«

Problems with DDR memory

--------------------------------------------------------------------------------

Bit of a long story so I will try to summarise.
PC came with 256m ram. DDR400 PC3200
I bought and exta gig and added into next slot to make 1.25 gig
All running well for many years.
Then upgraded to a X2 AMD 3800 processor.
I think I had reboot with that processor, anyway a month later
I got another gig of DDR400 PC3200

I had problems adding that, beeps etc.
However eventually I got running with the two 1 gig stick in the first two
(of 4) slots.
However I was getting reboots, about 1 every 4 hours on average (variable)

So I went to just the new 1 gig stick i slot 1.
That gave a reboot after 1.5 hours

So I tried just the old 1 gig stick in slot one - results -beeps!!!
I shifted it to slot 2 on a 'hunch' and it booted up correctly.

It has not rebbooted on that yet but I have only been running a couple of
hours
so not long enough to be conclusive.

Something I have noticed is that the first 1 gig stick I got appears to be
high
density, looking with the Everst utility (below) the module name 128M64K-40C
suggests it is arranged as high density.


However having said that it had been running fine along side what is a 256m
low
density module as a (rather badly) 'matched pair'

This is the Everest info for the old suspected high density stick.


Memory Module Properties:
Module Name 128M64K-40C
Serial Number 040413C3h
Module Size 1024 MB (2 ranks, 4 banks)
Module Type Unbuffered
Memory Type DDR SDRAM
Memory Speed PC3200 (200 MHz)
Module Width 64 bit
Module Voltage SSTL 2.5
Error Detection Method None
Refresh Rate Reduced (7.8 us), Self-Refresh

Memory Timings:
@ 200 MHz 3.0-3-3-8 (CL-RCD-RP-RAS)
@ 166 MHz 2.5-3-3-7 (CL-RCD-RP-RAS)
@ 133 MHz 2.0-2-2-6 (CL-RCD-RP-RAS)

Memory Module Features:
Early RAS# Precharge Not Supported
Auto-Precharge Not Supported
Precharge All Not Supported
Write1/Read Burst Not Supported
Buffered Address/Control Inputs Not Supported
Registered Address/Control Inputs Not Supported
On-Card PLL (Clock) Not Supported
Buffered DQMB Inputs Not Supported
Registered DQMB Inputs Not Supported
Differential Clock Input Supported
Redundant Row Address Not Supported



Anyway it is all rather confusing.

Some questions which occur are

1) Why does the above stick beep in slot one yet run fine (so far) in slot
2?
I though I put it in slot one OK, however it is possible slot 1 is 'dirty'.

I posted a long rambling thread about it here
http://www.techspot.com/vb/topic149949.html

Not getting any replies now, not surprising is it is so long and complicated
and involved
so I though I would try again here and try to summarise it a bit more
concisely.

I guess I will have to keep investigating and adding info until I can narrow
the problem
down more, even I am getting confused, I have 3 lots of memoryy from
different manufacturers - lol. I don't think it should matter who makes it,
however it seems a complex issue.

Also this may be related to my earlier thread here
"Sstem idle but CPU usage is high!!"

"Colin Trunt" <> wrote in message
news:2rq%n.113974$9c1.38686@hurricane...
> Problems with DDR memory
>
> --------------------------------------------------------------------------------
>
> Bit of a long story so I will try to summarise.
> PC came with 256m ram. DDR400 PC3200
> I bought and exta gig and added into next slot to make 1.25 gig
> All running well for many years.
> Then upgraded to a X2 AMD 3800 processor.
> I think I had reboot with that processor, anyway a month later
> I got another gig of DDR400 PC3200
>
> I had problems adding that, beeps etc.
> However eventually I got running with the two 1 gig stick in the first two
> (of 4) slots.
> However I was getting reboots, about 1 every 4 hours on average (variable)
>
> So I went to just the new 1 gig stick i slot 1.
> That gave a reboot after 1.5 hours
>
> So I tried just the old 1 gig stick in slot one - results -beeps!!!
> I shifted it to slot 2 on a 'hunch' and it booted up correctly.
>
> It has not rebbooted on that yet but I have only been running a couple of
> hours
> so not long enough to be conclusive.
>
> Something I have noticed is that the first 1 gig stick I got appears to be
> high
> density, looking with the Everst utility (below) the module name
> 128M64K-40C
> suggests it is arranged as high density.
>
>
> However having said that it had been running fine along side what is a
> 256m low
> density module as a (rather badly) 'matched pair'
>
> This is the Everest info for the old suspected high density stick.
>
>
> Memory Module Properties:
> Module Name 128M64K-40C
> Serial Number 040413C3h
> Module Size 1024 MB (2 ranks, 4 banks)
> Module Type Unbuffered
> Memory Type DDR SDRAM
> Memory Speed PC3200 (200 MHz)
> Module Width 64 bit
> Module Voltage SSTL 2.5
> Error Detection Method None
> Refresh Rate Reduced (7.8 us), Self-Refresh
>
> Memory Timings:
> @ 200 MHz 3.0-3-3-8 (CL-RCD-RP-RAS)
> @ 166 MHz 2.5-3-3-7 (CL-RCD-RP-RAS)
> @ 133 MHz 2.0-2-2-6 (CL-RCD-RP-RAS)
>
> Memory Module Features:
> Early RAS# Precharge Not Supported
> Auto-Precharge Not Supported
> Precharge All Not Supported
> Write1/Read Burst Not Supported
> Buffered Address/Control Inputs Not Supported
> Registered Address/Control Inputs Not Supported
> On-Card PLL (Clock) Not Supported
> Buffered DQMB Inputs Not Supported
> Registered DQMB Inputs Not Supported
> Differential Clock Input Supported
> Redundant Row Address Not Supported
>
>
>
> Anyway it is all rather confusing.
>
> Some questions which occur are
>
> 1) Why does the above stick beep in slot one yet run fine (so far) in slot
> 2?
> I though I put it in slot one OK, however it is possible slot 1 is
> 'dirty'.
>
> I posted a long rambling thread about it here
> http://www.techspot.com/vb/topic149949.html
>
> Not getting any replies now, not surprising is it is so long and
> complicated and involved
> so I though I would try again here and try to summarise it a bit more
> concisely.
>
> I guess I will have to keep investigating and adding info until I can
> narrow the problem
> down more, even I am getting confused, I have 3 lots of memoryy from
> different manufacturers - lol. I don't think it should matter who makes
> it, however it seems a complex issue.
>
> Also this may be related to my earlier thread here
> "Sstem idle but CPU usage is high!!"
>
>

OK I am jus going to stick down some info on the modules
for my own reference as much s anything else.

The 256meg module whihc came with the PC is:-

Infineon
It is labled one the module
HYS64D32300HU-5-C
This is the data sheet
http://pdf1.alldatasheet.net/datashe...300HU-5-C.html

The 8 chips on one side only are inscribed:-
hyb25d256800ce-5This ia data sheet
http://www.datasheetcatalog.org/data...11_rev1.11.pdf

That one seems OK the problemic one regaarding it may be high density is:

HYUIX 610AA A
HY5DU12822CPT 043
Kor

Hard to find info on thus.
The 128 in the description suggests it is high density.


