| rounding issues ? [message #73768] |
Mon, 12 June 2006 05:39  |
Eclipse User |
|
|
|
This is a multi-part message in MIME format.
--------------020706020704010006060909
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
one more from me:
I'm currently writing a xslt to extract info from *.trcxml files.
As a reference, I use the xml produced by the "Report > Export as XML
report" action.
Now I'm wondering that I have some rounding issues when calculation e.g.
cumulative time.
Example (from *.trcxml):
<methodEntry (...) time="1149838796,296922922" methodIdRef="229896"
ticket="4147030" stackDepth="13"/>
<methodExit (...) time="1149838796,310027122" methodIdRef="229896"
ticket="4147030" overhead="0.000292588"/>
This should result in a cumulative time of 0.0131042 (or 0.012811612 w/o
overhead) - at least according to my Windows calculator ;-)
But the XML exported by TPTP says: cumulativeTime="0.012906737001342773"
Am I missing some nanoseconds in my calculation somewhere, or should
this be considered normal rounding issues?
thanks again,
Florian
--------------020706020704010006060909
Content-Type: text/x-vcard; charset=utf-8;
name="Florian.Georg.vcf"
Content-Transfer-Encoding: 7bit
Content-Disposition: attachment;
filename="Florian.Georg.vcf"
begin:vcard
fn:Florian Georg
n:Georg;Florian
org:Innovations Softwaretechnologie GmbH;Product Development
adr:;;Ziegelei 7;Immenstaad;;88090;Germany
email;internet:Florian.Georg@innovations.de
title:Developer
tel;work:+49-7545-202-512
url:http://www.innovations.de
version:2.1
end:vcard
--------------020706020704010006060909--
|
|
|
| Re: rounding issues ? [message #73879 is a reply to message #73768] |
Mon, 12 June 2006 10:05  |
| Eclipse User |
|
|
|
Originally posted by: nmehrega.ca.ibm.com
This is a multipart message in MIME format.
--=_alternative 004D63558525718B_=
Content-Type: text/plain; charset="US-ASCII"
Hi Florian,
>I'm currently writing a xslt to extract info from *.trcxml files.
>As a reference, I use the xml produced by the "Report > Export as XML
>report" action.
There are easier ways to get to the XML Trace file. Please see section
5.0 of the following document:
http://www.eclipse.org/tptp/platform/documents/speed_profili ng/speedUpProfiling.html
Also, please consider donating your XSLT to the open source community when
completed.
> This should result in a cumulative time of 0.0131042 (or 0.012811612 w/o
> overhead) - at least according to my Windows calculator ;-)
Does 100th of a second really matter when profiling? This could be a
rounding problem on our part, but it's a minor issue. Developers don't
really care whether their method takes 100th of a second more to execute
so I don't think you should really invest time and effort trying to
resolve these rounding issues.
Navid Mehregani
--=_alternative 004D63558525718B_=
Content-Type: text/html; charset="US-ASCII"
<br><tt><font size=2>Hi Florian,</font></tt>
<br>
<br><tt><font size=2>>I'm currently writing a xslt to extract info from
*.trcxml files.<br>
>As a reference, I use the xml produced by the "Report > Export
as XML <br>
>report" action.</font></tt>
<br>
<br><font size=2 face="sans-serif">There are easier ways to get to the
XML Trace file. Please see section 5.0 of the following document:
http://www.eclipse.org/tptp/platform/documents/speed_profili ng/speedUpProfiling.html</font>
<br><font size=2 face="sans-serif">Also, please consider donating your
XSLT to the open source community when completed.</font>
<br>
<br><tt><font size=2>> This should result in a cumulative time of 0.0131042
(or 0.012811612 w/o <br>
> overhead) - at least according to my Windows calculator ;-)<br>
</font></tt>
<br><font size=2 face="sans-serif">Does 100th of a second really matter
when profiling? This could be a rounding problem on our part, but
it's a minor issue. Developers don't really care whether their method
takes 100th of a second more to execute so I don't think you should really
invest time and effort trying to resolve these rounding issues.</font>
<br>
<br><font size=2 face="sans-serif">Navid Mehregani</font>
--=_alternative 004D63558525718B_=--
|
|
|
Powered by
FUDForum. Page generated in 0.02596 seconds
Home