BOB: Business Objects Board
Not endorsed by or affiliated with SAP

Register | Login 

Streamline and extend SAP BusinessObjects software 
Streamline and extend SAP BusinessObjects software (Opens a new window)  

General Notice: No events within the next 45 days.

Benchmark YOUR hardware and compare the results V1.1
1 members found this topic helpful
Goto page Previous  1, 2, 3, 4, 5, 6, 7, 8, 9, 10  Next
 
Search this topic... | Search DI: Performance and Tuning... | Search Box
Register or Login to Post    Forum Index -> Data Integrator -> DI: Performance and Tuning  Previous TopicPrint TopicNext Topic
Author Message
kentto
Forum Member
Forum Member



Joined: 17 Sep 2009

Posts: 2



PostPosted: Mon Nov 30, 2009 6:08 pm 
Post subject: Re: Benchmark YOUR hardware and compare the results

Hi All,

Not sure if anyone still checks this, but here are some results from our environment.

Database: SQL Server 2008 Enterprise Edition 64 bit
Windows: Windows Server 2008 64 bit
Data Services: XI R3.2

DS and databases on the same server.

Server specs: 2x 3 Ghz Xeon, 32 GB RAM, RAID 10 disk configuration

4516 1972 DATAFLOW 1/12/2009 9:45:40 AM Process to execute data flow <DF_Benchmark_read_MS> is completed.
5988 5368 PRINTFN 1/12/2009 9:45:40 AM Time______________________________ 72 secounds

6136 4564 DATAFLOW 1/12/2009 9:47:41 AM Process to execute data flow <DF_Benchmark_API_bulkloader_MS> is completed.
5988 5368 PRINTFN 1/12/2009 9:47:41 AM Time______________________________ 121 secounds

1988 3232 DATAFLOW 1/12/2009 9:52:58 AM Process to execute data flow <DF_Benchmark_regular_load_MS> is completed.
5988 5368 PRINTFN 1/12/2009 9:52:58 AM Time______________________________ 316 secounds

2168 6012 DATAFLOW 1/12/2009 9:53:15 AM Process to execute data flow <DF_Benchmark_single_thread_MS> is completed.
5988 5368 PRINTFN 1/12/2009 9:53:15 AM Time______________________________ 17 secounds

4504 5592 DATAFLOW 1/12/2009 9:53:47 AM Process to execute data flow <DF_Benchmark_lookup_DOP1_MS> is completed.
5988 5368 PRINTFN 1/12/2009 9:53:47 AM Time______________________________ 32 secounds

5364 5584 DATAFLOW 1/12/2009 9:54:05 AM Process to execute data flow <DF_Benchmark_lookup_DOP10_MS> is completed.
5988 5368 PRINTFN 1/12/2009 9:54:05 AM Time______________________________ 17 secounds
Back to top
Werner Daehn
Forum Devotee
Forum Devotee



Joined: 17 Dec 2004

speaker.gif*6
Posts: 10590
Location: Germany


flag
PostPosted: Tue Dec 01, 2009 3:43 am 
Post subject: Re: Benchmark YOUR hardware and compare the results

Nice numbers. I'll add them to the overview asap!

Edit: Done.

@all: Noticed I have been lacy recently, there have been a few missing in the Wiki page. Corrected. Sorry for that.

@Doug & Data_Guy: And idea why you get so high numbers in the lookup_DOP10 case? Not in all cases the pageable cache was used.
btw, the reason pageable cache takes longer is because each cache has to be built 10 times and pageable cache has longer cache build times - working on that.

_________________
Werner Daehn
Product Manager Data Services
SAP AG
“The opinions expressed in this post are my own views and not those of SAP.”

PLEASE SUPPORT THIS: Run one ATL job and tell us the benchmark result. Details to be found here.
Back to top
avg_27
Forum Member
Forum Member



Joined: 26 Feb 2010

Posts: 7



PostPosted: Mon Mar 08, 2010 12:00 pm 
Post subject: Re: Benchmark YOUR hardware and compare the results

Hello All,

My results for this job are:

Source: Oracle 10g
Network source to DI: 1GB Ethernet
DI server Details:
OS Name: Microsoft(R) Windows(R) Server 2003, Enterprise Edition
OS Version:5.2.3790 Service Pack 2 Build 3790
System Manufacturer:GATEWAY
System Model:E-9722R
System Type:X86-based PC
Processor(s):8 Processor(s) Installed.