The last one the new 1 gig I got is

Simmtronics

64X8DDR P

Colin Trunt wrote:
> Problems with DDR memory
>
> --------------------------------------------------------------------------------
>
> Bit of a long story so I will try to summarise.
> PC came with 256m ram. DDR400 PC3200
> I bought and exta gig and added into next slot to make 1.25 gig
> All running well for many years.
> Then upgraded to a X2 AMD 3800 processor.
> I think I had reboot with that processor, anyway a month later
> I got another gig of DDR400 PC3200
>
> I had problems adding that, beeps etc.
> However eventually I got running with the two 1 gig stick in the first two
> (of 4) slots.
> However I was getting reboots, about 1 every 4 hours on average (variable)
>
> So I went to just the new 1 gig stick i slot 1.
> That gave a reboot after 1.5 hours
>
> So I tried just the old 1 gig stick in slot one - results -beeps!!!
> I shifted it to slot 2 on a 'hunch' and it booted up correctly.
>
> It has not rebbooted on that yet but I have only been running a couple of
> hours
> so not long enough to be conclusive.
>
> Something I have noticed is that the first 1 gig stick I got appears to be
> high
> density, looking with the Everst utility (below) the module name 128M64K-40C
> suggests it is arranged as high density.
>
>
> However having said that it had been running fine along side what is a 256m
> low
> density module as a (rather badly) 'matched pair'
>
> This is the Everest info for the old suspected high density stick.
>
>
> Memory Module Properties:
> Module Name 128M64K-40C
> Serial Number 040413C3h
> Module Size 1024 MB (2 ranks, 4 banks)
> Module Type Unbuffered
> Memory Type DDR SDRAM
> Memory Speed PC3200 (200 MHz)
> Module Width 64 bit
> Module Voltage SSTL 2.5
> Error Detection Method None
> Refresh Rate Reduced (7.8 us), Self-Refresh
>
> Memory Timings:
> @ 200 MHz 3.0-3-3-8 (CL-RCD-RP-RAS)
> @ 166 MHz 2.5-3-3-7 (CL-RCD-RP-RAS)
> @ 133 MHz 2.0-2-2-6 (CL-RCD-RP-RAS)
>
> Memory Module Features:
> Early RAS# Precharge Not Supported
> Auto-Precharge Not Supported
> Precharge All Not Supported
> Write1/Read Burst Not Supported
> Buffered Address/Control Inputs Not Supported
> Registered Address/Control Inputs Not Supported
> On-Card PLL (Clock) Not Supported
> Buffered DQMB Inputs Not Supported
> Registered DQMB Inputs Not Supported
> Differential Clock Input Supported
> Redundant Row Address Not Supported
>
>
>
> Anyway it is all rather confusing.
>
> Some questions which occur are
>
> 1) Why does the above stick beep in slot one yet run fine (so far) in slot
> 2?
> I though I put it in slot one OK, however it is possible slot 1 is 'dirty'.
>
> I posted a long rambling thread about it here
> http://www.techspot.com/vb/topic149949.html
>
> Not getting any replies now, not surprising is it is so long and complicated
> and involved
> so I though I would try again here and try to summarise it a bit more
> concisely.
>
> I guess I will have to keep investigating and adding info until I can narrow
> the problem
> down more, even I am getting confused, I have 3 lots of memoryy from
> different manufacturers - lol. I don't think it should matter who makes it,
> however it seems a complex issue.
>
> Also this may be related to my earlier thread here
> "Sstem idle but CPU usage is high!!"
>
>

A high density DDR400 would be "1 rank, (16)x 128Mx4", That is
the kind they sell on Ebay. The info you list above is two
ranks, meaning the module is double sided (x 64Mx8 for a total
of 16 chips. That means the above listing is for a 1GB low
density double sided module.

(Example of a high density 1GB, mentions 128Mx4)
http://cgi.ebay.com/PC3200-1GB-DDR-S...item35a6581d50

Remember, that on modern AMD sockets (S754, S929, AM2/AM2+/AM3), the
memory controller is on the processor. A "bad slot" can be bad because
of dirty contacts. You fix that, by inserting and removing the module
five times, with all power removed from the computer. (Don't try
pink erasers or sandpaper, as the metal plating and the contacts
within the slot are delicate.) Then try testing the memory again.

A bad control signal on the processor itself, is a second reason. So
it could be an actual bad processor, that causes a bad DIMM slot.

You mention a processor upgrade, so if you were really interested
in whether the slot was bad or not, you could switch back to the
original processor, then test the RAM again.

While the system is fully operational (using whatever RAM config
you need to achieve that), get info for your new processor. I expect
it is a Rev.E or later processor, but it would be good to verify
using CPUZ. A processor like Rev.D0 would not like mismatched
modules side by side. If the processor is Rev.E or later, then
it should accept any mixture of RAM, and then "make up its own mind"
whether to run in dual channel or not. Processors before that,
are more picky. Maybe the beeping is a sign of a processor before
Rev.E.

Paul

Colin Trunt wrote:
> "Colin Trunt" <> wrote in message
> news:2rq%n.113974$9c1.38686@hurricane...
>> Problems with DDR memory
>>
>> --------------------------------------------------------------------------------
>>
>> Bit of a long story so I will try to summarise.
>> PC came with 256m ram. DDR400 PC3200
>> I bought and exta gig and added into next slot to make 1.25 gig
>> All running well for many years.
>> Then upgraded to a X2 AMD 3800 processor.
>> I think I had reboot with that processor, anyway a month later
>> I got another gig of DDR400 PC3200
>>
>> I had problems adding that, beeps etc.
>> However eventually I got running with the two 1 gig stick in the first two
>> (of 4) slots.
>> However I was getting reboots, about 1 every 4 hours on average (variable)
>>
>> So I went to just the new 1 gig stick i slot 1.
>> That gave a reboot after 1.5 hours
>>
>> So I tried just the old 1 gig stick in slot one - results -beeps!!!
>> I shifted it to slot 2 on a 'hunch' and it booted up correctly.
>>
>> It has not rebbooted on that yet but I have only been running a couple of
>> hours
>> so not long enough to be conclusive.
>>
>> Something I have noticed is that the first 1 gig stick I got appears to be
>> high
>> density, looking with the Everst utility (below) the module name
>> 128M64K-40C
>> suggests it is arranged as high density.
>>
>>
>> However having said that it had been running fine along side what is a
>> 256m low
>> density module as a (rather badly) 'matched pair'
>>
>> This is the Everest info for the old suspected high density stick.
>>
>>
>> Memory Module Properties:
>> Module Name 128M64K-40C
>> Serial Number 040413C3h
>> Module Size 1024 MB (2 ranks, 4 banks)
>> Module Type Unbuffered
>> Memory Type DDR SDRAM
>> Memory Speed PC3200 (200 MHz)
>> Module Width 64 bit
>> Module Voltage SSTL 2.5
>> Error Detection Method None
>> Refresh Rate Reduced (7.8 us), Self-Refresh
>>
>> Memory Timings:
>> @ 200 MHz 3.0-3-3-8 (CL-RCD-RP-RAS)
>> @ 166 MHz 2.5-3-3-7 (CL-RCD-RP-RAS)
>> @ 133 MHz 2.0-2-2-6 (CL-RCD-RP-RAS)
>>
>> Memory Module Features:
>> Early RAS# Precharge Not Supported
>> Auto-Precharge Not Supported
>> Precharge All Not Supported
>> Write1/Read Burst Not Supported
>> Buffered Address/Control Inputs Not Supported
>> Registered Address/Control Inputs Not Supported
>> On-Card PLL (Clock) Not Supported
>> Buffered DQMB Inputs Not Supported
>> Registered DQMB Inputs Not Supported
>> Differential Clock Input Supported
>> Redundant Row Address Not Supported
>>
>>
>>
>> Anyway it is all rather confusing.
>>
>> Some questions which occur are
>>
>> 1) Why does the above stick beep in slot one yet run fine (so far) in slot
>> 2?
>> I though I put it in slot one OK, however it is possible slot 1 is
>> 'dirty'.
>>
>> I posted a long rambling thread about it here
>> http://www.techspot.com/vb/topic149949.html
>>
>> Not getting any replies now, not surprising is it is so long and
>> complicated and involved
>> so I though I would try again here and try to summarise it a bit more
>> concisely.
>>
>> I guess I will have to keep investigating and adding info until I can
>> narrow the problem
>> down more, even I am getting confused, I have 3 lots of memoryy from
>> different manufacturers - lol. I don't think it should matter who makes
>> it, however it seems a complex issue.
>>
>> Also this may be related to my earlier thread here
>> "Sstem idle but CPU usage is high!!"
>>
>>
>
> OK I am jus going to stick down some info on the modules
> for my own reference as much s anything else.
>
> The 256meg module whihc came with the PC is:-
>
> Infineon
> It is labled one the module
> HYS64D32300HU-5-C
> This is the data sheet
> http://pdf1.alldatasheet.net/datashe...300HU-5-C.html
>
> The 8 chips on one side only are inscribed:-
> hyb25d256800ce-5This ia data sheet
> http://www.datasheetcatalog.org/data...11_rev1.11.pdf
>
> That one seems OK the problemic one regaarding it may be high density is:
>
> HYUIX 610AA A
> HY5DU12822CPT 043
> Kor
>
> Hard to find info on thus.
> The 128 in the description suggests it is high density.
>
>
> The last one the new 1 gig I got is
>
> Simmtronics
>
> 64X8DDR P
>
>

