A7V400-MX and Sempron 2400+ problems

  • Thread starter Thread starter Tegger®
  • Start date Start date
T

Tegger®

Greetings,

Newly purchased motherboard, RAM and CPU:
A7V400-MX motherboard
generic DDR400 PC-3200 RAM
AMD Sempron 2400+ 1.67GHz CPU (SDA2400DUT3D).


Originally, the BIOS reported:
RAM at DDR133, FSB at 100MHz and CPU at 1000MHz.
That was with BIOS 1001.

I upgraded to BIOS 1003 (the latest), then loaded the Setup Defaults and
rebooted. This resulted in the following:

RAM now at DDR266, FSB still at 100MHz, CPU still at 1000MHz

The program CPU-Z shows this as well, but reports the RAM at 133 even
though the BIOS reports it as 266.

There is only one dipswitch that controls the core frequency and
multiplier. Apparently the Sempron 2400+ is a locked processor, so settings
to the dipswitch are ignored. This is confirmed by my observation and
experimentation.

Now to the crux of the matter:
How do I get the BIOS to see the CPU at 1.67GHz?
How do I get the BIOS to see the RAM at 400MHz?


--
TeGGeR®

How to find anything on the Internet:
www.google.com

or in Usenet Groups:
www.groups.google.com

Google is your friend. Learn how to use it.
 
"Tegger®" said:
Greetings,

Newly purchased motherboard, RAM and CPU:
A7V400-MX motherboard
generic DDR400 PC-3200 RAM
AMD Sempron 2400+ 1.67GHz CPU (SDA2400DUT3D).


Originally, the BIOS reported:
RAM at DDR133, FSB at 100MHz and CPU at 1000MHz.
That was with BIOS 1001.

I upgraded to BIOS 1003 (the latest), then loaded the Setup Defaults and
rebooted. This resulted in the following:

RAM now at DDR266, FSB still at 100MHz, CPU still at 1000MHz

The program CPU-Z shows this as well, but reports the RAM at 133 even
though the BIOS reports it as 266.

There is only one dipswitch that controls the core frequency and
multiplier. Apparently the Sempron 2400+ is a locked processor, so settings
to the dipswitch are ignored. This is confirmed by my observation and
experimentation.

Now to the crux of the matter:
How do I get the BIOS to see the CPU at 1.67GHz?
How do I get the BIOS to see the RAM at 400MHz?
Click to expand...

Your processor is demonstrating 10X multiplier, and that seems
correct. There is a DIP switch called "DSW" in the manual, and
it should be set to 166MHz. 10x166=1.67GHz. The processor cannot
"lock" this clock setting, as the clock comes from the clockgen
chip. The multiplier may be locked, just like they were this
year on the AthlonXP.

When you set the DSW to 166MHZ, the BIOS should report the clock
is 166MHz, which is an FSB of 333 (i.e. data transferred twice
per clock, so is double the clock rate).

The BIOS has a setting for the DRAM clock, and once the CPU
clock is set to 166MHz, the auto setting of the DRAM clock should
be setting itself to 166MHz as well. This gives DDR333. While
you may not be impressed by this, synchronous operation
(FSB333 = DDR333) is frequently the best performing setting, as
the Northbridge can have a couple of cycles of sync logic removed
from the access path when it knows the two sides of the chip
are synchronous.

It doesn't look to me like the BIOS has an option to clock the
memory higher. I mean, even if you overclock the processor by
setting DSW to 200MHz, the BIOS setting for the DRAM clock is
likely to still only offer 166MHz as the top clock speed.

I would say your best performing, non-overclocked config,
is DSW=166MHz Core=1.67GHz FSB=333 DDR=333.

If you attempt to overclock (and it may not work - I don't know
what kind of overclocking range the new Semprons have), the
likely settings would be DSW=200MHz Core=2.0GHz FSB=400 DDR=333.
That is, unless there is some way to "trick" the board into
continuing to operate the FSB and memory interfaces synchronous.
This could be possible with a program like clockgen, setfsb,
softfsb, but those programs have to be written for each
clock generator that is used. The authors of those programs
don't have a lot of incentive to update their program for
the thousands of different board designs that exist, so unless
the clock generator chip just happens to be one used on another
Athlon board, I wouldn't depend on that route working out for
you.

