Patch Factory - Full-featured byte-level patching and version management system for efficient software updating. Patch Factory - Full-featured byte-level patching and version management system for efficient software updating.
Home | Products | Downloads | Order/Purchase | Support | Forums




Pages:123 ( Go to first unread post ) Reply to this topicStart new topicStart Poll

> Difference in time between Demo tool & Dll
sue.biggs
  Posted: April 24, 2008 04:47 am
Reply to this POSTQuote Post


Member
*

Group: DF SDK Users
Posts: 4
Member No.: 167
Joined: April 11, 2008




Hi,
We recently purchased your DF SDK tool and we modified the sample example that was included. We tweaked it to make a differential file using byte-to-byte rather than smart difference. This generates us really nice & compact file [thanks for this], but we are having major concern with the time.
When we download the demo tool from this website and run with only smart difference to byte-by-byte difference on a set of file, we can generate the df file in 15 mins. This is quite good considering that the files are almost 1.7GB in size.

When we use the dll & the sample project included in our purchase on the same machine and on the same set of files, we are able to generate the files in 45 mins. I am not sure why the difference in time as much as 3.2x. Can you please help what we can do to make it run quicker?

I have tried to set the assume file as different flags in the sample example project along with no compression, where as the demo tool does compare and compress. This change has reduced the time by 5 - 7 mins on an average, but it is still far away than 15 mins that the demo tool takes with these options turned on.

Once thing that I noticed though, is that the file generated by the sample example is mostly half the size of the df file generated by the demo tool. Though this is greatly appreciated as it is a smaller file, yet we would like more speed rather than file size.

Here are the only codes that I have added / modified in the original sample example project.
CODE

TDfCmpOptions cmp_opt;
cmp_opt.iContentCmp = 1;
cmp_opt.iContentUpdateData = DF_CNTUPD_B2B_DIFF;
cmp_opt.iFileNameCmp = -1;
cmp_opt.iDateTimeCmp = 1;
cmp_opt.iAttrCmp = 1;
cmp_opt.iEqSearchMethod = DF_METHOD_FAST;


CODE

TDfBldOptions bld_opt;
 //Temp - Set compression to null
  bld_opt.iCompression = 0;
 //Temp End - Set compression to null


I would be gratefulf for your help. unsure.gif

This post has been edited by sue.biggs on April 24, 2008 04:47 am
PMEmail Poster
Top
Admin
Posted: April 28, 2008 12:33 am
Reply to this POSTQuote Post


Support Team
*******

Group: Forum Administrators
Posts: 102
Member No.: 1
Joined: March 30, 2004




To achieve the maximum speed you try setting the following parameters:
CODE

iContentUpdateData=DF_CNTUPD_SMART_DIFF
iEqSearchMethod=DF_METHOD_FASTEST
iMaxMemoryUsageUnit=0
nMaxMemoryUsage=80

Notes:
1. iEqSearchMethod defines the method of identical file content parts search process, in case of iContentUpdateData=DF_CONTENT_SMART_DIFF. Otherwise field value is ignored.
2. nMaxMemoryUsage defines max RAM size which can be used during identical content parts search process (actual only if iContentUpdateData=DF_CONTENT_SMART_DIFF).

However the size of the result patch-file and the speed of its building depends not only of the parameters of the comparing algorithm but also on the nature and the similarity ratio of files to be compared.


--------------------
Best regards, AgenSoft.Com Support Team
AgenSoft.Com - Complete Software Update Solutions
PMEmail PosterUsers Website
Top
sue.biggs
  Posted: April 28, 2008 02:49 am
Reply to this POSTQuote Post


Member
*

Group: DF SDK Users
Posts: 4
Member No.: 167
Joined: April 11, 2008




Hi,
Thanks for your reply. I agree that the patch size and time depends on the ratio of similar content in a file and also on system & algorithm. But why is there such a massive difference in time when same algorithm is used, on the same files. I am trying the demo tool and your sample project on the same set of files, one after the other. And the time difference is up to 3 times mostly. Can you please help us why such a major difference in time with the same files, algorithm etc.

