>There is a firmware issue whereby mismatched DIMMs won't be configured correctly.
>It is to do with timings and certain quirks of certain DRAM chipsets.
When using 2 DIMMs the firmware should surely be able to use the settings of the slowest module?
>The current recommendation is to use identical pairs of memory (two 256MB sticks of the same brand etc.) or single sticks.
I tried it
one year ago with 2x 512MB sticks, same brand, same type, same timings, same chip count (16) and it did not work with Peg2
while they run perfectly in a dual-channel AthlonXP board at full DDR400 speed with SPD-provided access timings of 2,5-3-3.
Has this matter changed now and how is your little RAM laboratory project going?
If possible, I would like bplan to
certify/validate the RAM from
MDT.
They are located between Frankfurt and Oberhausen (*hinthint*) and have a good reputation.
Datasheets with timings etc. are also available.
(I used one of their 256MB "PC150" SDR DIMM in Pegasos I and now a single 512MB DDR400 DIMM in Pegasos II.)