> - Takes full advantage of large (>4GB) memory configurations on
> numerous 64 bit platforms including AMD Opteron, HP/Compaq Alpha,
> Sun UltraSPARC, MIPS, PA-RISC, and RS6000*
This bullet is really lacking punch now. 64 bit CPUs aren't used
_just_ for additional memory access. I didn't look too closely to see
when they came in (before or after this email), but there were a few
quotes and bits about 64bit platforms that had better ring than this.
As it stands, I read, "we run on wherever GCC and the OS can run,"
which is true, but neglects that it's somewhat battle proven on
Opteron and the other CPU's/OS'es mentioned. We really do _support_
Opteron in the sense that there are #ifdef's in our code to make sure
that performance on Opteron doesn't suck(tm).
> - Adds support for the AMD Opteron, increasing the list of 64bit
> platforms supported. Supported 64bit platforms include HP/Compaq
> Alpha, Sun UltraSPARC, MIPS, PA-RISC, RS6000*, and AMD Opteron.
> - New wire protocol (version 3) increases the speed of data transfers
Did this ever get verified? The reason I'm pretty sure this is true
is there is a lack of need to encode/decode data off the wire.
src/backend/libpq/pqformat.c:
* Note in particular the distinction between "raw data" and "text"; raw data
* is message protocol characters and binary values that are not subject to
* character set conversion, while text is converted by character encoding
* rules.
Lastly, as someone else said, more quotes would be good if this is
going to mainstream/conventional media sources and not just Open
Source press outlets. I may have a connection to the Seattle Times,
fwiw... still exploring that avenue though... -sc
--
Sean Chittenden