HTH,
Paul
 
(e-mail address removed) (Paul) muttered darkly in

When you set the DSW to 166MHZ, the BIOS should report the clock
is 166MHz, which is an FSB of 333 (i.e. data transferred twice
per clock, so is double the clock rate).
Click to expand...



The DSW switch IS the dipswitch I referred to. Adjusting it according to
the manual has no effect: The clock is always reported as 100MHz.

I would say your best performing, non-overclocked config,
is DSW=166MHz Core=1.67GHz FSB=333 DDR=333.
Click to expand...



If I could to that, sure, but the BIOS does not respond to dipswitch
settings.

Oddly--and I wonder if this is related--I cannot change the system date or
time in the BIOS. No key or key combination will allow me to change the
values. Everything else can be changed just fine, but not the date and
time.

Maybe I have a defective board?

--
TeGGeR®

How to find anything on the Internet:
www.google.com

or in Usenet Groups:
www.groups.google.com

Google is your friend. Learn how to use it.
 
"Tegger®" said:
(e-mail address removed) (Paul) muttered darkly in


The DSW switch IS the dipswitch I referred to. Adjusting it according to
the manual has no effect: The clock is always reported as 100MHz.


If I could to that, sure, but the BIOS does not respond to dipswitch
settings.

Oddly--and I wonder if this is related--I cannot change the system date or
time in the BIOS. No key or key combination will allow me to change the
values. Everything else can be changed just fine, but not the date and
time.

Maybe I have a defective board?
Click to expand...

It could be a bad board.

A clock generator has two sets of inputs. It has some direct logic
signals, usually with names like FSB0, FSB1, FSB2. In this case, these
will be tied to the DSW switches. When the power first comes on,
the clock generator samples the signal levels on the FSB pins and
starts the clock running at that speed. That means, at T=0, the clock
is 166MHz.

After the BIOS has had a moment to run, it does an analysis of the
processor. There are tables stored in DMI/ESCD in the BIOS Flash chip,
and the BIOS records info about the processor in there. If the BIOS
recognizes that the processor is not the same type as it recorded in
DMI/ESCD the last time, it will examine the processor directly, and
record the new info in DMI/ESCD, by burning the info into the flash
chip.

It is my understanding, that if the BIOS doesn't understand the processor
type, it will use what it considers to be safe settings. This is where
a second input on the clock generator can be used. There should be
a I2C (I squared C) serial interface, which is tied to the SMBUS (System
Management Bus). This is the same bus that the SPD (serial presence
detect) device located on each DIMMs is connected. The BIOS can program
a new clock frequency via I2C. The 100MHz value sounds like a safe value,
or if the Sempron looks like some older processor, the BIOS could be
using the setup for the older processor type. The "name string" displayed
in the BIOS might give some clue as to what the BIOS thinks it is
dealing with.

Now, normally, this complicated bunch of functionality would have
another feature added to it. Which is, the ability to set the clock
frequency in the BIOS itself. That seems to be missing from this BIOS.
And, in such a situation, the BIOS should be able to read the value
of DSW, and leave the setting the way the user wants it via DSW.

It almost sounds like the ability to write settings is busted somehow.
Have you tried "Save Changes" followed by "Exit without saving",
instead of using "Save and Exit" ?

The time setting thing is slightly different, because that should be
written into the RTC (Real Time Clock), a chunk of circuitry in the
Southbridge. The same chunk of circuitry also contains the CMOS
RAM, which is powered by the coin cell battery on the motherboard,
whenever the PSU is switched off at the back or the computer is
unplugged. If your BIOS changes are still there, the next time you
power up, then the CMOS RAM must be OK. If the date and time are
not set, then the RTC could be forgetting its settings, or the
ability to write the clock is affected some how.

The first thing to look into, is the BIOS version. If you haven't
flashed the BIOS already, there should be a paper label on the
flash chip, with the version printed on it. The cpusupport web
page says Sempron support starts with BIOS version 1003. You may
want to visit the download page and pick up the 1003 BIOS, and
a copy of AWDflash, or even the 1003.002 beta BIOS. It is possible
at least some of your problems will change with a BIOS update.
And, since the board is brand new, if that doesn't help or change
the symptoms, then return the board to the vendor.