HY5DU12822CPT takes me here.

http://www.hynix.com/datasheet/eng/c...nuNo=1&m=7&s=2

"HY5DU12822C(L)TP-x*I 64Mx8"

The important thing there, is the "8" digit indicates the
width of the chip and 8 is good (low density). If the chip
was HY5DU124... that would be a four bit wide chip (nibble wide),
and that would be a high density chip.

Paul

"Paul" <> wrote in message
news:i1lrsu$hrv$...
> Colin Trunt wrote:
>> "Colin Trunt" <> wrote in message
>> news:2rq%n.113974$9c1.38686@hurricane...
>>> Problems with DDR memory
>>>
>>> --------------------------------------------------------------------------------
>>>
>>> Bit of a long story so I will try to summarise.
>>> PC came with 256m ram. DDR400 PC3200
>>> I bought and exta gig and added into next slot to make 1.25 gig
>>> All running well for many years.
>>> Then upgraded to a X2 AMD 3800 processor.
>>> I think I had reboot with that processor, anyway a month later
>>> I got another gig of DDR400 PC3200
>>>
>>> I had problems adding that, beeps etc.
>>> However eventually I got running with the two 1 gig stick in the first
>>> two (of 4) slots.
>>> However I was getting reboots, about 1 every 4 hours on average
>>> (variable)
>>>
>>> So I went to just the new 1 gig stick i slot 1.
>>> That gave a reboot after 1.5 hours
>>>
>>> So I tried just the old 1 gig stick in slot one - results -beeps!!!
>>> I shifted it to slot 2 on a 'hunch' and it booted up correctly.
>>>
>>> It has not rebbooted on that yet but I have only been running a couple
>>> of hours
>>> so not long enough to be conclusive.
>>>
>>> Something I have noticed is that the first 1 gig stick I got appears to
>>> be high
>>> density, looking with the Everst utility (below) the module name
>>> 128M64K-40C
>>> suggests it is arranged as high density.
>>>
>>>
>>> However having said that it had been running fine along side what is a
>>> 256m low
>>> density module as a (rather badly) 'matched pair'
>>>
>>> This is the Everest info for the old suspected high density stick.
>>>
>>>
>>> Memory Module Properties:
>>> Module Name 128M64K-40C
>>> Serial Number 040413C3h
>>> Module Size 1024 MB (2 ranks, 4 banks)
>>> Module Type Unbuffered
>>> Memory Type DDR SDRAM
>>> Memory Speed PC3200 (200 MHz)
>>> Module Width 64 bit
>>> Module Voltage SSTL 2.5
>>> Error Detection Method None
>>> Refresh Rate Reduced (7.8 us), Self-Refresh
>>>
>>> Memory Timings:
>>> @ 200 MHz 3.0-3-3-8 (CL-RCD-RP-RAS)
>>> @ 166 MHz 2.5-3-3-7 (CL-RCD-RP-RAS)
>>> @ 133 MHz 2.0-2-2-6 (CL-RCD-RP-RAS)
>>>
>>> Memory Module Features:
>>> Early RAS# Precharge Not Supported
>>> Auto-Precharge Not Supported
>>> Precharge All Not Supported
>>> Write1/Read Burst Not Supported
>>> Buffered Address/Control Inputs Not Supported
>>> Registered Address/Control Inputs Not Supported
>>> On-Card PLL (Clock) Not Supported
>>> Buffered DQMB Inputs Not Supported
>>> Registered DQMB Inputs Not Supported
>>> Differential Clock Input Supported
>>> Redundant Row Address Not Supported
>>>
>>>
>>>
>>> Anyway it is all rather confusing.
>>>
>>> Some questions which occur are
>>>
>>> 1) Why does the above stick beep in slot one yet run fine (so far) in
>>> slot 2?
>>> I though I put it in slot one OK, however it is possible slot 1 is
>>> 'dirty'.
>>>
>>> I posted a long rambling thread about it here
>>> http://www.techspot.com/vb/topic149949.html
>>>
>>> Not getting any replies now, not surprising is it is so long and
>>> complicated and involved
>>> so I though I would try again here and try to summarise it a bit more
>>> concisely.
>>>
>>> I guess I will have to keep investigating and adding info until I can
>>> narrow the problem
>>> down more, even I am getting confused, I have 3 lots of memoryy from
>>> different manufacturers - lol. I don't think it should matter who makes
>>> it, however it seems a complex issue.
>>>
>>> Also this may be related to my earlier thread here
>>> "Sstem idle but CPU usage is high!!"
>>>
>>>
>>
>> OK I am jus going to stick down some info on the modules
>> for my own reference as much s anything else.
>>
>> The 256meg module whihc came with the PC is:-
>>
>> Infineon
>> It is labled one the module
>> HYS64D32300HU-5-C
>> This is the data sheet
>> http://pdf1.alldatasheet.net/datashe...300HU-5-C.html
>>
>> The 8 chips on one side only are inscribed:-
>> hyb25d256800ce-5This ia data sheet
>> http://www.datasheetcatalog.org/data...11_rev1.11.pdf
>>
>> That one seems OK the problemic one regaarding it may be high density is:
>>
>> HYUIX 610AA A
>> HY5DU12822CPT 043
>> Kor
>>
>> Hard to find info on thus.
>> The 128 in the description suggests it is high density.
>>
>>
>> The last one the new 1 gig I got is
>>
>> Simmtronics
>>
>> 64X8DDR P
>
> HY5DU12822CPT takes me here.
>
> http://www.hynix.com/datasheet/eng/c...nuNo=1&m=7&s=2
>
> "HY5DU12822C(L)TP-x*I 64Mx8"
>
> The important thing there, is the "8" digit indicates the
> width of the chip and 8 is good (low density). If the chip
> was HY5DU124... that would be a four bit wide chip (nibble wide),
> and that would be a high density chip.
>
> Paul

Thanks for that thatis very helpful as it had confused me a lot.

It was this report from Everest which confused me.
--------[
SPD ]---------------------------------------------------------------------------------------------------------

[ DIMM2: 128M64K-40C ]

Memory Module Properties:
Module Name 128M64K-40C
Serial Number 040413C3h
Module Size 1024 MB (2 ranks, 4
banks)
Module Type Unbuffered
Memory Type DDR SDRAM
Memory Speed PC3200 (200 MHz)
Module Width 64 bit
Module Voltage SSTL 2.5
Error Detection Method None
Refresh Rate Reduced (7.8 us),
Self-Refresh

-----------------------------------------------------------------------------------------------------------------------------------

It was the module name which made me think it was high density 128M64K-40C
unfortunately I could not find anything useful with google to confirm but
your link is a great help!!

Anyway I have been running very smoothly for 10 hours with just that module
in my PC
so it seems fine.
The only percularity is that when I first tried it in slot one it gave
beeps, so I put it slot 2 instead
and I have had fine running from it.

I am going to try it in slot one again, it should work in there really.

