WinXP sees 1.5 GB memory as 1GB: Solution?

  • Thread starter Thread starter htchess
  • Start date Start date
htchess said:
Thanks, Paul, for the very detailed eloboration.

I have one more point. I am from Antalya, Turkey. Yesterday,
I queried a few computer stores. They say Kingston RAMs
dominate the market, in 1 GB RAM, Plus some OEM RAMs
with lower price. Other memory modules from MT, APACER,
CORSAIR, ELPIDA, INFINEON, NANYA, GEIL, ELIXIR and
VDATA are virtually absent on the market. Now, my questions:

1. Specifically, which 1 GB RAM should I purchase totalling
1.5 GB on my computer? Or

2. Are two 512 MB modules totalling 1 GB better to employ?
In this case, which model should I purchase?

Best regards,
HT
Click to expand...

The first source of information, would be the motherboard manual.
The Foxconn P4M800P7MB manual has no advice on DIMM selection.

If I select an MSI motherboard based on P4M800-Pro chipset, this
is the document MSI provides with information on memory choices.
The problem with a document like this, is the memory products listed
may not be for sale where you are. Some of the documents I've
looked at like this one, all the memories are no longer for sale.
So, while some users consult these documents rigorously, the memory
document is not always that practical.

http://www.msi.com.tw/html/products/mainboard/testreport_pdf/7222/memory.pdf

My selection algorithm would be as follows:

1) Figure out which memory speeds the board supports. Sometimes
that can be harder to figure out than it should be.
2) Find a vendor who has customer reviews. For example, here are
some reviews for a Geil 1GB module

The reviews are not very good for this stick. In addition, the product
is apparently only for use in Athlon64 AM2 systems.

http://www.newegg.com/Product/CustratingReview.asp?Item=N82E16820144049

This is another stick of Geil RAM. Here there is only one problem report
out of sixteen.

http://www.newegg.com/Product/CustratingReview.asp?Item=N82E16820144038

So, what you want, is a memory product of the right speed (and in
your case density), and the product should have good reviews.
Also, note in the specification for the second product, it says "1.8V".
DDR2 memory is supposed to operate at 1.8V. Some memory products have
a higher minimum than the 1.8V number. For example, some overclocker
RAM might be rated at 2.0-2.1V. The reason that can be a problem, is if
the motherboard you are using, uses closer to the standard value of
voltage, the motherboard may not start up when the memory is installed.

If Kingston is available, it is as good as any. As for what speed or timings
to buy, you could match what you have currently for the 512MB stick.

I don't think your board is intended to go too high in speed on the memory,
and perhaps DDR2-533 is as far as it would go. This MSI manual notes that
the memory speed is independent of the processor FSB, so that is not a
problem. You could use DDR2-533 with any choice of processor FSB.

http://us1.msi.com.tw/support/mnu_exe/mbd_mnu/M7222v2.0.zip

And no one seemed to have any problems with this Kingston 1GB DDR2-533. It
is $85 USD.

Kingston KVR533D2N4/1G (1.8V) - customer reviews
http://www.newegg.com/Product/CustratingReview.asp?Item=N82E16820144151

Oh, I did do a little more research into the memory chips. Apparently
64Mx8 chips are 14 x 10 x 4_internal_banks, while the next larger chips
at 128Mx8 are 14 x 10 x 8_internal_banks. So I should have been looking
more closely at the "bank" field (byte number 17) in the SPD.

I still haven't been able to find any more good info on the P4M800 Pro
chipset. The VIA site has conflicting information, and various manuals
aren't being very helpful either. So for the time being, the "high density"
issue remains an unproven theory. Good luck with the next 1GB stick you
buy.

As far as I know, your motherboard contains a single memory channel, so
there wouldn't be a speed advantage from using 2x512MB versus 512MB + 1GB.
On many other motherboards, there would be an advantage to using 2x512MB
(i.e. dual channel mode), and on the Intel chipset motherboards, you might
have four DIMM sockets to work with. That is a bit more convenient for
future memory growth.

HTH,
Paul
 
The Foxconn P4M800P7MB manual has no advice on DIMM selection.

If I select an MSI motherboard based on P4M800-Pro chipset, this
is the document MSI provides with information on memory choices.
The problem with a document like this, is the memory products listed
may not be for sale where you are. Some of the documents I've
looked at like this one, all the memories are no longer for sale.
So, while some users consult these documents rigorously, the memory
document is not always that practical.

http://www.msi.com.tw/html/products/mainboard/testreport_pdf/7222/mem...

