| From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
|---|---|
| To: | Manfred Spraul <manfred(at)colorfullife(dot)com> |
| Cc: | "pgsql-hackers(at)postgresql(dot)org" <pgsql-hackers(at)postgresql(dot)org> |
| Subject: | Re: Memory buffer alignment |
| Date: | 2003-09-18 21:46:34 |
| Message-ID: | 3149.1063921594@sss.pgh.pa.us |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
Manfred Spraul <manfred(at)colorfullife(dot)com> writes:
> Unfortunately it's the slower one: Intel cpus have a microcode fastpath
> for memcopies that are 8-byte aligned. This fastpath is around 50%
> faster than the manual copy that is used for "misaligned" (i.e. only
> 4-byte aligned) pointers.
Maybe it'd be worth setting MAXIMUM_ALIGNOF to 8 on such CPUs? Or at
least hacking ShmemAlloc and friends to use 8-byte alignment. I assume
the major issue here is that the shared buffers don't get 8-byte-aligned
within the shared memory segment.
Are there any machines where it'd be worth forcing an even larger
alignment for the buffers?
regards, tom lane
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Gaetano Mendola | 2003-09-18 22:17:58 | Re: 7.4beta2 vs 7.3.3 |
| Previous Message | Carlos Guzman Alvarez | 2003-09-18 21:39:56 | Array Parameters on protocol 3.0 |