Strange memory problem

A

Asko

I have a problem in DDR memory speed. When I do a normal cold boot the
memory speed is 200 MHz but when the computer is rebooted from the
Start-menu ( "Start Again" I guess that might be in an English version of
Windows XP) the memory speed is 266 MHz, which would be the figure I would
it like to be. The hardware specs allow that.

I get information of the system through a couple of programs such as cpu-z,
Sisoftware Sandra and PCWizard 2006. Here is listed information about this
issue. There is two DDR DIMMs 256 MB each. One is PC-2100, the other
PC-2700. My question is why the memory speed is lower when cold booted and
what to do in order to get the higher speed out without rebooting? Take a
look at the asynchronous memory speed multiplier, which is 1/1 when cold
booted, but 4/3 when rebooted and it should be that 4/3.

Mainboard : Asus P4GE-FSC
Chipset : Intel i845GE
Processor : Intel Celeron @ 2600 MHz
Physical Memory : 512 MB (2 x 256 DDR-SDRAM )

NorthBridge Information :
Architecture : HUB
Manufacturer : Intel (ASUSTeK Computer Inc)
Codename : Brookdale-G
Revision : B1
Bus Speed : 100.9 MHz THIS BUS SPEED IS NOT TWEAKABLE AT BIOS
FSB Frequency : 403.6 MHz (QDR)
FSB max. Support : 533 MHz
RAM max. Support : DDR (333 MHz)

Memory at Slot 1 :
Manufacturer : Kingston
Part Number : KVR266X64C25/256
Type : DDR-SDRAM PC-2100 (133 MHz) - [DDR-266]
Size : 256 MB (1 rows, 4 banks)
Supported Frequencies : 100 MHz, 133 MHz
CAS Latency (tCL) : 2 clocks @100 MHz, 2.5 clocks @133 MHz
RAS to CAS (tRCD) : 2 clocks @100 MHz, 3 clocks @133 MHz
RAS Precharge (tRP) : 2 clocks @100 MHz, 3 clocks @133 MHz
Cycle Time (tRAS) : 5 clocks @100 MHz, 6 clocks @133 MHz

Memory at Slot 2:
Manufacturer : Nanya Technology
Part Number : NT256D64S88B1G-6K
Serial Number : 13150807
Type : DDR-SDRAM PC-2700 (166 MHz) - [DDR-333]
Size : 256 MB (1 rows, 4 banks)
Supported Frequencies : 133 MHz, 166 MHz
CAS Latency (tCL) : 2 clocks @133 MHz, 2.5 clocks @166 MHz
RAS to CAS (tRCD) : 3 clocks @133 MHz, 3 clocks @166 MHz
RAS Precharge (tRP) : 3 clocks @133 MHz, 3 clocks @166 MHz
Cycle Time (tRAS) : 6 clocks @133 MHz, 7 clocks @166 MHz

Memory Information/ Chipset :
Type : DDR-SDRAM PC1600
Frequency : 100.9 MHz
DRAM/FSB Ratio : 1/1 THIS RATIO VALUE IS 4/3 WHEN REBOOTED
making memory speed: 4/3 x 100.9 Mhz = 133Mhz, which means 266 MHz DDR
speed. No I'm going with 200 MHz, which isn't nice.

Activated Channels : Single
ECC Diagnostic : No
CAS Latency (tCL) : 2.5 clocks
RAS to CAS (tRCD) : 3 clocks
RAS Precharge (tRP) : 3 clocks
Cycle Time (tRAS) : 6 clocks
DRAM Idle Timer : 64 clocks
Shared Memory (video) : Yes

Any help? Thanks,
Asko
 
J

Jim Macklin

You really should have more RAM, why not buy two identical
512 PC2700 or even better, buy 1 PC3200 1 GB.