Thanks.

This post has been edited by sue.biggs on April 28, 2008 04:48 am
PMEmail Poster
Top
nwakerrDem
Posted: June 08, 2015 04:51 am
Reply to this POSTQuote Post


Member
*

Group: Members
Posts: 1
Member No.: 5416
Joined: June 04, 2015




Hooray. Time to get ripped of again, & this time Waveney are having a laugh
Anybody know of good offers lately? Here's hoping....

link


--------------------
PMEmail PosterUsers WebsiteICQAOLYahoo
Top
SpringCc
Posted: June 14, 2015 08:55 pm
Reply to this POSTQuote Post


Advanced Member
**

Group: Members
Posts: 17
Member No.: 5414
Joined: May 27, 2015




Thanks for your reply. I agree that the patch size and time depends on the ratio of similar content in a file and also on system & algorithm. But why is there such a massive difference in time when same algorithm is used, on the same files. I am trying the demo tool and your sample project on the same set of files, one after the other. And the time difference is up to 3 times mostly. Can you please help us why such a major difference in time with the same files, algorithm etc.

meizu m2 note
meizu m1 note
OPPO Find 5
meizu mx5
PMEmail Poster
Top
futglory
Posted: November 24, 2016 02:15 am
Reply to this POSTQuote Post


Member
*

Group: Members
Posts: 3
Member No.: 5649
Joined: November 24, 2016




Fifa 17 Points Account Additionally feature stomped when Gunners additionally Space looked like there was just for midweek Champions Types activities a difficulty improbable destabilized or merely sacrificed approximately Jose Mourinhos a fabulous important righthand individual.
fut packs Combined with up to now cloth nearly amongst eachother, Karanka listed connected to Chelsea.
PMEmail Poster
Top
Nedforspidlouch
Posted: January 16, 2017 01:49 pm
Reply to this POSTQuote Post


Member
*

Group: Members
Posts: 8
Member No.: 5846
Joined: January 12, 2017




Hooray. Time to get ripped of again, & this time Waveney are having a laugh
Anybody know of good offers lately? Here's hoping....

http://nedforspid.moy.su/
PMEmail Poster
Top
MaryAnnBef
Posted: January 19, 2017 02:14 am
Reply to this POSTQuote Post


Member
*

Group: Members
Posts: 3
Member No.: 6011
Joined: January 18, 2017




Im interested in the Primo Ramdisk Professional Edition
but my question is what is the difference between Business License & personal licence.
I work from home on occasion. so its also my personal computer.
is there any advantage to having a business license?

Thanks.


--------------------
PMEmail PosterUsers WebsiteICQAOLYahoo
Top
HarriettBef
Posted: January 27, 2017 08:38 am
Reply to this POSTQuote Post


Member
*

Group: Members
Posts: 6
Member No.: 6078
Joined: January 19, 2017




Im interested in the Primo Ramdisk Professional Edition
but my question is what is the difference between Business License & personal licence.
I work from home on occasion. so its also my personal computer.
is there any advantage to having a business license?

Thanks.


--------------------
PMEmail PosterUsers WebsiteICQAOLYahoo
Top
RafaelNuh
Posted: March 09, 2017 12:46 pm
Reply to this POSTQuote Post


Member
*

Group: Members
Posts: 5
Member No.: 7663
Joined: March 01, 2017




Dear JS-Support,
What is the planned difference features between debug and release builds ?

Br,
P.
PMEmail PosterICQAOLYahoo
Top

Topic OptionsPages:123 Reply to this topicStart new topicStart Poll

 




[ Home ]   [ About Us ]   [ News ]   [ Awards ]   [ PatchFactory ]   [ ScreenShots ]   [ Downloads ]   [ Order/Register ]   [ Affiliate Program ]
[ Support ]   [ Feedback ]   [ FAQ ]   [ Online Help ]   [ Support Forums ]

© 2002-2006 AgenSoft. All rights reserved.   |   Terms of Use   |   Privacy Statement