Overhead of cputrack on UltraSPARC T1

Hi,
I'm trying to use performance counters on the SUN T1 to measure cache misses in a multithreaded application.
I launch the program with:
cputrack -o /tmp/PERF_OUT -c DC_miss,Instr_cnt <application args>When run via cputrack the execution time is almost 4x as long as when run without. Is this typical? Are the performance counters returned by cputrack reliable?
Am I correct in reading the totals next to "exit" in the cputrack output as the totals for the entire execution?
Is there a better way to measure these events than with cputrack?
Thanks so much in advance.
-John

You might want to try Solaris part of the forums to get better answer on cputrack.
Performance counters returned by cputrack are reliable to the extent of hardware counters themselves. On intel it is not that reliable, as counters are system wide and it has to resort to a sampling. I have an impression that on sparc it is better, but honestly I have no real experience.
I doubt there is a better way to measure, but there is a bunch of tools that provide different statistics and different views.
Solaris tools - cputrack, cpustat, mpstat, vmstat
Sun Studio - collect/analyzer
Analyzer(/collector) uses the same hardware counters system interface as cputrack, but it is much more powerful in terms of data representation - it groks your sourcecode and attributes counters to it.
I do not remember that much of a slowdown when running cputrack (as you report - 4x). Though these are quite typical numbers for collector.
Might very well depend on exact counters you gather.
regards,
__Fedor.

Similar Messages

Link aggregation - performance overhead?

Does anybody know if Solaris link aggregation incurs any performance degredation compared with non-redundant network connections?
We've recently upgraded a client system and have enabled link aggregation to bind two interfaces (bge) to a logical aggregated interface.
Apart from the server hardware upgrade, which brought a change from ce to bge interfaces, this is the only other significant network change.
On this heavily used system, database network performance has degraded significantly, and being the only significant network change I'm wondering whether link aggregation could be a cause.
Reading through Sunsolve articles and man pages, there doesn't appear to be anything categorically stating aggregation imposes a cost overhead.
However, I'd be interested to hear from others if they've experienced this.
For what it's worth, this is Solaris 10 10/09 on an Ultrasparc VII.

So many considerations when you talk about performance: what system hardware in use? how many interfaces? is this a CoolThreads "T" server (where single threaded programs don't run as well as multi-threaded); what do you mean by "database network performance has degraded"; does the DB need to be tuned or adjusted for this? If using add-on cards, are they in the most optimal slot (some platforms) and don't go over the recommended # of interface boards supported by the system? different storage array or config for you DB?
when cases like these come in, there typically is a discussion that needs to take place around those questions as well as what/when/where and expectations like those described here: http://blogs.sun.com/hippy/entry/what_s_the_answer_to
also understand that link aggregation does not necessarily mean that you will see a balancing of the load evenly b/t 2 or more aggregated interfaces. The default policy is "L4" which decides the outbound interface by TCP and UDP info if found in the packet, not by simple source/destination or mac address hashing, although you can set those with -P if you want. You can also set multiple policies. (-P L2,L3).
So it /could/ be related to the aggregation change you made, but you also upgraded this DB machine from some other system, so there ARE other differences. Were you using link aggregation on the previous system or Sun Trunking? What type of CPU was that other system?
that's all I can think of, but hopefully you get the idea.

Error while entering the Overhead percentage 0.3693 in Costing Sheet

HI,
While defining costing sheet in the system,i need to give the overhead percentage as 0.3692 as per the client requirement.but the system is giving the below error.
Input should be in the form __,___,__~.___V
Message no. 00088
Diagnosis
Your entry does not match the specified input format.
System response
The entry in this field was rejected.
Procedure
The entry must comply with the edit format. The following edit format characters have a special meaning:
"_" (underscore)
There should be an input character at this point; this should be a number for numeric fields.
"." (decimal point) (applies to numeric fields)
The decimal point occurs here (setting in the user master record).
"," (thousands separator) (applies to numeric fields)
This separator occurs (optionally) for more than three figures. Depending on the setting in the user master record, it can be a period or a comma.
"V" (applies to numeric fields)
The operational sign appears here. If used, it must by at the right margin of the field. The sign is either "-" or " "(space).
"~" (tilde) (applies to numeric fields)
As of and including this character, leading zeros must also be entered. Otherwise, this character has the same meaning as an underscore. Leading zeros need not be entered on the left of the tilde. They are not output at this position.
All other characters have their normal meanings and must be entered in the same position as in the edit format.
Thanks
Sunitha

hi Sunitha,
this error is related to Decimal Format pls check your setting in country
For 4.7C
SPRO > IMG > General Setting > Define countries in mySAPsystem
For ECC 6.0
SPRO > IMG > SAP Netweaver > General Setting > Define countries in mySAPsystem
select your countries and double click in that there is last field for Decimal Pt. format check that format and try to enter value according to that format because in error it showing 4 decimal and there you may have maintained in two decimal. also check decimal currecy setting. One more reason could be there instead of full stop you may have maintained comma in setting.
thx.
Ganu

What is the overhead of Supplemental Logging?

