don't "need" over 128MB memory unless details are
set to maximum. Future games are bound to use more though,
if it were my purchase and I already had a Radeon 9800 Pro,
I'd go with 256MB on any >$150 card, IF gaming performance
is among the most important indexes for the new system.



>
>> Never been a fan of Epox boards, but never tried that one
>> either. Harder to know if it's reasonable except in
>> retrospect.
>
>Alright, I'll look for someone different. Any brands you tend to
>prefer?

Asus, Abit, sometimes MSI, Gigabyte on their mid-to-high end
boards. I dont' know how long you need it to last or
feature support though... I try to avoid recommending
specific boards as there are too many variables involved in
what's best for any specific person's needs. You can
probably get a lot of info on some in the motherboard
subsections of forums like http://forums.amdmb.com



>> For best performance, yes you will need new memory. Most
>> boards support async, slower memory bus but buying all new
>> parts just to underclock the memory bus for older memory
>> seems a bit limiting.
>
>In terms of memory I figure I want 2GB. I tend to not stress the CPU
>so much in every day to day usage but I stress the CRAP out of memory.
>You know, having like twenty things open. Corsair 512MB DDR 400 memory
>is only $44 so I could get four or those and stick 'em in a board with
>4 slots. Or would it be better to get 2 sticks of 1GB, which is
>slightly more expensive?

Yes it would be better to get 1GB per module when expecting
to use 2GB total. Get them both at same time and test
together to be sure they're stable.

>
>> >Ugh, this is all looking very expensive.
>>
>> Yep, you could always get another skt A board now and buy
>> the other parts a piece at a time, preferribly the
>> motherboard last so you had all other parts available to be
>> sure it isn't problematic- and motherboard is usually most
>> likely defective or unworkable in some way, IF anything
>> were.
>
>But I'd have to get another Socket A board that supports my current AGP
>video card. To be able to upgrade either the CPU or GPU would require
>getting another mobo anyway. About the only thing I could upgrade with
>another Socket A mobo is the memory (up to DDR 400). This is, unless,
>I choose one of those dual AGP/PCI-X boards, but those look kind of
>iffy.

My point was not to try to max-out skt A, rather to be able
to reuse the parts you already had- CPU, memory, video. If
the resulting system isn't enough performance, then sure,
you'll have to build another one too... but getting that one
running again for around $30-40 (typical price of some of
the refurb'd boards) is fairly cost-effective. I didn't
even consider it but does your board have a warranty in
effect?

>I really can't decide whether to spend a lot of money to upgrade my
>overall system or just replace the mobo and go on for awhile with what
>I have.

Well, for the cost of a refurb'd board you can get by for
awhile, and even if you wanted to do the whole system
upgrade in a couple months, given the $ depreciation on the
newer parts and more time to hunt down deals, it could end
up being same or lower total cost... just depends on how
much of a need there is at this very moment for higher
performanace, since buying for the future is never a good
investment.

On 17 Oct 2005 02:16:10 -0700, "Cyde Weys"
<cydeweys@gmail.com> wrote:


>I'm now looking at a differential in price between upgrading and
>replacing that is greater than one order of magnitude. It just doesn't
>make sense to upgrade. I found the following mobo at $36:
>
>Foxconn 600A01-6LRS Socket A (Socket 462) VIA KT600 ATX AMD Motherboard
>- Retail
>
> * Model #: 600A01-6LRS
> * Item #: N82E16813186025
>
>http://www.newegg.com/Product/Produc...82E16813186025
>
>It has everything I need ... 5 PCI slots, DDR 400/333 FSB, and AGP 8X
>slot. Unless there's some other mobo that has better features than
>this one for not much more of a price.
>
>Thanks for the help, everyone.

I would go for nForce2 instead of KT600. I have both here,
the KT600 may test out better in synthetic benches than it
ends up doing in real world uses. I"m not too fond of the
OSI capacitors it uses (blue and gold ones) either, as I've
had to replace a few of those- not necessarily defective,
just lower quality, higher ESR caps that don't like running
hot or leave as much margin when overclocking. That board
also has an incredibly cheap VRM subcircuit, it's rare to
see 2 phase in TO252 'Fets.

Basically I"m suggesting that if you only want to spend
~$40, better to go with a refurb'd board. "Maybe" it is a
refurb but it's hard to tell because the part number doesn't
have "R" on the end like any others do. I think it's just a
dirt-cheaply manufactured board. One of the issues might be
whether you need integrated SATA. Most of the boards at the
'egg can handle DDR333 & AGP8X and have 5 PCI slots so long
as they're not mATX.

I'd probably get the MSI K7N2 Delta2-FSR @ $45,

http://www.newegg.com/Product/Produc...2E16813130512R
but is it right for you? I can only speak for what I'd buy
for myself in this situation. It's northbridge fan may need
throttled back or a passive 'sink installed, and the CPU
socket doesn't have the 4 mounting holes around it for
through-board types of heatsink mounts. In the picture it
also looks like they left a few pin-headers off the board
but I"m too lazy ATM to find out if it's important or not.
Integrated GbE is nice though, especially if you wanted to
turn it into a fileserver some day.

