> The BS layman's speak they gave me at the store was that they had seen a lot
> of cases where having "double sided SIMMs" (they were oh-so-usefully
> classifying the memory based on whether there were chips on just one side,
> or on both sides of the circuit card 8-) in the second and subsequent slots
> never worked.
This is actually a perfectly resonable thing to do. Twice the number of
chips (and for double-sided modules with stacked chips four times) will
load the address bus with twice (our four times) the capacitance. Thus the
siganl flanks will become less and less well defined when adding more
chips on the bus until some access patterns start to give bad results.
Some motherboards, mostly servers, specify registered memories that have a
buffer circuit to greatly reduce the load on the line drivers. More
serious technical details from at least some motherboards will specify
that only a maximum number of chip-loads are possible. Since there are
only a few address bus driver chipsets the same limitations apply to most
motherboards. The manefacturer can play some games with voltages and
buffer strentghs but only so much. At high speeds and several stacked-chip
modules you really need registered modules.
Peter
-- Peter Svensson ! Pgp key available by finger, fingerprint: <petersv@psv.nu> ! 8A E9 20 98 C1 FF 43 E3 07 FD B9 0A 80 72 70 AF ------------------------------------------------------------------------ Remember, Luke, your source will be with you... always...
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/