From: | Qingqing Zhou <zhouqq(at)cs(dot)toronto(dot)edu> |
---|---|
To: | Luke Lonergan <llonergan(at)greenplum(dot)com> |
Cc: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, Neil Conway <neilc(at)samurai(dot)com>, Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us>, pgsql-hackers(at)postgresql(dot)org |
Subject: | Re: Which qsort is used |
Date: | 2005-12-13 01:08:47 |
Message-ID: | Pine.LNX.4.58.0512122000040.3553@eon.cs |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
On Mon, 12 Dec 2005, Luke Lonergan wrote:
>
> Do you have a test suite you can recommend with those edge cases?
>
I have at least those factors in mind:
+ f1: number of elements
- in {10^3, 10^4, 10^5, 10^6, 10^7}
+ f2: key comparators measured by cpu cost
- in {1, 10, 100+};
+ f3: data distribution
- in {uniform, Gussian, 95% sorted, 95% reverse sorted}
+ f4: data value range
- in {2, 32, 1024, unlimited}: radix sort might be better for small
range
The element size doesn't matter since the main usage of our qsort is
on pointer array. Element data type is covered by f2 and f4.
This will gives us a 5*3*4*4 = 240 tests ...
Regards,
Qingqing
From | Date | Subject | |
---|---|---|---|
Next Message | Tom Lane | 2005-12-13 01:14:11 | Re: Anyone for adding -fwrapv to our standard CFLAGS? |
Previous Message | Michael Fuhr | 2005-12-13 01:07:51 | Re: How much expensive are row level statistics? |