|I have a problem in DDR memory speed. When I do a normal
cold boot the
| memory speed is 200 MHz but when the computer is rebooted
from the
| Start-menu ( "Start Again" I guess that might be in an
English version of
| Windows XP) the memory speed is 266 MHz, which would be
the figure I would
| it like to be. The hardware specs allow that.
|
| I get information of the system through a couple of
programs such as cpu-z,
| Sisoftware Sandra and PCWizard 2006. Here is listed
information about this
| issue. There is two DDR DIMMs 256 MB each. One is PC-2100,
the other
| PC-2700. My question is why the memory speed is lower
when cold booted and
| what to do in order to get the higher speed out without
rebooting? Take a
| look at the asynchronous memory speed multiplier, which is
1/1 when cold
| booted, but 4/3 when rebooted and it should be that 4/3.
|
| Mainboard : Asus P4GE-FSC
| Chipset : Intel i845GE
| Processor : Intel Celeron @ 2600 MHz
| Physical Memory : 512 MB (2 x 256 DDR-SDRAM )
|
| NorthBridge Information :
| Architecture : HUB
| Manufacturer : Intel (ASUSTeK Computer Inc)
| Codename : Brookdale-G
| Revision : B1
| Bus Speed : 100.9 MHz THIS BUS SPEED IS NOT TWEAKABLE
AT BIOS
| FSB Frequency : 403.6 MHz (QDR)
| FSB max. Support : 533 MHz
| RAM max. Support : DDR (333 MHz)
|
| Memory at Slot 1 :
| Manufacturer : Kingston
| Part Number : KVR266X64C25/256
| Type : DDR-SDRAM PC-2100 (133 MHz) - [DDR-266]
| Size : 256 MB (1 rows, 4 banks)
| Supported Frequencies : 100 MHz, 133 MHz
| CAS Latency (tCL) : 2 clocks @100 MHz, 2.5 clocks @133 MHz
| RAS to CAS (tRCD) : 2 clocks @100 MHz, 3 clocks @133 MHz
| RAS Precharge (tRP) : 2 clocks @100 MHz, 3 clocks @133 MHz
| Cycle Time (tRAS) : 5 clocks @100 MHz, 6 clocks @133 MHz
|
| Memory at Slot 2:
| Manufacturer : Nanya Technology
| Part Number : NT256D64S88B1G-6K
| Serial Number : 13150807
| Type : DDR-SDRAM PC-2700 (166 MHz) - [DDR-333]
| Size : 256 MB (1 rows, 4 banks)
| Supported Frequencies : 133 MHz, 166 MHz
| CAS Latency (tCL) : 2 clocks @133 MHz, 2.5 clocks @166 MHz
| RAS to CAS (tRCD) : 3 clocks @133 MHz, 3 clocks @166 MHz
| RAS Precharge (tRP) : 3 clocks @133 MHz, 3 clocks @166 MHz
| Cycle Time (tRAS) : 6 clocks @133 MHz, 7 clocks @166 MHz
|
| Memory Information/ Chipset :
| Type : DDR-SDRAM PC1600
| Frequency : 100.9 MHz
| DRAM/FSB Ratio : 1/1 THIS RATIO VALUE IS
4/3 WHEN REBOOTED
| making memory speed: 4/3 x 100.9 Mhz = 133Mhz, which means
266 MHz DDR
| speed. No I'm going with 200 MHz, which isn't nice.
|
| Activated Channels : Single
| ECC Diagnostic : No
| CAS Latency (tCL) : 2.5 clocks
| RAS to CAS (tRCD) : 3 clocks
| RAS Precharge (tRP) : 3 clocks
| Cycle Time (tRAS) : 6 clocks
| DRAM Idle Timer : 64 clocks
| Shared Memory (video) : Yes
|
| Any help? Thanks,
| Asko
|
|
 
P

peter

Take a look at your CAS latencys of both chips......2.5 gives you 133 on the
Kingston and 166 on the Nanya..that called a mismatch.Since you have 2
different speeds of Ram it will default to the lowest...CAS 2 and again you
have a mismatch.
Suggest you replace the Kingston with another stick of the same Nanya....and
if your doing that you might as well upgrade to more memory........say
2x512??
peter
Asko said:
I have a problem in DDR memory speed. When I do a normal cold boot the
memory speed is 200 MHz but when the computer is rebooted from the
Start-menu ( "Start Again" I guess that might be in an English version of
Windows XP) the memory speed is 266 MHz, which would be the figure I would
it like to be. The hardware specs allow that.

I get information of the system through a couple of programs such as
cpu-z, Sisoftware Sandra and PCWizard 2006. Here is listed information
about this issue. There is two DDR DIMMs 256 MB each. One is PC-2100, the
other PC-2700. My question is why the memory speed is lower when cold
booted and what to do in order to get the higher speed out without
rebooting? Take a look at the asynchronous memory speed multiplier, which
is 1/1 when cold booted, but 4/3 when rebooted and it should be that 4/3.

