Tom,
On 12/9/05 2:14 PM, "Tom Lane" <[EMAIL PROTECTED]> wrote:
> Andrew Dunstan <[EMAIL PROTECTED]> writes:
>> Luke Lonergan wrote:
>>> It (the compute intensity optimization) is what we did for copy parsing, and
>>> it sped up by a factor of 100+.
>
>> The changes made to COPY were portable, though.
>
> In fact, the changes made to COPY had absolutely nada to do with any of
> the things discussed in this thread.
Yes, they do, you must not have read my post in this thread on compute
intensity and removing pipeline stalls.
> BTW, "sped up by 100%" (which is already an overstatement of what was
> actually accomplished) is a long way from "sped up by a factor of 100".
Wrong again - the code section that did parsing sped up by 100x, but the
overall improvement was *only* 100% in our version and 60% in your version.
- Luke
---------------------------(end of broadcast)---------------------------
TIP 9: In versions below 8.0, the planner will ignore your desire to
choose an index scan if your joining column's datatypes do not
match