I will also try the one which seems to cause reboots in slot 2 which seems
to be a 'good slot'.

Maybe both are good modules, I am hopinig so anyway, however id I get
reboots on slot 2 with the
other module (whihc I just bought) that looks pretty conclusive.

Incidently according to the motherboard manual my motherboard should take
either high
or low density modules as long as they are not mixed on a channel.
http://217.110.237.70/Manuals/7184-engl%20v1.x.pdf

It also seems to say it should not boot with just a module in slot 2,
however, it booted and
ran fine for me!!

"Paul" <> wrote in message
news:i1lr5j$f4l$...
> Colin Trunt wrote:
>> Problems with DDR memory
>>
>> --------------------------------------------------------------------------------
>>
>> Bit of a long story so I will try to summarise.
>> PC came with 256m ram. DDR400 PC3200
>> I bought and exta gig and added into next slot to make 1.25 gig
>> All running well for many years.
>> Then upgraded to a X2 AMD 3800 processor.
>> I think I had reboot with that processor, anyway a month later
>> I got another gig of DDR400 PC3200
>>
>> I had problems adding that, beeps etc.
>> However eventually I got running with the two 1 gig stick in the first
>> two (of 4) slots.
>> However I was getting reboots, about 1 every 4 hours on average
>> (variable)
>>
>> So I went to just the new 1 gig stick i slot 1.
>> That gave a reboot after 1.5 hours
>>
>> So I tried just the old 1 gig stick in slot one - results -beeps!!!
>> I shifted it to slot 2 on a 'hunch' and it booted up correctly.
>>
>> It has not rebbooted on that yet but I have only been running a couple of
>> hours
>> so not long enough to be conclusive.
>>
>> Something I have noticed is that the first 1 gig stick I got appears to
>> be high
>> density, looking with the Everst utility (below) the module name
>> 128M64K-40C
>> suggests it is arranged as high density.
>>
>>
>> However having said that it had been running fine along side what is a
>> 256m low
>> density module as a (rather badly) 'matched pair'
>>
>> This is the Everest info for the old suspected high density stick.
>>
>>
>> Memory Module Properties:
>> Module Name 128M64K-40C
>> Serial Number 040413C3h
>> Module Size 1024 MB (2 ranks, 4 banks)
>> Module Type Unbuffered
>> Memory Type DDR SDRAM
>> Memory Speed PC3200 (200 MHz)
>> Module Width 64 bit
>> Module Voltage SSTL 2.5
>> Error Detection Method None
>> Refresh Rate Reduced (7.8 us), Self-Refresh
>>
>> Memory Timings:
>> @ 200 MHz 3.0-3-3-8 (CL-RCD-RP-RAS)
>> @ 166 MHz 2.5-3-3-7 (CL-RCD-RP-RAS)
>> @ 133 MHz 2.0-2-2-6 (CL-RCD-RP-RAS)
>>
>> Memory Module Features:
>> Early RAS# Precharge Not Supported
>> Auto-Precharge Not Supported
>> Precharge All Not Supported
>> Write1/Read Burst Not Supported
>> Buffered Address/Control Inputs Not Supported
>> Registered Address/Control Inputs Not Supported
>> On-Card PLL (Clock) Not Supported
>> Buffered DQMB Inputs Not Supported
>> Registered DQMB Inputs Not Supported
>> Differential Clock Input Supported
>> Redundant Row Address Not Supported
>>
>>
>>
>> Anyway it is all rather confusing.
>>
>> Some questions which occur are
>>
>> 1) Why does the above stick beep in slot one yet run fine (so far) in
>> slot 2?
>> I though I put it in slot one OK, however it is possible slot 1 is
>> 'dirty'.
>>
>> I posted a long rambling thread about it here
>> http://www.techspot.com/vb/topic149949.html
>>
>> Not getting any replies now, not surprising is it is so long and
>> complicated and involved
>> so I though I would try again here and try to summarise it a bit more
>> concisely.
>>
>> I guess I will have to keep investigating and adding info until I can
>> narrow the problem
>> down more, even I am getting confused, I have 3 lots of memoryy from
>> different manufacturers - lol. I don't think it should matter who makes
>> it, however it seems a complex issue.
>>
>> Also this may be related to my earlier thread here
>> "Sstem idle but CPU usage is high!!"
>
> A high density DDR400 would be "1 rank, (16)x 128Mx4", That is
> the kind they sell on Ebay. The info you list above is two
> ranks, meaning the module is double sided (x 64Mx8 for a total
> of 16 chips. That means the above listing is for a 1GB low
> density double sided module.
>
> (Example of a high density 1GB, mentions 128Mx4)
> http://cgi.ebay.com/PC3200-1GB-DDR-S...item35a6581d50
>
> Remember, that on modern AMD sockets (S754, S929, AM2/AM2+/AM3), the
> memory controller is on the processor. A "bad slot" can be bad because
> of dirty contacts. You fix that, by inserting and removing the module
> five times, with all power removed from the computer. (Don't try
> pink erasers or sandpaper, as the metal plating and the contacts
> within the slot are delicate.) Then try testing the memory again.
>
> A bad control signal on the processor itself, is a second reason. So
> it could be an actual bad processor, that causes a bad DIMM slot.
>
> You mention a processor upgrade, so if you were really interested
> in whether the slot was bad or not, you could switch back to the
> original processor, then test the RAM again.
>
> While the system is fully operational (using whatever RAM config
> you need to achieve that), get info for your new processor. I expect
> it is a Rev.E or later processor, but it would be good to verify
> using CPUZ. A processor like Rev.D0 would not like mismatched
> modules side by side. If the processor is Rev.E or later, then
> it should accept any mixture of RAM, and then "make up its own mind"
> whether to run in dual channel or not. Processors before that,
> are more picky. Maybe the beeping is a sign of a processor before
> Rev.E.
>
> Paul

CPU Z says BH-4E

This is Everests CPU info:-

CPU Properties:
CPU Type DualCore AMD Athlon
64 X2, 2000 MHz (10 x 200) 3800+
CPU Alias Manchester S939
CPU Stepping BH-E4
Instruction Set x86, x86-64, MMX,
3DNow!, SSE, SSE2, SSE3
Min / Max CPU Multiplier 4x / 10x
L1 Code Cache 64 KB (Parity)
L1 Data Cache 64 KB (ECC)
L2 Cache 512 KB (On-Die,
ECC, Full-Speed)

Multi CPU:
Motherboard ID OEM00000
PROD00000000
CPU #0 AMD Athlon(tm) 64 X2
Dual Core Processor 3800+, 1989 MHz
CPU #1 AMD Athlon(tm) 64 X2
Dual Core Processor 3800+, 1989 MHz

CPU Physical Info:
Package Type 939 Pin uOPGA
Package Size 4.00 cm x 4.00 cm
Transistors 233.2 million
Process Technology 11Mi, 90 nm, CMOS,
Cu, SOI
Die Size 199 mm2
Core Voltage 1.100 - 1.400 V
I/O Voltage 1.2 V + 2.5 V
Maximum Power 89 - 110 W
(depending on clock speed)

CPU Manufacturer:
Company Name Advanced Micro
Devices, Inc.
Product Information
http://www.amd.com/us-en/Processors/...30_118,00.html

CPU Utilization:
CPU #1 / Core #1 0 %
CPU #1 / Core #2 3 %


-----------------------


Also the motherboard manual (page 1 says it will not work with 3 memeory
modules (1, 2 or 4 only)
That seems to be true as when I first added the new 1 gig to slot 3 (
already 256 meg in slot 1, 1 gig in slot2)
it was not very happy PC!!

Anyway I think I will reboot now and see if I can get this stick to work in
slot 1.
It should work because I had the other new 1 gig stick working in slot 1,
although
it did reboot occasionally. But then with it being a new module I guess it
would have better
contacts so that might explain it (and perhaps the occasional reboots).