Mainboard : Asus P4GE-FSC
Chipset : Intel i845GE
Processor : Intel Celeron @ 2600 MHz
Physical Memory : 512 MB (2 x 256 DDR-SDRAM )

NorthBridge Information :
Architecture : HUB
Manufacturer : Intel (ASUSTeK Computer Inc)
Codename : Brookdale-G
Revision : B1
Bus Speed : 100.9 MHz THIS BUS SPEED IS NOT TWEAKABLE AT BIOS
FSB Frequency : 403.6 MHz (QDR)
FSB max. Support : 533 MHz
RAM max. Support : DDR (333 MHz)

Memory at Slot 1 :
Manufacturer : Kingston
Part Number : KVR266X64C25/256
Type : DDR-SDRAM PC-2100 (133 MHz) - [DDR-266]
Size : 256 MB (1 rows, 4 banks)
Supported Frequencies : 100 MHz, 133 MHz
CAS Latency (tCL) : 2 clocks @100 MHz, 2.5 clocks @133 MHz
RAS to CAS (tRCD) : 2 clocks @100 MHz, 3 clocks @133 MHz
RAS Precharge (tRP) : 2 clocks @100 MHz, 3 clocks @133 MHz
Cycle Time (tRAS) : 5 clocks @100 MHz, 6 clocks @133 MHz

Memory at Slot 2:
Manufacturer : Nanya Technology
Part Number : NT256D64S88B1G-6K
Serial Number : 13150807
Type : DDR-SDRAM PC-2700 (166 MHz) - [DDR-333]
Size : 256 MB (1 rows, 4 banks)
Supported Frequencies : 133 MHz, 166 MHz
CAS Latency (tCL) : 2 clocks @133 MHz, 2.5 clocks @166 MHz
RAS to CAS (tRCD) : 3 clocks @133 MHz, 3 clocks @166 MHz
RAS Precharge (tRP) : 3 clocks @133 MHz, 3 clocks @166 MHz
Cycle Time (tRAS) : 6 clocks @133 MHz, 7 clocks @166 MHz

Memory Information/ Chipset :
Type : DDR-SDRAM PC1600
Frequency : 100.9 MHz
DRAM/FSB Ratio : 1/1 THIS RATIO VALUE IS 4/3 WHEN
REBOOTED making memory speed: 4/3 x 100.9 Mhz = 133Mhz, which means 266
MHz DDR speed. No I'm going with 200 MHz, which isn't nice.

Activated Channels : Single
ECC Diagnostic : No
CAS Latency (tCL) : 2.5 clocks
RAS to CAS (tRCD) : 3 clocks
RAS Precharge (tRP) : 3 clocks
Cycle Time (tRAS) : 6 clocks
DRAM Idle Timer : 64 clocks
Shared Memory (video) : Yes

Any help? Thanks,
Asko
 
A

Asko

Jim Macklin said:
You really should have more RAM, why not buy two identical
512 PC2700 or even better, buy 1 PC3200 1 GB.

Not exactly an answer to my question. Yes, we all should have more this and
that. I go pretty well with this 512 MB total RAM. Here we don't have an
habit to buy just for the heck of it nor leave food on plate. Your
kilometriage may vary, though.

Next.

Asko
 
A

Asko

peter said:
Take a look at your CAS latencys of both chips......2.5 gives you 133 on
the Kingston and 166 on the Nanya..that called a mismatch.Since you have 2
different speeds of Ram it will default to the lowest...CAS 2 and again
you have a mismatch.
Suggest you replace the Kingston with another stick of the same
Nanya....and if your doing that you might as well upgrade to more
memory........say 2x512??

I was aware of the fact that it would be better if memory modules were
identical, at the moment when I purchased the extra module (Kingston, used,
20 e). My fault, I admit that. But, basically there should not be any
problem with these two memory modules put together in a computer. I said
basically.

CAS latency is 2.5 as you see in the info listing in my original posting. I
launched just cpu-z and it says: CAS latency is currently 2.5. RAS to CAS
Delay is 3, RAS Precharge is 3, Cycle Time (Tras) is 6. When compared these
values with the memory modules' specs no mismatch should be present. With
these values both memory modules should work at 133 MHz (266 MHz as a
DDR-value). But they do not. The multiplier is 1/1, not 4/3 as it should be.
They work now at 100 MHz (200 Mhz DDR speed).

