On Dec 14, 2006, at 20:28 , Ron wrote:
> Can you do runs with just CFLAGS="-O3" and just CFLAGS="-msse2 -
> mfpmath=sse -funroll-loops -m64 - march=opteron -pipe" as well ?
All right. From my perspective, the effect of -O3 is significant,
whereas architecture-related optimizations have no statistically
significant effect. As far as I'm aware, though, there's no other
arch targets on the Opteron that will make sense, there being no
predecessor CPU instruction set to choose from; -march=pentium4
> Also, what HDs are you using? How many in what config?
I believe the volume is a two-drive RAID 1 configuration, but I'm not
managing these servers, so I'll ask the company's support people.
Interestingly enough I see that PostgreSQL seems to be writing around
1MB/s during the pgbench run, even though I'm running pgbench in the -
S mode. I haven't had the chance to look at the source yet; is it
really only doing selects?
In response to
pgsql-performance by date
|Next:||From: Kelly Burkhart||Date: 2006-12-14 23:43:51|
|Subject: Re: EXPLAIN ANALYZE on 8.2|
|Previous:||From: Tom Lane||Date: 2006-12-14 21:23:16|
|Subject: Re: Insertion to temp table deteriorating over time |