HTH,
Paul
 
(e-mail address removed) (Paul) muttered darkly in
It could be a bad board.

A clock generator has two sets of inputs. It has...
Click to expand...


<snip very informative technical info>


An update: The problems have been solved.

The trouble turned out to be a loose nut, that nut being me.

When I was reading the manual, I was assuming that the DSW switch drawings
were like the jumper drawings elsewhere in the book, where the dark part
represents the jumper itself. The DSW drawings actually use WHITE to
indicate the position of the switch. It makes sense once you think about
it, the jumpers being black and the DSW switches being white, but I was
thinking along different lines. I therefore had the switches backwards from
where they should have been.

The RAM is now seen as 333. It's still not 400, but close enough. Maybe a
later BIOS flash will allow 400. I'm already up to BIOS version 1003, but
do not want to try the beta BIOS.

As far as the date and time goes, to set that in the BIOS it seems you need
to use the + and - keys on the *numeric keypad*, not the ones on the main
part of the keyboard, something that never crossed my mind and was not
clarified in the manual.

Thanks for the help, Paul.


--
TeGGeR®

How to find anything on the Internet:
www.google.com

or in Usenet Groups:
www.groups.google.com

Google is your friend. Learn how to use it.
 
Tegger® said:
(e-mail address removed) (Paul) muttered darkly in


An update: The problems have been solved.

The trouble turned out to be a loose nut, that nut being me.

When I was reading the manual, I was assuming that the DSW switch drawings
were like the jumper drawings elsewhere in the book, where the dark part
represents the jumper itself. The DSW drawings actually use WHITE to
indicate the position of the switch. It makes sense once you think about
it, the jumpers being black and the DSW switches being white, but I was
thinking along different lines. I therefore had the switches backwards
from
where they should have been.
Click to expand...
You are not the only one! Today I assembled an identical system (A7V400-MX
+ Sempron 2400) for a friend of mine. I went through exactly the same steps
as you did, and also made the same mistake. Since I could not figure out
what was wrong, I starting browsing this newsgroup and read your posting in
this newsgroup, and Paul's replies. Only when I re-read the manual, and
compared the 166MHz DSW settings to the default DSW settings, I noticed that
not switch 1, but switched 2-5 should have been changed. That solved the
problem for me as well..
The RAM is now seen as 333. It's still not 400, but close enough. Maybe a
later BIOS flash will allow 400. I'm already up to BIOS version 1003, but
do not want to try the beta BIOS.
Click to expand...
I have 333Mhz RAM. The Sempron has a 333Mhz FSB. Can the RAM than run at
400? I though that was not possible.
Thanks for the help, Paul.
Click to expand...
Same here. Thanks Paul!

Tommy

(pardon my lame english, I'm Dutch).
 
"Tommy" said:
You are not the only one! Today I assembled an identical system (A7V400-MX
+ Sempron 2400) for a friend of mine. I went through exactly the same steps
as you did, and also made the same mistake. Since I could not figure out
what was wrong, I starting browsing this newsgroup and read your posting in
this newsgroup, and Paul's replies. Only when I re-read the manual, and
compared the 166MHz DSW settings to the default DSW settings, I noticed that
not switch 1, but switched 2-5 should have been changed. That solved the
problem for me as well..

I have 333Mhz RAM. The Sempron has a 333Mhz FSB. Can the RAM than run at
400? I though that was not possible.
Same here. Thanks Paul!

Tommy

(pardon my lame english, I'm Dutch).
Click to expand...

The worst part about DIP switches, is there are different
styles. There are slide switches (like the ones on the A7V400-MX) and
there are rocker switches. The best way for a manual to be written, is
to only show whether the switch is to be set "ON" or "OFF" for each
position, without showing any pictures of the switch itself. The
switch will usually have a legend printed on it for "ON" and "OFF",
so the user can figure out how to set them.

Glad you guys are up and running :-)

Enjoy,
Paul
 
Back
Top