Any idea about the fact that if I reboot the computer I get the speed 266
MHz? I have tried it several times and that change is a fact. The multiplier
changes to 4/3.

Asko
 
P

Paul

Asko said:
Not exactly an answer to my question. Yes, we all should have more this
and that. I go pretty well with this 512 MB total RAM. Here we don't
have an habit to buy just for the heck of it nor leave food on plate.
Your kilometriage may vary, though.

Next.

Asko

So your theory is, that the first time the memory
clock was 100MHz, and the second time the memory
clock was 133MHz ?

What I would do, is get a copy of memtest86+ from
memtest.org. Power on the computer and boot the computer
with the memtest86+ test floppy. There are some
memory bandwidth indicators on the upper left hand
corner of the screen.

Record the bandwidth number (the slowest of the three
is the main memory bandwidth).

Now, press escape to quit memtest86+. Leave the test
floppy in the floppy drive, and memtest86+ will boot
again. Since your theory is that the memory is different,
memtest86+ should report a faster memory bandwidth on
the second test.

If the test results are exactly the same, on the second
test, then you know that CPUZ is wrong.

I did some experiments with a different board, a motherboard
that uses an 845 chipset (Asus P4B motherboard). I got a
copy of WPCREDIT and dumped the memory controller map.
There is a bit in the 845, that controls whether the memory
ratio is 1:1 or 3:4. The Asus BIOS was doing the right thing,
but CPUZ got the wrong memory speed, when I did my tests.
So it could be that CPUZ still has a bug in it, when it comes
to the 845 Northbridge and reported memory speed. But your
memtest86+ test results should help you figure out whether
the speed is different or not.

So measure the memory bandwidth, as a way to confirm or
deny your theory.

Paul
 
A

Asko

Paul said:
So your theory is, that the first time the memory
clock was 100MHz, and the second time the memory
clock was 133MHz ?

Exactly. Those values can be read from the output of the beforementioned
analysing programs.
What I would do, is get a copy of memtest86+ from
memtest.org. Power on the computer and boot the computer
with the memtest86+ test floppy. There are some
memory bandwidth indicators on the upper left hand
corner of the screen.

Record the bandwidth number (the slowest of the three
is the main memory bandwidth).

Now, press escape to quit memtest86+. Leave the test
floppy in the floppy drive, and memtest86+ will boot
again. Since your theory is that the memory is different,
memtest86+ should report a faster memory bandwidth on
the second test.

If the test results are exactly the same, on the second
test, then you know that CPUZ is wrong.

I did some experiments with a different board, a motherboard
that uses an 845 chipset (Asus P4B motherboard). I got a
copy of WPCREDIT and dumped the memory controller map.
There is a bit in the 845, that controls whether the memory
ratio is 1:1 or 3:4. The Asus BIOS was doing the right thing,
but CPUZ got the wrong memory speed, when I did my tests.
So it could be that CPUZ still has a bug in it, when it comes
to the 845 Northbridge and reported memory speed. But your
memtest86+ test results should help you figure out whether
the speed is different or not.

So measure the memory bandwidth, as a way to confirm or
deny your theory.

Thanks for your posting. I will study it. Unfortunately this computer has no
floppy drive. ALL three programs that I have installed inform the same
values (cpu-z, Sandra, PC Wizard 2006). And all three inform the same better
values (speed and multiplier) when the computer is rebooted.

Asko
 
P

peter

But there is a CAS mismatch........2.5 on the Kingston is 133 on the Nanya
its 166....the Kingston cant run that high so the chipset reverts to CAS 2
which is 100 on the Kingston and 133 on the nanya...........confusion
reigns.
So the 1/1 multiplier gives you the lower(Kingston) value and the 4/3
multiplier somehow seems to only apply to the Nanya chip and you get
266...........how I dont know
peter
 
J

Jackie

Maby cmos bat is low, on boot defalts to 100 and for whatever reason to 133
on warm boot.
 
K

Kerry Brown

When you mix and match different speeds of RAM this can happen. Turn of any
auto settings for the RAM in your BIOS and configure the settings manually
for the slowest RAM you have installed.