Alright guys I've sorted out my troubles. I was able to pretty much
rule out everything except the mobo as not working, so I bought the
mobo kony suggested, I just installed it, and I'm now running memtest86
and not getting the errors.

Windows isn't booting, mind you, but that's a Windows driver issue, not
a hardware issue.

Grrrr, stupid Windows. It should be able to hot-swap mobos.

Cyde Weys wrote:
> Alright guys I've sorted out my troubles. I was able to pretty much
> rule out everything except the mobo as not working, so I bought the
> mobo kony suggested, I just installed it, and I'm now running memtest86
> and not getting the errors.
>
> Windows isn't booting, mind you, but that's a Windows driver issue, not
> a hardware issue.

I think I ran into this issue before, and I ended up just reinstalling
Windows. I'm not sure if there even is an alternate solution. I'm
downloading Windows XP now (I always lose CDs, dammit), and hopefully
I'll have my system back up and running at the end of the night.

Cyde Weys wrote:
> Cyde Weys wrote:
>
>>Alright guys I've sorted out my troubles. I was able to pretty much
>>rule out everything except the mobo as not working, so I bought the
>>mobo kony suggested, I just installed it, and I'm now running memtest86
>>and not getting the errors.
>>
>>Windows isn't booting, mind you, but that's a Windows driver issue, not
>>a hardware issue.
>
>
> I think I ran into this issue before, and I ended up just reinstalling
> Windows. I'm not sure if there even is an alternate solution. I'm
> downloading Windows XP now (I always lose CDs, dammit), and hopefully
> I'll have my system back up and running at the end of the night.
>

Do a repair install. Not the first repair option but continue on as if
doing a fresh install but select repair the existing installation. You'll
need to reinstall all service packs before all applications will work
properly because they'll expect the service pack files to be there.

Cyde Weys wrote:

> Alright guys I've sorted out my troubles. I was able to pretty much
> rule out everything except the mobo as not working, so I bought the
> mobo kony suggested, I just installed it, and I'm now running memtest86
> and not getting the errors.
>
> Windows isn't booting, mind you, but that's a Windows driver issue, not
> a hardware issue.
>
> Grrrr, stupid Windows. It should be able to hot-swap mobos.
>

Why should it? A motherboard is not a 'mobile' item and, in fact,
considered essentially 'the computer' (with a processor and memory).
Everything else is an 'add-on' to it.

Btw, "hot swap" means while power is on and I'd love to see how you did
that with a motherboard ;)

On Sat, 22 Oct 2005 18:36:59 -0500, David Maynard
<nospam@private.net> wrote:

>Cyde Weys wrote:
>
>> Alright guys I've sorted out my troubles. I was able to pretty much
>> rule out everything except the mobo as not working, so I bought the
>> mobo kony suggested, I just installed it, and I'm now running memtest86
>> and not getting the errors.
>>
>> Windows isn't booting, mind you, but that's a Windows driver issue, not
>> a hardware issue.
>>
>> Grrrr, stupid Windows. It should be able to hot-swap mobos.
>>
>
>Why should it? A motherboard is not a 'mobile' item and, in fact,
>considered essentially 'the computer' (with a processor and memory).
>Everything else is an 'add-on' to it.

Because it's a desirable feature, to many even if you don't
care either way. Because what they "have" to allow is the
opposite of what they would allow, in a free market to
remain competitive.

Full retail licenses are not tied to "the computer", only
one system which would obviously not be the case here. They
would have the feature because without it the customer will
spend a minimum of 1/2 hour if not more (actually a LOT more
time if a person is prudent and makes system backups
inbetween the windows updates as they should've previously).
Their choice to cost customers time is clearly a disregard
for them.

David Maynard wrote:
> Cyde Weys wrote:
> > Grrrr, stupid Windows. It should be able to hot-swap mobos.
>
> Btw, "hot swap" means while power is on and I'd love to see how you did
> that with a motherboard ;)

I'm well aware of what hot swapping means, and it should easily be
possible if you can swap the CPU between the mobos within one clock
cycle.

kony wrote:

> On Sat, 22 Oct 2005 18:36:59 -0500, David Maynard
> <nospam@private.net> wrote:
>
>
>>Cyde Weys wrote:
>>
>>
>>>Alright guys I've sorted out my troubles. I was able to pretty much
>>>rule out everything except the mobo as not working, so I bought the
>>>mobo kony suggested, I just installed it, and I'm now running memtest86
>>>and not getting the errors.
>>>
>>>Windows isn't booting, mind you, but that's a Windows driver issue, not
>>>a hardware issue.
>>>
>>>Grrrr, stupid Windows. It should be able to hot-swap mobos.
>>>
>>
>>Why should it? A motherboard is not a 'mobile' item and, in fact,
>>considered essentially 'the computer' (with a processor and memory).
>>Everything else is an 'add-on' to it.
>
>
> Because it's a desirable feature, to many even if you don't
> care either way.