So I will do what you said about reinserting and see what happens.

Thanks for your help, much appreciated!!

"Colin Trunt" <> wrote in message
news:LPw%n.218651$Yb4.105721@hurricane...
>
> "Paul" <> wrote in message
> news:i1lr5j$f4l$...
>> Colin Trunt wrote:
>>> Problems with DDR memory
>>>
>>> --------------------------------------------------------------------------------
>>>
>>> Bit of a long story so I will try to summarise.
>>> PC came with 256m ram. DDR400 PC3200
>>> I bought and exta gig and added into next slot to make 1.25 gig
>>> All running well for many years.
>>> Then upgraded to a X2 AMD 3800 processor.
>>> I think I had reboot with that processor, anyway a month later
>>> I got another gig of DDR400 PC3200
>>>
>>> I had problems adding that, beeps etc.
>>> However eventually I got running with the two 1 gig stick in the first
>>> two (of 4) slots.
>>> However I was getting reboots, about 1 every 4 hours on average
>>> (variable)
>>>
>>> So I went to just the new 1 gig stick i slot 1.
>>> That gave a reboot after 1.5 hours
>>>
>>> So I tried just the old 1 gig stick in slot one - results -beeps!!!
>>> I shifted it to slot 2 on a 'hunch' and it booted up correctly.
>>>
>>> It has not rebbooted on that yet but I have only been running a couple
>>> of hours
>>> so not long enough to be conclusive.
>>>
>>> Something I have noticed is that the first 1 gig stick I got appears to
>>> be high
>>> density, looking with the Everst utility (below) the module name
>>> 128M64K-40C
>>> suggests it is arranged as high density.
>>>
>>>
>>> However having said that it had been running fine along side what is a
>>> 256m low
>>> density module as a (rather badly) 'matched pair'
>>>
>>> This is the Everest info for the old suspected high density stick.
>>>
>>>
>>> Memory Module Properties:
>>> Module Name 128M64K-40C
>>> Serial Number 040413C3h
>>> Module Size 1024 MB (2 ranks, 4 banks)
>>> Module Type Unbuffered
>>> Memory Type DDR SDRAM
>>> Memory Speed PC3200 (200 MHz)
>>> Module Width 64 bit
>>> Module Voltage SSTL 2.5
>>> Error Detection Method None
>>> Refresh Rate Reduced (7.8 us), Self-Refresh
>>>
>>> Memory Timings:
>>> @ 200 MHz 3.0-3-3-8 (CL-RCD-RP-RAS)
>>> @ 166 MHz 2.5-3-3-7 (CL-RCD-RP-RAS)
>>> @ 133 MHz 2.0-2-2-6 (CL-RCD-RP-RAS)
>>>
>>> Memory Module Features:
>>> Early RAS# Precharge Not Supported
>>> Auto-Precharge Not Supported
>>> Precharge All Not Supported
>>> Write1/Read Burst Not Supported
>>> Buffered Address/Control Inputs Not Supported
>>> Registered Address/Control Inputs Not Supported
>>> On-Card PLL (Clock) Not Supported
>>> Buffered DQMB Inputs Not Supported
>>> Registered DQMB Inputs Not Supported
>>> Differential Clock Input Supported
>>> Redundant Row Address Not Supported
>>>
>>>
>>>
>>> Anyway it is all rather confusing.
>>>
>>> Some questions which occur are
>>>
>>> 1) Why does the above stick beep in slot one yet run fine (so far) in
>>> slot 2?
>>> I though I put it in slot one OK, however it is possible slot 1 is
>>> 'dirty'.
>>>
>>> I posted a long rambling thread about it here
>>> http://www.techspot.com/vb/topic149949.html
>>>
>>> Not getting any replies now, not surprising is it is so long and
>>> complicated and involved
>>> so I though I would try again here and try to summarise it a bit more
>>> concisely.
>>>
>>> I guess I will have to keep investigating and adding info until I can
>>> narrow the problem
>>> down more, even I am getting confused, I have 3 lots of memoryy from
>>> different manufacturers - lol. I don't think it should matter who makes
>>> it, however it seems a complex issue.
>>>
>>> Also this may be related to my earlier thread here
>>> "Sstem idle but CPU usage is high!!"
>>
>> A high density DDR400 would be "1 rank, (16)x 128Mx4", That is
>> the kind they sell on Ebay. The info you list above is two
>> ranks, meaning the module is double sided (x 64Mx8 for a total
>> of 16 chips. That means the above listing is for a 1GB low
>> density double sided module.
>>
>> (Example of a high density 1GB, mentions 128Mx4)
>> http://cgi.ebay.com/PC3200-1GB-DDR-S...item35a6581d50
>>
>> Remember, that on modern AMD sockets (S754, S929, AM2/AM2+/AM3), the
>> memory controller is on the processor. A "bad slot" can be bad because
>> of dirty contacts. You fix that, by inserting and removing the module
>> five times, with all power removed from the computer. (Don't try
>> pink erasers or sandpaper, as the metal plating and the contacts
>> within the slot are delicate.) Then try testing the memory again.
>>
>> A bad control signal on the processor itself, is a second reason. So
>> it could be an actual bad processor, that causes a bad DIMM slot.
>>
>> You mention a processor upgrade, so if you were really interested
>> in whether the slot was bad or not, you could switch back to the
>> original processor, then test the RAM again.
>>
>> While the system is fully operational (using whatever RAM config
>> you need to achieve that), get info for your new processor. I expect
>> it is a Rev.E or later processor, but it would be good to verify
>> using CPUZ. A processor like Rev.D0 would not like mismatched
>> modules side by side. If the processor is Rev.E or later, then
>> it should accept any mixture of RAM, and then "make up its own mind"
>> whether to run in dual channel or not. Processors before that,
>> are more picky. Maybe the beeping is a sign of a processor before
>> Rev.E.
>>
>> Paul
>
> CPU Z says BH-4E
>
> This is Everests CPU info:-
>
> CPU Properties:
> CPU Type DualCore AMD Athlon
> 64 X2, 2000 MHz (10 x 200) 3800+
> CPU Alias Manchester S939
> CPU Stepping BH-E4
> Instruction Set x86, x86-64, MMX,
> 3DNow!, SSE, SSE2, SSE3
> Min / Max CPU Multiplier 4x / 10x
> L1 Code Cache 64 KB (Parity)
> L1 Data Cache 64 KB (ECC)
> L2 Cache 512 KB (On-Die,
> ECC, Full-Speed)
>
> Multi CPU:
> Motherboard ID OEM00000
> PROD00000000
> CPU #0 AMD Athlon(tm) 64
> X2 Dual Core Processor 3800+, 1989 MHz
> CPU #1 AMD Athlon(tm) 64
> X2 Dual Core Processor 3800+, 1989 MHz
>
> CPU Physical Info:
> Package Type 939 Pin uOPGA
> Package Size 4.00 cm x 4.00 cm
> Transistors 233.2 million
> Process Technology 11Mi, 90 nm, CMOS,
> Cu, SOI
> Die Size 199 mm2
> Core Voltage 1.100 - 1.400 V
> I/O Voltage 1.2 V + 2.5 V
> Maximum Power 89 - 110 W
> (depending on clock speed)
>
> CPU Manufacturer:
> Company Name Advanced Micro
> Devices, Inc.
> Product Information
> http://www.amd.com/us-en/Processors/...30_118,00.html
>
> CPU Utilization:
> CPU #1 / Core #1 0 %
> CPU #1 / Core #2 3 %
>
>
> -----------------------
>
>
> Also the motherboard manual (page 1 says it will not work with 3 memeory
> modules (1, 2 or 4 only)
> That seems to be true as when I first added the new 1 gig to slot 3 (
> already 256 meg in slot 1, 1 gig in slot2)
> it was not very happy PC!!
>
> Anyway I think I will reboot now and see if I can get this stick to work
> in slot 1.
> It should work because I had the other new 1 gig stick working in slot 1,
> although
> it did reboot occasionally. But then with it being a new module I guess it
> would have better
> contacts so that might explain it (and perhaps the occasional reboots).
>
> So I will do what you said about reinserting and see what happens.
>
> Thanks for your help, much appreciated!!