--
Kerry Brown
Microsoft MVP - Shell/User
http://www.vistahelp.ca


Asko said:
I have a problem in DDR memory speed. When I do a normal cold boot the
memory speed is 200 MHz but when the computer is rebooted from the
Start-menu ( "Start Again" I guess that might be in an English version of
Windows XP) the memory speed is 266 MHz, which would be the figure I would
it like to be. The hardware specs allow that.

I get information of the system through a couple of programs such as
cpu-z, Sisoftware Sandra and PCWizard 2006. Here is listed information
about this issue. There is two DDR DIMMs 256 MB each. One is PC-2100, the
other PC-2700. My question is why the memory speed is lower when cold
booted and what to do in order to get the higher speed out without
rebooting? Take a look at the asynchronous memory speed multiplier, which
is 1/1 when cold booted, but 4/3 when rebooted and it should be that 4/3.

Mainboard : Asus P4GE-FSC
Chipset : Intel i845GE
Processor : Intel Celeron @ 2600 MHz
Physical Memory : 512 MB (2 x 256 DDR-SDRAM )

NorthBridge Information :
Architecture : HUB
Manufacturer : Intel (ASUSTeK Computer Inc)
Codename : Brookdale-G
Revision : B1
Bus Speed : 100.9 MHz THIS BUS SPEED IS NOT TWEAKABLE AT BIOS
FSB Frequency : 403.6 MHz (QDR)
FSB max. Support : 533 MHz
RAM max. Support : DDR (333 MHz)

Memory at Slot 1 :
Manufacturer : Kingston
Part Number : KVR266X64C25/256
Type : DDR-SDRAM PC-2100 (133 MHz) - [DDR-266]
Size : 256 MB (1 rows, 4 banks)
Supported Frequencies : 100 MHz, 133 MHz
CAS Latency (tCL) : 2 clocks @100 MHz, 2.5 clocks @133 MHz
RAS to CAS (tRCD) : 2 clocks @100 MHz, 3 clocks @133 MHz
RAS Precharge (tRP) : 2 clocks @100 MHz, 3 clocks @133 MHz
Cycle Time (tRAS) : 5 clocks @100 MHz, 6 clocks @133 MHz

Memory at Slot 2:
Manufacturer : Nanya Technology
Part Number : NT256D64S88B1G-6K
Serial Number : 13150807
Type : DDR-SDRAM PC-2700 (166 MHz) - [DDR-333]
Size : 256 MB (1 rows, 4 banks)
Supported Frequencies : 133 MHz, 166 MHz
CAS Latency (tCL) : 2 clocks @133 MHz, 2.5 clocks @166 MHz
RAS to CAS (tRCD) : 3 clocks @133 MHz, 3 clocks @166 MHz
RAS Precharge (tRP) : 3 clocks @133 MHz, 3 clocks @166 MHz
Cycle Time (tRAS) : 6 clocks @133 MHz, 7 clocks @166 MHz

Memory Information/ Chipset :
Type : DDR-SDRAM PC1600
Frequency : 100.9 MHz
DRAM/FSB Ratio : 1/1 THIS RATIO VALUE IS 4/3 WHEN
REBOOTED making memory speed: 4/3 x 100.9 Mhz = 133Mhz, which means 266
MHz DDR speed. No I'm going with 200 MHz, which isn't nice.

Activated Channels : Single
ECC Diagnostic : No
CAS Latency (tCL) : 2.5 clocks
RAS to CAS (tRCD) : 3 clocks
RAS Precharge (tRP) : 3 clocks
Cycle Time (tRAS) : 6 clocks
DRAM Idle Timer : 64 clocks
Shared Memory (video) : Yes

Any help? Thanks,
Asko
 
A

Asko

Jackie said:
Maby cmos bat is low, on boot defalts to 100 and for whatever reason to
133 on warm boot.

Thanks! I must check that. Never thought about that possibility.

Asko
 
A

Asko

Kerry Brown said:
When you mix and match different speeds of RAM this can happen. Turn of
any auto settings for the RAM in your BIOS and configure the settings
manually for the slowest RAM you have installed.

In this computer's BIOS only the four timing settings are manually
configurable. I have tried a variety of values taken from the specs of these
DDR memory sticks. It did not change anything. At the moment there are Auto
(SPD) -setting enabled. I must try again.

Asko
 

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.

Ask a Question

Top