[01]: x86 Family 15 Model 65 Stepping 2 AuthenticAMD ~2799 Mhz

[02]: x86 Family 15 Model 65 Stepping 2 AuthenticAMD ~2799 Mhz

[03]: x86 Family 15 Model 65 Stepping 2 AuthenticAMD ~2799 Mhz

[04]: x86 Family 15 Model 65 Stepping 2 AuthenticAMD ~2799 Mhz

[05]: x86 Family 15 Model 65 Stepping 2 AuthenticAMD ~2800 Mhz

[06]: x86 Family 15 Model 65 Stepping 2 AuthenticAMD ~2800 Mhz

[07]: x86 Family 15 Model 65 Stepping 2 AuthenticAMD ~2800 Mhz

[08]: x86 Family 15 Model 65 Stepping 2 AuthenticAMD ~2800 Mhz

Total Physical Memory:16,383 MB
Page File: Max Size:18,182 MB
DI Version: XIR3.2
Network DI to Target: shared with above
Target: Oracle 10g

RESULTS:

<DF_Benchmark_read_MS> -------------------- 54 sec

<DF_Benchmark_API_bulkloader_MS> ---------- 262 sec

<DF_Benchmark_regular_load_MS> ------------ 187 sec

<DF_Benchmark_single_thread_MS> ----------- 52 sec

<DF_Benchmark_lookup_DOP1_MS> ------------- 79 sec

<DF_Benchmark_lookup_DOP10_MS> ------------ 362 sec

Please let me know your thoughts on it .
Back to top
Werner Daehn
Forum Devotee
Forum Devotee



Joined: 17 Dec 2004

speaker.gif*6
Posts: 10590
Location: Germany


flag
PostPosted: Mon Mar 08, 2010 12:13 pm 
Post subject: Re: Benchmark YOUR hardware and compare the results

please set all dataflows to in-memory caching, especially the DOP ones, and the job property to use statistics for optimization disable.

DOP10 should be way faster than DOP1.

_________________
Werner Daehn
Product Manager Data Services
SAP AG
“The opinions expressed in this post are my own views and not those of SAP.”

PLEASE SUPPORT THIS: Run one ATL job and tell us the benchmark result. Details to be found here.
Back to top
dnewton
Forum Addict
Forum Addict



Joined: 30 Jan 2004
ASUG Icon
Posts: 3456
Location: Massachusetts, USA


flag
PostPosted: Mon Mar 08, 2010 12:28 pm 
Post subject: Re: Benchmark YOUR hardware and compare the results

wdaehn wrote:
DOP10 should be way faster than DOP1.


Hooray, someone else finally is seeing the same thing. I wonder what our environments (infrastructure) have in common...
Back to top
avg_27
Forum Member
Forum Member



Joined: 26 Feb 2010

Posts: 7



PostPosted: Mon Mar 08, 2010 12:53 pm 
Post subject: Re: Benchmark YOUR hardware and compare the results

Hello Werner,

Thank You very much for the quick reply. I have re-run the job selecting IN-MEMORY and I've unchecked the collect statics for optimization option. My results now are:


<DF_Benchmark_read_MS> -------------------- 59 sec

<DF_Benchmark_API_bulkloader_MS> ---------- 340 sec

<DF_Benchmark_regular_load_MS> ------------ 415 sec

<DF_Benchmark_single_thread_MS> ----------- 213 sec

<DF_Benchmark_lookup_DOP1_MS> ------------- 38 sec

<DF_Benchmark_lookup_DOP10_MS> ------------ 13 sec
Back to top
Werner Daehn
Forum Devotee
Forum Devotee



Joined: 17 Dec 2004

speaker.gif*6
Posts: 10590
Location: Germany


flag
PostPosted: Mon Mar 08, 2010 3:33 pm 
Post subject: Re: Benchmark YOUR hardware and compare the results

icon_cool.gif

Just don't draw the conclusion pageable cache is slower. The DOP=10 case has to read ten times the full data into memory and the cache build time is slower for pageable cache, not the lookup as such.

_________________
Werner Daehn
Product Manager Data Services
SAP AG
“The opinions expressed in this post are my own views and not those of SAP.”

PLEASE SUPPORT THIS: Run one ATL job and tell us the benchmark result. Details to be found here.
Back to top
ErikR
Forum Associate
Forum Associate



Joined: 10 Jan 2007

Posts: 965
Location: Wellington, NZ


flag
PostPosted: Thu Apr 01, 2010 6:20 pm 
Post subject: Re: Benchmark YOUR hardware and compare the results

Werner, are you compiling the results of these stats into a report of sorts?
Back to top
Werner Daehn
Forum Devotee
Forum Devotee



Joined: 17 Dec 2004

speaker.gif*6
Posts: 10590
Location: Germany


flag
PostPosted: Sat Apr 03, 2010 5:52 am 
Post subject: Re: Benchmark YOUR hardware and compare the results

Define "Report". I tried once to come up with a graph but there have been too many variables to get anything meaningful.

I collect them in a table...

https://wiki.sdn.sap.com/wiki/display/BOBJ/Results

_________________
Werner Daehn
Product Manager Data Services
SAP AG
“The opinions expressed in this post are my own views and not those of SAP.”

PLEASE SUPPORT THIS: Run one ATL job and tell us the benchmark result. Details to be found here.
Back to top
ErikR
Forum Associate
Forum Associate



Joined: 10 Jan 2007

Posts: 965
Location: Wellington, NZ


flag
PostPosted: Tue May 04, 2010 8:43 pm 
Post subject: Re: Benchmark YOUR hardware and compare the results

Source: Oracle 10g 10.2.0.4.0 64-bit
Network source to DI: 1GBit
DI Server: VMWare Win2003 R2 SP2 32-bit Standard Edition
2 CPU @ 2.3Ghz (AMD 2380), 3.75 GB PAE
DI: 12.2.2
Network target to DI: 1GBit
Target: Oracle 10g 10.2.0.4.0 64-bit

Virtual Win2003 server provided by EDS (v4.60.001)
VMWare version and configuration unknown.