My selection algorithm would be as follows:

1) Figure out which memory speeds the board supports. Sometimes
that can be harder to figure out than it should be.
2) Find a vendor who has customer reviews. For example, here are
some reviews for a Geil 1GB module

The reviews are not very good for this stick. In addition, the product
is apparently only for use in Athlon64 AM2 systems.

http://www.newegg.com/Product/CustratingReview.asp?Item=N82E16820144049

This is another stick of Geil RAM. Here there is only one problem report
out of sixteen.

http://www.newegg.com/Product/CustratingReview.asp?Item=N82E16820144038

So, what you want, is a memory product of the right speed (and in
your case density), and the product should have good reviews.
Also, note in the specification for the second product, it says "1.8V".
DDR2 memory is supposed to operate at 1.8V. Some memory products have
a higher minimum than the 1.8V number. For example, some overclocker
RAM might be rated at 2.0-2.1V. The reason that can be a problem, is if
the motherboard you are using, uses closer to the standard value of
voltage, the motherboard may not start up when the memory is installed.

If Kingston is available, it is as good as any. As for what speed or timings
to buy, you could match what you have currently for the 512MB stick.

I don't think your board is intended to go too high in speed on the memory,
and perhaps DDR2-533 is as far as it would go. This MSI manual notes that
the memory speed is independent of the processor FSB, so that is not a
problem. You could use DDR2-533 with any choice of processor FSB.

http://us1.msi.com.tw/support/mnu_exe/mbd_mnu/M7222v2.0.zip

And no one seemed to have any problems with this Kingston 1GB DDR2-533. It
is $85 USD.

Kingston KVR533D2N4/1G (1.8V) - customer reviewshttp://www.newegg.com/Product/CustratingReview.asp?Item=N82E16820144151
Click to expand...

I am just back from computer store! Thank you very much
for the very thorough write-up again.

Well, I just got mine. Kingston, 1 GB, 16 chips as you wrote.
120 USD. This means the sum of transportation and distribution
costs are about 50% :( They said that this memory module has
lifetime warranty.

We tried another OEM memory module before I bought this one.
We saw the same problem: computer sees 1 GB, ie 1024 MB
memory, although 1.5 GB physical memory installed. And next,
we tried with Kingston, well, finally saw 1536 MB while booting
and later in desktop with right click on my computer icon.


Excerpt from FreshDiagnose:

Memory Device
Memory Array Handle $1C
Memory Error Info Handle $FFFE
Total Width 65535 Bits
Data Width 65535 Bits
Memory Size 1024 MB
Form Factor DIMM
Device Set Not part of a set
Device Locator A0
Bank Locator Bank0/1
Memory Type Unknown
Memory Speed Unknown
Manufacturer None
Serial Number None
Asset Tag None
Part Number None

Memory Device
Memory Array Handle $1C
Memory Error Info Handle $FFFE
Total Width 65535 Bits
Data Width 65535 Bits
Memory Size 512 MB
Form Factor DIMM
Device Set Not part of a set
Device Locator A1
Bank Locator Bank2/3
Memory Type Unknown
Memory Speed Unknown
Manufacturer None
Serial Number None
Asset Tag None
Part Number None

I understand from above instead of four slots with 8 chips memory
modules,
mainboard is designed to utilize two slots with 16 chips memory
modules.

Hmmmm.

This is from Everset Ultimate:

BilgiAlaný Deðer
Fiziksel Bellek
Toplam 1534 MB
Kullanýlan 577 MB
Boþ 956 MB
Kullaným 38 %

Takas Alaný
Toplam 2152 MB
Kullanýlan 425 MB
Boþ 1726 MB
Kullaným 20 %

Sanal Bellek
Toplam 3687 MB
Kullanýlan 1003 MB
Boþ 2683 MB
Kullaným 27 %

Physical Address Extension (PAE)
Ýþletim Sistemi Destekli Evet
Ýþlemci Destekli Evet
Etkin Evet

Well, I am happy now.

Thank you all.

Best regards,
HT
 
Good to see it is all sorted out :-)

Don't forget to test the memory with Memtest86+ (memtest.org)
and Prime95 "Torture Test". Prime95 is a pretty good test
that your computer is healthy (download from mersenne.org).

Paul
 
htchess wrote: >

Good to see it is all sorted out :-)

Don't forget to test the memory with Memtest86+ (memtest.org)
and Prime95 "Torture Test". Prime95 is a pretty good test
that your computer is healthy (download from mersenne.org).

Paul
Click to expand...

Oh!, incredible mersenne numbers, just another prime numbers.
20 to 30 days 24-h-full-run, and we possibly get another bloody
mersenne number!

I carried out two benchmarks:

[Tue Jan 30 17:51:27 2007]
Compare your results to other computers at http://www.mersenne.org/
bench.htm
That web page also contains instructions on how your results can be
included.

