Wednesday Mar 18, 2015

Re: Exadata X4 vs. X5 DB node CPU comparison

This is a follow-up to my past entry.

Exadata X4 vs. X5 DB node CPU comparison (INOUE Katsumi @ Tokyo)

Now with Database In-Memory, DB node's CPU and memory specs more directly influences DB performance.

Firstly, HW spec is not the only change going from X4-2 -> X5-2.
I heard default kernel parameter is changed which will influence performance.
There may be kernel code change for performance, too.

Secondly, our product manager wrote about related info here.

Do I still need Exadata now that I have Oracle Database In-Memory? (Oracle Database In-Memory)

  • You get a complete fault-tolerant In-Memory solution



Tuesday Mar 17, 2015

Exadata 1000th column challenge

This is a continuation of my past entry.

Database In-Memory accessing 1000th column (INOUE Katsumi @ Tokyo)

"Buffer Cache" result was close to Charles's result and In-Memory result matched what we all expect from "columnar" database.

I ran same script twice on Exadata. 1st with default setting and 2nd with "hidden parameter" described here.

Forcing Smart Scans on Exadata – is the serialdirect_read parameter safe to use in production? | Tanel Poder's Performance & Troubleshooting blog

direct path reads (and thus also Smart Scans) don’t sometimes kick in when running full scans in serial sessions.

This is the result. YMMV

I also changed the perl one-liner a little readable.

$ p-e-r-l -ne '
$c=5 if $_ =~ /^ *C *$/;
--$c;
if ($c==0)
  {++$d; $_ =~ /00:00:(.*)/; $total+=$1;
    if($d==10)
      {printf "%f\n", $total/10;$d=0;$total=0}
  }
' COLUMN_ORDER2.TXT
1.157000
1.280000
About

The views expressed on this blog are my own and do not necessarily reflect the views of Oracle.

Search

Archives
« March 2015
SunMonTueWedThuFriSat
1
2
3
5
6
7
8
9
14
15
21
22
24
25
26
27
28
29
30
31
    
       
Today