Source and Target on the same Oracle 10g database.
The Oracle 10g database runs in a virtual 64-bit Red Hat Linux environment of unknown hardware specifications.

<DF_Benchmark_read_MS> -------------------- 52 sec

<DF_Benchmark_API_bulkloader_MS> ---------- 587 sec

<DF_Benchmark_regular_load_MS> ------------ 1189 sec

<DF_Benchmark_single_thread_MS> ----------- 87 sec

<DF_Benchmark_lookup_DOP1_MS> ------------- 41 sec

<DF_Benchmark_lookup_DOP10_MS> ------------ 41 sec


@Werner: I meant, besides collection all this information - are you going to base some conclusions on this? Or point out which factors seem to have the greatest impact on DI performance?
Back to top
jeff55
Forum Member
Forum Member



Joined: 11 Sep 2006

Posts: 4



PostPosted: Wed May 26, 2010 11:28 am 
Post subject: Re: Benchmark YOUR hardware and compare the results

Hi Werner,
I'd like to do this benchmark test... where can I get the atl file? (it says will be added shortly?)

Thanks! This is a great initiative icon_smile.gif
Jeff
Back to top
Werner Daehn
Forum Devotee
Forum Devotee



Joined: 17 Dec 2004

speaker.gif*6
Posts: 10590
Location: Germany


flag
PostPosted: Wed May 26, 2010 11:34 am 
Post subject: Re: Benchmark YOUR hardware and compare the results

Got me. I am currently changing it to version V1.01, wanted to have the file ready last Monday already, so every second.
_________________
Werner Daehn
Product Manager Data Services
SAP AG
“The opinions expressed in this post are my own views and not those of SAP.”

PLEASE SUPPORT THIS: Run one ATL job and tell us the benchmark result. Details to be found here.
Back to top
jeff55
Forum Member
Forum Member



Joined: 11 Sep 2006

Posts: 4



PostPosted: Wed May 26, 2010 11:55 am 
Post subject: Re: Benchmark YOUR hardware and compare the results

Okay, thanks. Looking forward to the updates!

Jeff
Back to top
Werner Daehn
Forum Devotee
Forum Devotee



Joined: 17 Dec 2004

speaker.gif*6
Posts: 10590
Location: Germany


flag
PostPosted: Thu May 27, 2010 6:54 am 
Post subject: Re: Benchmark YOUR hardware and compare the results

Done.

@all: I have released version 1.1 of the benchmark. It does not differ from the initial version significantly, however some numbers cannot be compared.

    1. I have upgraded the ATL to 11.7 version. This should avoid people using different cache settings (pageable cache instead of in-memory cache method) and utilizes datastore configurations.

    2. It is partitioned more for those cases where larger disk arrays are used for the databases.

    3. The Reading dataflow was very likely flawed for most of us. First because one reader thread might have been the bottleneck. But for Oracle the data was not even transferred! I didn't know that Oracle SQL*Net does compress multiple identical rows. As a result the network was never busy in such a case. So the source table has different rows now.

    4. And I have added a monitor chapter in case you want to validate the resource utilization.



Please let me know if you have troubles using the new version so I can fix it asap.

_________________
Werner Daehn
Product Manager Data Services
SAP AG
“The opinions expressed in this post are my own views and not those of SAP.”

PLEASE SUPPORT THIS: Run one ATL job and tell us the benchmark result. Details to be found here.
Back to top
dnewton
Forum Addict
Forum Addict



Joined: 30 Jan 2004
ASUG Icon
Posts: 3456
Location: Massachusetts, USA


flag
PostPosted: Thu May 27, 2010 6:58 am 
Post subject: Re: Benchmark YOUR hardware and compare the results

Werner, would you like all of us to re-test using this newer version, so we have consistent numbers for comparison?
Back to top
Display posts from previous:   
Register or Login to Post    Forum Index -> Data Integrator -> DI: Performance and Tuning  Previous TopicPrint TopicNext Topic
Page 7 of 10 All times are GMT - 5 Hours
Goto page Previous  1, 2, 3, 4, 5, 6, 7, 8, 9, 10  Next
 
Jump to:  

Index | About | FAQ | RAG | Privacy | Search |  Register |  Login 

Get community updates via Twitter:

Not endorsed by or affiliated with SAP
Powered by phpBB © phpBB Group
Generated in 0.0736 seconds using 17 queries. (SQL 0.0195 Parse 0.0355 Other 0.0186)
CCBot/2.0 (http://commoncrawl.org/faq/)
Hosted by ForumTopics.com | Terms of Service
phpBB Customizations by the phpBBDoctor.com
Shameless plug for MomentsOfLight.com Moments of Light Logo