Intel(R) Pentium(R) 4 CPU 3.00GHz
CPU speed: 2987.33 MHz
CPU features: RDTSC, CMOV, Prefetch, MMX, SSE, SSE2
L1 cache size: 16 KB
L2 cache size: 1024 KB
L1 cache line size: 64 bytes
L2 cache line size: 128 bytes
TLBS: 64
Prime95 32-bit version 24.14, RdtscTiming=1
Best time for 512K FFT length: 16.779 ms.
Best time for 640K FFT length: 21.447 ms.
Best time for 768K FFT length: 25.998 ms.
Best time for 896K FFT length: 32.098 ms.
Best time for 1024K FFT length: 36.268 ms.
Best time for 1280K FFT length: 45.102 ms.
Best time for 1536K FFT length: 54.317 ms.
Best time for 1792K FFT length: 65.747 ms.
Best time for 2048K FFT length: 73.505 ms.
Best time for 2560K FFT length: 96.680 ms.
Best time for 3072K FFT length: 115.942 ms.
Best time for 3584K FFT length: 141.267 ms.
Best time for 4096K FFT length: 156.048 ms.
Best time for 58 bit trial factors: 9.369 ms.
Best time for 59 bit trial factors: 9.391 ms.
Best time for 60 bit trial factors: 9.337 ms.
Best time for 61 bit trial factors: 9.327 ms.
Best time for 62 bit trial factors: 13.012 ms.
Best time for 63 bit trial factors: 13.031 ms.
Best time for 64 bit trial factors: 14.790 ms.
Best time for 65 bit trial factors: 14.958 ms.
Best time for 66 bit trial factors: 14.839 ms.
Best time for 67 bit trial factors: 14.701 ms.
[Tue Jan 30 17:57:13 2007]
Compare your results to other computers at http://www.mersenne.org/
bench.htm
That web page also contains instructions on how your results can be
included.

Intel(R) Pentium(R) 4 CPU 3.00GHz
CPU speed: 2987.39 MHz
CPU features: RDTSC, CMOV, Prefetch, MMX, SSE, SSE2
L1 cache size: 16 KB
L2 cache size: 1024 KB
L1 cache line size: 64 bytes
L2 cache line size: 128 bytes
TLBS: 64
Prime95 32-bit version 24.14, RdtscTiming=1
Best time for 512K FFT length: 16.699 ms.
Best time for 640K FFT length: 21.399 ms.
Best time for 768K FFT length: 26.112 ms.
Best time for 896K FFT length: 32.103 ms.
Best time for 1024K FFT length: 36.556 ms.
Best time for 1280K FFT length: 44.989 ms.
Best time for 1536K FFT length: 54.265 ms.
Best time for 1792K FFT length: 65.406 ms.
Best time for 2048K FFT length: 73.195 ms.
Best time for 2560K FFT length: 96.152 ms.
Best time for 3072K FFT length: 115.817 ms.
Best time for 3584K FFT length: 140.031 ms.
Best time for 4096K FFT length: 156.978 ms.
Best time for 58 bit trial factors: 9.454 ms.
Best time for 59 bit trial factors: 9.424 ms.
Best time for 60 bit trial factors: 9.278 ms.
Best time for 61 bit trial factors: 9.413 ms.
Best time for 62 bit trial factors: 12.983 ms.
Best time for 63 bit trial factors: 13.041 ms.
Best time for 64 bit trial factors: 14.861 ms.
Best time for 65 bit trial factors: 14.898 ms.
Best time for 66 bit trial factors: 14.855 ms.
Best time for 67 bit trial factors: 14.875 ms.

Two more questions:

1. How stand these numbers?

2. Is increasing L1 and L2 cache sizes possible? if yes how?
What I saw from the benchmark page. The higher L1 and L2
cache sizes, the better the performance.

Best regards,
HT
 
Second test was the torture test:

Small FFT 7 min 0 err 0 warn
In-place large FFT 3 min 0 err 0 warn
Blend 7 min 0 err 0 warn

A 6-hours-run is suggested. But as I understand
the text, few minutes are also enough.

Best regards,
HT
 
Second test was the torture test:

Small FFT 7 min 0 err 0 warn
In-place large FFT 3 min 0 err 0 warn
Blend 7 min 0 err 0 warn

A 6-hours-run is suggested. But as I understand
the text, few minutes are also enough.

Best regards,
HT
Click to expand...

well, actually, another 4-hours-torture-test of my computer during and
after:

keyboard-lock;
software-mismatch;
unloading programs;
deleting temp files;
loading drivers again;
updating system;

I am more happy than ever!

Best regards,
HT
 
Back
Top