Right I tried that reinsertion 'trick' and it has worked a treat, booted-up
no problem.
So all I need to do now it retest the new module I bought, I will put it in
by itself to avoid any compatabilty problem, I don't want to send it back
for them to say it works fine for them.
Hopefully I won't have to send it back at all, if it works by itself but not
with
my module I can sell it on ebay (or sell the old one), and buy another of
the same type.

It will take a while to test though as the reboots were sporadic, I would
want a good 12
hours without a reboot.


Thanks one again, you certainly know your stuff!!
>
>
>
>

Colin Trunt wrote:

>
> It also seems to say it should not boot with just a module in slot 2,
> however, it booted and
> ran fine for me!!
>

That depends on the revision of processor. For processors before Revision E,
such as D0 or CG, the statement in the manual would be correct. For Revision
E (E3 or E6 are examples), then anything goes. You should be able to stick the
DIMM in any slot and have it work with Revision E. That is why I wanted you to
check the CPU revision, if you can find a utility to do it. CPUZ is available from
cpuid.com , but there are likely others that can identify the processor.

Paul

Colin Trunt wrote:
> "Colin Trunt" <> wrote in message
> news:LPw%n.218651$Yb4.105721@hurricane...
>> "Paul" <> wrote in message
>> news:i1lr5j$f4l$...
>>> Colin Trunt wrote:
>>>> Problems with DDR memory
>>>>
>>>> --------------------------------------------------------------------------------
>>>>
>>>> Bit of a long story so I will try to summarise.
>>>> PC came with 256m ram. DDR400 PC3200
>>>> I bought and exta gig and added into next slot to make 1.25 gig
>>>> All running well for many years.
>>>> Then upgraded to a X2 AMD 3800 processor.
>>>> I think I had reboot with that processor, anyway a month later
>>>> I got another gig of DDR400 PC3200
>>>>
>>>> I had problems adding that, beeps etc.
>>>> However eventually I got running with the two 1 gig stick in the first
>>>> two (of 4) slots.
>>>> However I was getting reboots, about 1 every 4 hours on average
>>>> (variable)
>>>>
>>>> So I went to just the new 1 gig stick i slot 1.
>>>> That gave a reboot after 1.5 hours
>>>>
>>>> So I tried just the old 1 gig stick in slot one - results -beeps!!!
>>>> I shifted it to slot 2 on a 'hunch' and it booted up correctly.
>>>>
>>>> It has not rebbooted on that yet but I have only been running a couple
>>>> of hours
>>>> so not long enough to be conclusive.
>>>>
>>>> Something I have noticed is that the first 1 gig stick I got appears to
>>>> be high
>>>> density, looking with the Everst utility (below) the module name
>>>> 128M64K-40C
>>>> suggests it is arranged as high density.
>>>>
>>>>
>>>> However having said that it had been running fine along side what is a
>>>> 256m low
>>>> density module as a (rather badly) 'matched pair'
>>>>
>>>> This is the Everest info for the old suspected high density stick.
>>>>
>>>>
>>>> Memory Module Properties:
>>>> Module Name 128M64K-40C
>>>> Serial Number 040413C3h
>>>> Module Size 1024 MB (2 ranks, 4 banks)
>>>> Module Type Unbuffered
>>>> Memory Type DDR SDRAM
>>>> Memory Speed PC3200 (200 MHz)
>>>> Module Width 64 bit
>>>> Module Voltage SSTL 2.5
>>>> Error Detection Method None
>>>> Refresh Rate Reduced (7.8 us), Self-Refresh
>>>>
>>>> Memory Timings:
>>>> @ 200 MHz 3.0-3-3-8 (CL-RCD-RP-RAS)
>>>> @ 166 MHz 2.5-3-3-7 (CL-RCD-RP-RAS)
>>>> @ 133 MHz 2.0-2-2-6 (CL-RCD-RP-RAS)
>>>>
>>>> Memory Module Features:
>>>> Early RAS# Precharge Not Supported
>>>> Auto-Precharge Not Supported
>>>> Precharge All Not Supported
>>>> Write1/Read Burst Not Supported
>>>> Buffered Address/Control Inputs Not Supported
>>>> Registered Address/Control Inputs Not Supported
>>>> On-Card PLL (Clock) Not Supported
>>>> Buffered DQMB Inputs Not Supported
>>>> Registered DQMB Inputs Not Supported
>>>> Differential Clock Input Supported
>>>> Redundant Row Address Not Supported
>>>>
>>>>
>>>>
>>>> Anyway it is all rather confusing.
>>>>
>>>> Some questions which occur are
>>>>
>>>> 1) Why does the above stick beep in slot one yet run fine (so far) in
>>>> slot 2?
>>>> I though I put it in slot one OK, however it is possible slot 1 is
>>>> 'dirty'.
>>>>
>>>> I posted a long rambling thread about it here
>>>> http://www.techspot.com/vb/topic149949.html
>>>>
>>>> Not getting any replies now, not surprising is it is so long and
>>>> complicated and involved
>>>> so I though I would try again here and try to summarise it a bit more
>>>> concisely.
>>>>
>>>> I guess I will have to keep investigating and adding info until I can
>>>> narrow the problem
>>>> down more, even I am getting confused, I have 3 lots of memoryy from
>>>> different manufacturers - lol. I don't think it should matter who makes
>>>> it, however it seems a complex issue.
>>>>
>>>> Also this may be related to my earlier thread here
>>>> "Sstem idle but CPU usage is high!!"
>>> A high density DDR400 would be "1 rank, (16)x 128Mx4", That is
>>> the kind they sell on Ebay. The info you list above is two
>>> ranks, meaning the module is double sided (x 64Mx8 for a total
>>> of 16 chips. That means the above listing is for a 1GB low
>>> density double sided module.
>>>
>>> (Example of a high density 1GB, mentions 128Mx4)
>>> http://cgi.ebay.com/PC3200-1GB-DDR-S...item35a6581d50
>>>
>>> Remember, that on modern AMD sockets (S754, S929, AM2/AM2+/AM3), the
>>> memory controller is on the processor. A "bad slot" can be bad because
>>> of dirty contacts. You fix that, by inserting and removing the module
>>> five times, with all power removed from the computer. (Don't try
>>> pink erasers or sandpaper, as the metal plating and the contacts
>>> within the slot are delicate.) Then try testing the memory again.
>>>
>>> A bad control signal on the processor itself, is a second reason. So
>>> it could be an actual bad processor, that causes a bad DIMM slot.
>>>
>>> You mention a processor upgrade, so if you were really interested
>>> in whether the slot was bad or not, you could switch back to the
>>> original processor, then test the RAM again.
>>>
>>> While the system is fully operational (using whatever RAM config
>>> you need to achieve that), get info for your new processor. I expect
>>> it is a Rev.E or later processor, but it would be good to verify
>>> using CPUZ. A processor like Rev.D0 would not like mismatched
>>> modules side by side. If the processor is Rev.E or later, then
>>> it should accept any mixture of RAM, and then "make up its own mind"
>>> whether to run in dual channel or not. Processors before that,
>>> are more picky. Maybe the beeping is a sign of a processor before
>>> Rev.E.
>>>
>>> Paul
>> CPU Z says BH-4E
>>
>> This is Everests CPU info:-
>>
>> CPU Properties:
>> CPU Type DualCore AMD Athlon
>> 64 X2, 2000 MHz (10 x 200) 3800+
>> CPU Alias Manchester S939
>> CPU Stepping BH-E4
>> Instruction Set x86, x86-64, MMX,
>> 3DNow!, SSE, SSE2, SSE3
>> Min / Max CPU Multiplier 4x / 10x
>> L1 Code Cache 64 KB (Parity)
>> L1 Data Cache 64 KB (ECC)
>> L2 Cache 512 KB (On-Die,
>> ECC, Full-Speed)
>>
>> Multi CPU:
>> Motherboard ID OEM00000
>> PROD00000000
>> CPU #0 AMD Athlon(tm) 64
>> X2 Dual Core Processor 3800+, 1989 MHz
>> CPU #1 AMD Athlon(tm) 64
>> X2 Dual Core Processor 3800+, 1989 MHz
>>
>> CPU Physical Info:
>> Package Type 939 Pin uOPGA
>> Package Size 4.00 cm x 4.00 cm
>> Transistors 233.2 million
>> Process Technology 11Mi, 90 nm, CMOS,
>> Cu, SOI
>> Die Size 199 mm2
>> Core Voltage 1.100 - 1.400 V
>> I/O Voltage 1.2 V + 2.5 V
>> Maximum Power 89 - 110 W
>> (depending on clock speed)
>>
>> CPU Manufacturer:
>> Company Name Advanced Micro
>> Devices, Inc.
>> Product Information
>> http://www.amd.com/us-en/Processors/...30_118,00.html
>>
>> CPU Utilization:
>> CPU #1 / Core #1 0 %
>> CPU #1 / Core #2 3 %
>>
>>
>> -----------------------
>>
>>
>> Also the motherboard manual (page 1 says it will not work with 3 memeory
>> modules (1, 2 or 4 only)
>> That seems to be true as when I first added the new 1 gig to slot 3 (
>> already 256 meg in slot 1, 1 gig in slot2)
>> it was not very happy PC!!
>>
>> Anyway I think I will reboot now and see if I can get this stick to work
>> in slot 1.
>> It should work because I had the other new 1 gig stick working in slot 1,
>> although
>> it did reboot occasionally. But then with it being a new module I guess it
>> would have better
>> contacts so that might explain it (and perhaps the occasional reboots).
>>
>> So I will do what you said about reinserting and see what happens.
>>
>> Thanks for your help, much appreciated!!
>
>
> Right I tried that reinsertion 'trick' and it has worked a treat, booted-up
> no problem.
> So all I need to do now it retest the new module I bought, I will put it in
> by itself to avoid any compatabilty problem, I don't want to send it back
> for them to say it works fine for them.
> Hopefully I won't have to send it back at all, if it works by itself but not
> with
> my module I can sell it on ebay (or sell the old one), and buy another of
> the same type.
>
> It will take a while to test though as the reboots were sporadic, I would
> want a good 12
> hours without a reboot.
>
>
> Thanks one again, you certainly know your stuff!!