Not to 'many'. To a comparatively small few.

> Because what they "have" to allow is the
> opposite of what they would allow, in a free market to
> remain competitive.

There is no 'market incentive' to provide features, at significant cost,
that only a handful of people would even care about. And, just as it might
be 'convenient' to a hand full of car 'enthusiasts' if automobiles came
with a sufficiently sized engine compartment, 'self configuring' electrical
systems, 'auto-fit' exhaust piping, 'universal' cooling, 'infinitely
variable' transmission ratios, and built in mounts to hold every motor made
so they can 'hot swap' a 'vette 454 for a Ford 289 in 10 minutes, it's
ridiculously cost prohibitive.


> Full retail licenses are not tied to "the computer", only
> one system which would obviously not be the case here.

I didn't say a thing about licenses.

> They
> would have the feature because without it the customer will
> spend a minimum of 1/2 hour if not more (actually a LOT more
> time if a person is prudent and makes system backups
> inbetween the windows updates as they should've previously).
> Their choice to cost customers time is clearly a disregard
> for them.

That is simply nonsense. It isn't done for the reason I just stated. It
would be an incredibly complex 'feature', costing significant time and
money to implement (if even possible), for a hand full of 'enthusiasts' who
exchange motherboards, or the rare failure, and all to simply save an hour
or so of time for those few. Not to mention it's doubtful it would even
save time because the process of redetecting and installing everything
would still need to take place.

But then you'd bitch incessantly about the "software bloat" needed to
support a 'feature' that perhaps 1 out of a thousand use once every 10 years.

Actually, the capability exists. It's called a 'repair install'.

Cyde Weys wrote:

> David Maynard wrote:
>
>>Cyde Weys wrote:
>>
>>>Grrrr, stupid Windows. It should be able to hot-swap mobos.
>>
>>Btw, "hot swap" means while power is on and I'd love to see how you did
>>that with a motherboard ;)
>
>
> I'm well aware of what hot swapping means, and it should easily be
> possible if you can swap the CPU between the mobos within one clock
> cycle.
>

Oh I can easily swap the CPU in one clock cycle, it's getting it to
remember which cycle it was in when it powers back up that has me stumped.

David Maynard wrote:
> Cyde Weys wrote:
> > David Maynard wrote:
> >>Cyde Weys wrote:
> >>>Grrrr, stupid Windows. It should be able to hot-swap mobos.
> >>
> >>Btw, "hot swap" means while power is on and I'd love to see how you did
> >>that with a motherboard ;)
> >
> > I'm well aware of what hot swapping means, and it should easily be
> > possible if you can swap the CPU between the mobos within one clock
> > cycle.
>
> Oh I can easily swap the CPU in one clock cycle, it's getting it to
> remember which cycle it was in when it powers back up that has me stumped.

But then it's not hot-swapping if you end up powering down the computer
and then bringing it back up. :-P

On Sun, 23 Oct 2005 17:01:34 -0500, David Maynard
<nospam@private.net> wrote:

>> Because it's a desirable feature, to many even if you don't
>> care either way.
>
>Not to 'many'. To a comparatively small few.

You might be surprised just how many people want the
performance boost that comes with newer hardware OR that
have a system failure and either buy new box or new board
and do not want to reinstall windows, customize it, or
install all their software 'n such all over again. In fact,
I suspect most people would find the ability to do that,
useful... far more useful than some of the so-called
features in XP.


>
>> Because what they "have" to allow is the
>> opposite of what they would allow, in a free market to
>> remain competitive.
>
>There is no 'market incentive' to provide features, at significant cost,

Well here we go again... I doubt it is a significant cost.
I suspect they know how and could do it right now if they
chose, and may have even tested it. They DID have to have
the basic functionality working in the first place just to
install the OS, too.

There is a market incentive. It is a very desirable
feature. They just didn't need to do it because they're not
competing with anyone else in the retail PC biz. Nobody
claimed that "all" users "need" it, but such is the case
with any marketable feature of (take your pick- any
product).


>that only a handful of people would even care about. And, just as it might
>be 'convenient' to a hand full of car 'enthusiasts' if automobiles came
>with a sufficiently sized engine compartment,

handful of enthusiasts? I take it you either don't remember
how roomy cars used to be under the hood, or don't do any
repairs yourself anymore. Regardless, with that space
reduction there was gain, a size and aerodynamics benefit.

>'self configuring' electrical
>systems, 'auto-fit' exhaust piping, 'universal' cooling, 'infinitely
>variable' transmission ratios, and built in mounts to hold every motor made
>so they can 'hot swap' a 'vette 454 for a Ford 289 in 10 minutes, it's
>ridiculously cost prohibitive.

.... if you choose to compare only ridiculously cost
prohibitive things, sure, but that is not the case with XP
being able to PNP a drive controller, for example, if/when
it can't find the OS boot drive.


>> They
>> would have the feature because without it the customer will
>> spend a minimum of 1/2 hour if not more (actually a LOT