Should RAM timings have to be set manually?

[Rod Speed]

That's just it, I can, and do on most memory

Irrelevant to whether the kingston value ram is BAD RAM.

and it is fine running at the SPD values. The
exception is memory with less conservative ratings.

Doesnt mean its BAD RAM.

 

[larry moe 'n curly]

larry moe 'n curly <[email protected]> wrote

Not all bios can setup the chipset properly with some ram.


Corse it does, they have different detail in the spd that the bios can handle.


I dont believe the spd data was identical and if it was, all
that proves is that the bios setup timing details that suited the
Fxxxxxx chips better, NOT that the unmarked chips are BAD.

The SPD contents were the same, except for things like date of
manufacture and checksum.

How do you know it proves that the BIOS setup timings were unsuitable
rather than that the modules were bad, especially when some modules
wouldn't work with even the slowest setup timings?

 

[Rod Speed]

Rod Speed wrote

The SPD contents were the same, except for
things like date of manufacture and checksum.

How do you know it proves that the BIOS setup timings
were unsuitable rather than that the modules were bad,

I just dont believe that Kingston would ship ram that produces
that obscene result you got in all systems that ram is used in.

especially when some modules wouldn't
work with even the slowest setup timings?

See above. Bet you've just made that last one up now that your nose has
been rubbed in the fact that you dont have any evidence that its BAD RAM.

 

[larry moe 'n curly]

larry moe 'n curly <[email protected]> wrote

I just dont believe that Kingston would ship ram that produces
that obscene result you got in all systems that ram is used in.

Not necessarily with all of them but with at least one of them.

See above. Bet you've just made that last one up now that your nose has
been rubbed in the fact that you dont have any evidence that its BAD RAM.

You lose your bet. My evidence is that those modules failed either
Gold Memory or MemTest86 at even the slowest settings and that the
results were repeatable. OTOH you have no evidence that the BIOS setup
SPD timings were incompatible with the modules.

 

[Rod Speed]

Rod Speed wrote

Not necessarily with all of them but with at least one of them.

Then its likely that its the bios in that one that is the problem.

You lose your bet.

Nope, you would have said that before now it it had been true.

My evidence is that those modules failed either Gold Memory or MemTest86
at even the slowest settings and that the results were repeatable.

More evidence of something seriously ****ed in the bios
of that system if that result was only seen in that system.

OTOH you have no evidence that the BIOS setup
SPD timings were incompatible with the modules.

Wrong again, its there above.

 

[larry moe 'n curly]

I just dont believe that Kingston would ship ram that produces
that obscene result you got in all systems that ram is used in.


Then its likely that its the bios in that one that is the problem.


See above. Bet you've just made that last one up now that your nose has
been rubbed in the fact that you dont have any evidence that its BAD RAM.


Nope, you would have said that before now it it had been true.

I had mentioned it before, about PC2100 modules that showed errors at
266 MHz, even with the slowest timings, but not at 200 MHz.

More evidence of something seriously ****ed in the bios
of that system if that result was only seen in that system.

Errors disappear when an identical model DIMM is substituted, so by
your reasoning the BIOS is bad, not the first module.

I counted a module as bad if it failed in either system, one system
having a VIA KT400 chipset, the other an nVidia nForce 3. Almost
always, when a module failed with the KT400 it also failed with the
nForce.

Wrong again, its there above.

You also haven't explained why setting the timings manually to slower
than the SPD values didn't help in most cases. MemTest86 reported
errors at 3-3-3-8-1T (SPD defaults) and even 3-4-4-8-1T, but changing
1T to 2T usually cleared up the errors, except with the Taiwan
Kingstons with completely unmarked chips and both Mushkins (Spectek
chips, but IDed as Kingston. Mushkin said they didn't use Kingston,
but they also said they didn't use Spectek).

 

[Rod Speed]

Rod Speed wrote

I had mentioned it before, about PC2100 modules that showed errors
at 266 MHz, even with the slowest timings, but not at 200 MHz.

Thats nothing like the previous claim.

Errors disappear when an identical model DIMM is substituted,

Irrelevant, the marking on the ram chips wasnt identical.

so by your reasoning the BIOS is bad, not the first module.

The timing detail setup by the bios with the ram you claim is bad
is clearly not appropriate if the ram works fine in a different system.

I counted a module as bad if it failed in either system, one
system having a VIA KT400 chipset, the other an nVidia nForce 3.

More fool you. Thats not BAD RAM, thats just a bios that
doesnt use appropriate timing for that particular ram.

Almost always, when a module failed with
the KT400 it also failed with the nForce.

Irrelevant if it worked fine in other systems.

You also haven't explained why setting the timings manually
to slower than the SPD values didn't help in most cases.

What matters is if it works fine in some systems. If it does it isnt BAD RAM.

MemTest86 reported errors at 3-3-3-8-1T (SPD defaults) and even
3-4-4-8-1T, but changing 1T to 2T usually cleared up the errors,

So clearly the bios isnt doing much of a job with the spd data.

except with the Taiwan Kingstons with completely unmarked chips
and both Mushkins (Spectek chips, but IDed as Kingston. Mushkin
said they didn't use Kingston, but they also said they didn't use Spectek).

Irrelevant to what is being discussed, whether its actually BAD RAM.

Clearly if it works fine in ANY system, it isnt BAD RAM.