I think the above report, is saying your processor revision is E4.
That means the processor supports 1,2,3, or 4 stick configurations,
and the single stick can be placed in any slot.

The best configuration is dual channel, and personally, if I had
such a system, I'd probably stick 2x1GB matched in it. You can run the
best timings, with two sticks in dual channel configuration.
If I had a third stick, I'd just store it in the drawer, as it
would only slow things down.

A lot of the more recent designs, support one degree more freedom
than your S939 system. Some now, support dual channel whenever the
total quantity of memory on both channels is the same. On such systems,
you could mix unmatched 1GB modules, and it would still run dual channel.
(Or, you could put 2x512MB DIMMs on one channel, and a single 1GB DIMM
on the other channel, and that would still be dual channel.)

But your E4 processor doesn't do that. It had one sin correct in its
design, but they didn't go "all the way" and fix it right. Still, on
balance, because a user can plug any configuration into your E4, and get it
to work, I don't count that as all bad. You can plug in any configuration,
but some of them work better than others (higher bandwidth in dual channel).

I find the AMD documentation dreadful, which is one reason I don't keep
up to date on their stuff. I hate having to confirm info about memory
configurations, because their documentation is not focused on a single
product per document.

Paul

"Paul" <> wrote in message
news:i1m8pm$d5t$...
> Colin Trunt wrote:
>> "Colin Trunt" <> wrote in message
>> news:LPw%n.218651$Yb4.105721@hurricane...
>>> "Paul" <> wrote in message
>>> news:i1lr5j$f4l$...
>>>> Colin Trunt wrote:
>>>>> Problems with DDR memory
>>>>>
>>>>> --------------------------------------------------------------------------------
>>>>>
>>>>> Bit of a long story so I will try to summarise.
>>>>> PC came with 256m ram. DDR400 PC3200
>>>>> I bought and exta gig and added into next slot to make 1.25 gig
>>>>> All running well for many years.
>>>>> Then upgraded to a X2 AMD 3800 processor.
>>>>> I think I had reboot with that processor, anyway a month later
>>>>> I got another gig of DDR400 PC3200
>>>>>
>>>>> I had problems adding that, beeps etc.
>>>>> However eventually I got running with the two 1 gig stick in the first
>>>>> two (of 4) slots.
>>>>> However I was getting reboots, about 1 every 4 hours on average
>>>>> (variable)
>>>>>
>>>>> So I went to just the new 1 gig stick i slot 1.
>>>>> That gave a reboot after 1.5 hours
>>>>>
>>>>> So I tried just the old 1 gig stick in slot one - results -beeps!!!
>>>>> I shifted it to slot 2 on a 'hunch' and it booted up correctly.
>>>>>
>>>>> It has not rebbooted on that yet but I have only been running a couple
>>>>> of hours
>>>>> so not long enough to be conclusive.
>>>>>
>>>>> Something I have noticed is that the first 1 gig stick I got appears
>>>>> to be high
>>>>> density, looking with the Everst utility (below) the module name
>>>>> 128M64K-40C
>>>>> suggests it is arranged as high density.
>>>>>
>>>>>
>>>>> However having said that it had been running fine along side what is a
>>>>> 256m low
>>>>> density module as a (rather badly) 'matched pair'
>>>>>
>>>>> This is the Everest info for the old suspected high density stick.
>>>>>
>>>>>
>>>>> Memory Module Properties:
>>>>> Module Name 128M64K-40C
>>>>> Serial Number 040413C3h
>>>>> Module Size 1024 MB (2 ranks, 4 banks)
>>>>> Module Type Unbuffered
>>>>> Memory Type DDR SDRAM
>>>>> Memory Speed PC3200 (200 MHz)
>>>>> Module Width 64 bit
>>>>> Module Voltage SSTL 2.5
>>>>> Error Detection Method None
>>>>> Refresh Rate Reduced (7.8 us), Self-Refresh
>>>>>
>>>>> Memory Timings:
>>>>> @ 200 MHz 3.0-3-3-8 (CL-RCD-RP-RAS)
>>>>> @ 166 MHz 2.5-3-3-7 (CL-RCD-RP-RAS)
>>>>> @ 133 MHz 2.0-2-2-6 (CL-RCD-RP-RAS)
>>>>>
>>>>> Memory Module Features:
>>>>> Early RAS# Precharge Not Supported
>>>>> Auto-Precharge Not Supported
>>>>> Precharge All Not Supported
>>>>> Write1/Read Burst Not Supported
>>>>> Buffered Address/Control Inputs Not Supported
>>>>> Registered Address/Control Inputs Not Supported
>>>>> On-Card PLL (Clock) Not Supported
>>>>> Buffered DQMB Inputs Not Supported
>>>>> Registered DQMB Inputs Not Supported
>>>>> Differential Clock Input Supported
>>>>> Redundant Row Address Not Supported
>>>>>
>>>>>
>>>>>
>>>>> Anyway it is all rather confusing.
>>>>>
>>>>> Some questions which occur are
>>>>>
>>>>> 1) Why does the above stick beep in slot one yet run fine (so far) in
>>>>> slot 2?
>>>>> I though I put it in slot one OK, however it is possible slot 1 is
>>>>> 'dirty'.
>>>>>
>>>>> I posted a long rambling thread about it here
>>>>> http://www.techspot.com/vb/topic149949.html
>>>>>
>>>>> Not getting any replies now, not surprising is it is so long and
>>>>> complicated and involved
>>>>> so I though I would try again here and try to summarise it a bit more
>>>>> concisely.
>>>>>
>>>>> I guess I will have to keep investigating and adding info until I can
>>>>> narrow the problem
>>>>> down more, even I am getting confused, I have 3 lots of memoryy from
>>>>> different manufacturers - lol. I don't think it should matter who
>>>>> makes it, however it seems a complex issue.
>>>>>
>>>>> Also this may be related to my earlier thread here
>>>>> "Sstem idle but CPU usage is high!!"
>>>> A high density DDR400 would be "1 rank, (16)x 128Mx4", That is
>>>> the kind they sell on Ebay. The info you list above is two
>>>> ranks, meaning the module is double sided (x 64Mx8 for a total
>>>> of 16 chips. That means the above listing is for a 1GB low
>>>> density double sided module.
>>>>
>>>> (Example of a high density 1GB, mentions 128Mx4)
>>>> http://cgi.ebay.com/PC3200-1GB-DDR-S...item35a6581d50
>>>>
>>>> Remember, that on modern AMD sockets (S754, S929, AM2/AM2+/AM3), the
>>>> memory controller is on the processor. A "bad slot" can be bad because
>>>> of dirty contacts. You fix that, by inserting and removing the module
>>>> five times, with all power removed from the computer. (Don't try
>>>> pink erasers or sandpaper, as the metal plating and the contacts
>>>> within the slot are delicate.) Then try testing the memory again.
>>>>
>>>> A bad control signal on the processor itself, is a second reason. So
>>>> it could be an actual bad processor, that causes a bad DIMM slot.
>>>>
>>>> You mention a processor upgrade, so if you were really interested
>>>> in whether the slot was bad or not, you could switch back to the
>>>> original processor, then test the RAM again.
>>>>
>>>> While the system is fully operational (using whatever RAM config
>>>> you need to achieve that), get info for your new processor. I expect
>>>> it is a Rev.E or later processor, but it would be good to verify
>>>> using CPUZ. A processor like Rev.D0 would not like mismatched
>>>> modules side by side. If the processor is Rev.E or later, then
>>>> it should accept any mixture of RAM, and then "make up its own mind"
>>>> whether to run in dual channel or not. Processors before that,
>>>> are more picky. Maybe the beeping is a sign of a processor before
>>>> Rev.E.
>>>>
>>>> Paul
>>> CPU Z says BH-4E
>>>
>>> This is Everests CPU info:-
>>>
>>> CPU Properties:
>>> CPU Type DualCore AMD
>>> Athlon 64 X2, 2000 MHz (10 x 200) 3800+
>>> CPU Alias Manchester S939
>>> CPU Stepping BH-E4
>>> Instruction Set x86, x86-64, MMX,
>>> 3DNow!, SSE, SSE2, SSE3
>>> Min / Max CPU Multiplier 4x / 10x
>>> L1 Code Cache 64 KB (Parity)
>>> L1 Data Cache 64 KB (ECC)
>>> L2 Cache 512 KB (On-Die,
>>> ECC, Full-Speed)
>>>
>>> Multi CPU:
>>> Motherboard ID OEM00000
>>> PROD00000000
>>> CPU #0 AMD Athlon(tm) 64
>>> X2 Dual Core Processor 3800+, 1989 MHz
>>> CPU #1 AMD Athlon(tm) 64
>>> X2 Dual Core Processor 3800+, 1989 MHz
>>>
>>> CPU Physical Info:
>>> Package Type 939 Pin uOPGA
>>> Package Size 4.00 cm x 4.00 cm
>>> Transistors 233.2 million
>>> Process Technology 11Mi, 90 nm,
>>> CMOS, Cu, SOI
>>> Die Size 199 mm2
>>> Core Voltage 1.100 - 1.400 V
>>> I/O Voltage 1.2 V + 2.5 V
>>> Maximum Power 89 - 110 W
>>> (depending on clock speed)
>>>
>>> CPU Manufacturer:
>>> Company Name Advanced Micro
>>> Devices, Inc.
>>> Product Information
>>> http://www.amd.com/us-en/Processors/...30_118,00.html
>>>
>>> CPU Utilization:
>>> CPU #1 / Core #1 0 %
>>> CPU #1 / Core #2 3 %
>>>
>>>
>>> -----------------------
>>>
>>>
>>> Also the motherboard manual (page 1 says it will not work with 3
>>> memeory modules (1, 2 or 4 only)
>>> That seems to be true as when I first added the new 1 gig to slot 3 (
>>> already 256 meg in slot 1, 1 gig in slot2)
>>> it was not very happy PC!!
>>>
>>> Anyway I think I will reboot now and see if I can get this stick to work
>>> in slot 1.
>>> It should work because I had the other new 1 gig stick working in slot
>>> 1, although
>>> it did reboot occasionally. But then with it being a new module I guess
>>> it would have better
>>> contacts so that might explain it (and perhaps the occasional reboots).
>>>
>>> So I will do what you said about reinserting and see what happens.
>>>
>>> Thanks for your help, much appreciated!!
>>
>>
>> Right I tried that reinsertion 'trick' and it has worked a treat,
>> booted-up
>> no problem.
>> So all I need to do now it retest the new module I bought, I will put it
>> in
>> by itself to avoid any compatabilty problem, I don't want to send it back
>> for them to say it works fine for them.
>> Hopefully I won't have to send it back at all, if it works by itself but
>> not with
>> my module I can sell it on ebay (or sell the old one), and buy another of
>> the same type.
>>
>> It will take a while to test though as the reboots were sporadic, I would
>> want a good 12
>> hours without a reboot.
>>
>>
>> Thanks one again, you certainly know your stuff!!
>
> I think the above report, is saying your processor revision is E4.
> That means the processor supports 1,2,3, or 4 stick configurations,
> and the single stick can be placed in any slot.
>
> The best configuration is dual channel, and personally, if I had
> such a system, I'd probably stick 2x1GB matched in it. You can run the
> best timings, with two sticks in dual channel configuration.
> If I had a third stick, I'd just store it in the drawer, as it
> would only slow things down.
>
> A lot of the more recent designs, support one degree more freedom
> than your S939 system. Some now, support dual channel whenever the
> total quantity of memory on both channels is the same. On such systems,
> you could mix unmatched 1GB modules, and it would still run dual channel.
> (Or, you could put 2x512MB DIMMs on one channel, and a single 1GB DIMM
> on the other channel, and that would still be dual channel.)
>
> But your E4 processor doesn't do that. It had one sin correct in its
> design, but they didn't go "all the way" and fix it right. Still, on
> balance, because a user can plug any configuration into your E4, and get
> it
> to work, I don't count that as all bad. You can plug in any configuration,
> but some of them work better than others (higher bandwidth in dual
> channel).
>
> I find the AMD documentation dreadful, which is one reason I don't keep
> up to date on their stuff. I hate having to confirm info about memory
> configurations, because their documentation is not focused on a single
> product per document.
>
> Paul

OK so I have got the result I really did not want, I had it running for
over 7 hours on the new module in slot one, I was beginning to think
the problem was cured when *whoosh* reboot.
I let it boot up into windows then I started a chkdsk, but then it rebooted
again.

I think I only had it running about 10 hours on the other processor so it is
possible a reboot might have occured on that processor had I left it long
enough.

So it looks like it is faulty but I would need more time to be 100% sure,
I am about 90% sure at the moment, I mean it could possible be something
else.

Anyway I think I will burn one of those memory test CDS and hope
that produces something conclusive.
I expect it will find nothing unless I leave it running for hours and hours.