 

[larry moe 'n curly]

larry moe 'n curly <[email protected]> wrote

See above. Bet you've just made that last one up now that your nose has
been rubbed in the fact that you dont have any evidence that its BAD RAM.


Nope, you would have said that before now it it had been true.


Thats nothing like the previous claim.

Here's what I wrote on 11/21/2006:

It seems a lot like my previous claim.

Irrelevant, the marking on the ram chips wasnt identical.

I'm including modules with identical chip markings..

The timing detail setup by the bios with the ram you claim is bad
is clearly not appropriate if the ram works fine in a different system.

That doesn't explain why manually choosing slower settings didn't help.
BTW, the VIA KT400 mobo has a BIOS setting for relaxed memory
timings. I don't know what it changes, but using it didn't change the
results.

Irrelevant if it worked fine in other systems.

Wrong again, its there above.

You haven't presented any evidence, but I have tried settings that were
slower than the defaults..

What matters is if it works fine in some systems. If it does it isnt BAD RAM.

So if 9 out of 10 patients of a certain surgeon die, that surgeon is OK
because of the other patient?

So clearly the bios isnt doing much of a job with the spd data.

Why should 1T memory have to be run at 2T to be reliable?

 

[Jon Danniken]

I wanted to post this followup to my original post.

I received the Crucial Ballistix on Monday, and they perform at 1T
flawlessly (24 hrs memtest), and at tighter timings (2-2-2-8-1T @ 200MHz)
than the HyperX specs to boot.

The conclusion is that the fault was with the Kingston HyperX, which failed
to run at their advertised speed (2-3-2-6-1T @ 200MHz), at least in this
board (EPOX 9NDA3I) and CPU (AMD 3700+ San Diego).

I do not have the time nor inclination to pay more shipping charges to test
out any more Kingston sticks for them.

The Crucial Ballistix do run very hot though, so I will be needing to put a
fan on them, but the Kingstons are going back to Newegg, and I will not be
purchasing any more memory from Kingston in the future.

Jon

 

[Rod Speed]

Rod Speed wrote

Here's what I wrote on 11/21/2006:

It seems a lot like my previous claim.

That wasnt the previous claim I was referring to.

I'm including modules with identical chip markings..

Thats not what you said previously.

That doesn't explain why manually choosing slower settings didn't help.

Yes it does, the bios presumably stuffed up with the manual settings too.

BTW, the VIA KT400 mobo has a BIOS setting for relaxed memory timings.
I don't know what it changes, but using it didn't change the results.

Irrelevant if that ram works fine in some systems. If it does, its clearly not BAD RAM.

You haven't presented any evidence,

Yes I have, the fact that I dont believe that Kingston
is shipping ram that does not work in any system.

but I have tried settings that were slower than the defaults..

In just a couple of systems.

So if 9 out of 10 patients of a certain surgeon die,
that surgeon is OK because of the other patient?

Different matter entirely.

It aint BAD RAM if it works fine in some systems.

Why should 1T memory have to be run at 2T to be reliable?

You dont even know what the bios is doing, just what its told to do.

 

[larry moe 'n curly]

larry moe 'n curly <[email protected]> wrote

Then its likely that its the bios in that one that is the problem.


See above. Bet you've just made that last one up now that your nose has
been rubbed in the fact that you dont have any evidence that its BAD RAM.


Nope, you would have said that before now it it had been true.





That wasn't the previous claim I was referring to.

Then what did I write?

Irrelevant, the marking on the ram chips wasnt identical.


Thats not what you said previously.

Why would I have argued about modules with non-identical chips when
your contention was that the chips were different?

The timing detail setup by the bios with the ram you claim is bad
is clearly not appropriate if the ram works fine in a different system.


Yes it does, the bios presumably stuffed up with the manual settings too.

How can you presume that?

A memory failure with just one system is enough to prove that the module is bad.

What matters is if it works fine in some systems. If it does it isnt BAD RAM.


Different matter entirely.

It's the same matter -- low quality standards.

So clearly the bios isnt doing much of a job with the spd data.


You dont even know what the bios is doing, just what its told to do.

Then why did most of the modules fail when BIOS was told to use 1T but
not when it was told to use 2T? Isn't it likely that the BIOS is
changing the 1T/2T setting?

 

[Rod Speed]

Rod Speed wrote

Then what did I write?

You keep hacking back the quoting, go back and check that for yourself.

Why would I have argued about modules with non-identical chips
when your contention was that the chips were different?

It wasnt clear whether you meant identical in the sense
of identical chips or just identical in the sense of the
module specs, particularly when you used the word model.

How can you presume that?

Because, like I said, I just dont believe that Kingston
would be shipping ram which doesnt work in any system.

I just dont believe that their quality control is that bad.

It's the same matter -- low quality standards.

Nope. You dont know that it is low quality standard with the ram.

ALL you know is that you got the result you got in a COUPLE OF SYSTEMS.

Then why did most of the modules fail when BIOS was
told to use 1T but not when it was told to use 2T?

You dont know what the bios chose to do when told to do that.

Isn't it likely that the BIOS is changing the 1T/2T setting?

Likely, but since I just dont believe that Kingston is shipping ram that
gets that obscene result in all systems that are speced for that ram,
its much more likely to be a bios problem and not BAD RAM at all.