We would like to copy data from a 600 GB Oracle database (9i) to a separate database for reporting. We want to use Data Guard in Logical Standby mode. The source database is heavily used and we can't afford a significant increase in system load (e.g. I/O activity) on that system.
To set up a Logical Standby, we need to put the JD Edwards database into Supplemental Logging mode. I am concerned that this will noticibly increase the load on the source server.
Has anyone analyzed the additional overhead of Supplemental Logging?
I have done some testing using Oracle 10.2 (Oracle XE on my computer) which indicates that when I turn on Supplemental Logging, the size of the archive logs grows by 40%. I have not yet tested this on our 9i database.
Thank you in advance for your help!
Best Regards,
Mike
=================================
The code below demonstrates the symptoms mentioned above:
RESULTS - size of archive logs generated:
- With Supplemental Logging: 120 MB
- Without: 80 MB
=================================
CREATE TABLE "EMP"
(     "EMPLOYEE_ID" NUMBER(6,0),
     "FIRST_NAME" VARCHAR2(20),
     "LAST_NAME" VARCHAR2(25) NOT NULL ENABLE,
     "EMAIL" VARCHAR2(25) NOT NULL ENABLE,
     "PHONE_NUMBER" VARCHAR2(20),
     "HIRE_DATE" DATE NOT NULL ENABLE,
     "JOB_ID" VARCHAR2(10) NOT NULL ENABLE,
     "SALARY" NUMBER(8,2),
     "COMMISSION_PCT" NUMBER(2,2),
     "MANAGER_ID" NUMBER(6,0),
     "DEPARTMENT_ID" NUMBER(4,0)
alter table emp add CONSTRAINT "EMP_EMP_ID_PK" PRIMARY KEY ("EMPLOYEE_ID")
CREATE TABLE "STAT"
(     "F1" NUMBER,
     "ID" VARCHAR2(10)
The "employee" table is from Oracle XE samples
The procedure below generates transactions to test archive log size.
To run, put the database in archive log mode. Then pop an archive log by executing
ALTER SYSTEM ARCHIVE LOG CURRENT;
To flip between running with Supplemental Logging, use one of:
ALTER DATABASE drop SUPPLEMENTAL LOG DATA (PRIMARY KEY, UNIQUE INDEX) COLUMNS;
ALTER DATABASE ADD SUPPLEMENTAL LOG DATA (PRIMARY KEY, UNIQUE INDEX) COLUMNS;
declare
i number;
begin
i := 0;
while i < 1000
loop
delete from emp;
insert into emp select * from employees;
update emp set COMMISSION_PCT = COMMISSION_PCT * .5;
update stat set f1 = i where id = 'UPD';
commit;
if i mod 1000 = 0 then
dbms_output.put_line(i);
end if;
i := i + 1;
end loop;
end;
/***********************************************/

Unless the bottleneck of your system is related in any way to the redo log files, I don't see any risk in generating supplemental logging. A good way to find out is to look at a Statspack report, and see which events are the top 5 time-wise.
Daniel

What is difference between SUNW,UltraSPARC-IIIi and US-III+

Hi All,
I am trying to finalise a VxVM version for a Solaris 8 version. On the Solaris machine I see,
System Configuration: Sun Microsystems sun4u Sun Fire 280R (UltraSPARC-III+)
System clock frequency: 150 MHz
Memory size: 4096 Megabytes
========================= CPUs ===============================================
Run E$ CPU CPU
Brd CPU MHz MB Impl. Mask
A 0 1200 8.0 US-III+ 11.0
On the compatibility matrix I see only Ultra Sparc IIIi and no III+ . Are they both same or hugely diffrent?
Please help.
Regards,
Anand

Google
["SPARC IIIi versus SPARC III"|http://www.google.com/search?hl=en&q=SPARC+iiii+versus+sparc+iii&btnG=Google+Search&aq=f&oq=]
Find Ben Rockford's blog comparing Sun's cpu architectures.
[http://www.cuddletech.com/blog/pivot/entry.php?id=332|http://www.cuddletech.com/blog/pivot/entry.php?id=332]

Overhead Percentage is not getting updated in KONX

Hi All,
The Overhead percentage which we maintain in KZZ2 is not getting updated in KONX table. Can anybody suggest whether the OH percentage is available in KONX table only or anyother table also? Or Is there any program which needs to be run to update the KONX table?
Regards,
Rajeswari Shankar.

Hi,
U can also check konx1 table.
Check whether overhead calculations are run. Check whether CO42(in case it Product Cost Collector), KGI2(in case of costing by order), VA44(incase of costing by salesorder) run as applicable.
Regards
Sudhakar Reddy

Overhead is not getting updated in the production order

Hi all,
I am running actual overhead calculation with KGI2.
But the overheads are not getting updated in the order.
Order contains relevant overhead key also.
What could be the possible causes?
SmanS

Arun/ Suleman,
Thank you for your reply.
I have maintained the % in the costing sheet against the 'Overhead rate' but I didn't get the secondary cost element.
Request you to elaborate it further & explain me how to maintain the same.
Also I will be thankful to you, if you could explain me the exact process of loading the overhead on the order i.e. how it loads the overhead form the cost center to order? What happens to the total cost which was lying on the cost center etc, in details.
Mean while I had maintained the required values in costing sheet in KZS2.
Maintained the % against material overhead rate as well as manufacturing overhead rate.
Maintained the credit values for the overhead rate.
Maintained the calculation base values also.
Now while overhead calculation with KGI2, I am getting two errors as follows,
1. S:K5:162 E02 20090831
2. Error CREDIT_SEGMENTS_ADD COIOB NOT FOUND for object
I can not figure out the reason.
SmanS
Edited by: SAP PP Consultant on Aug 10, 2009 6:47 AM

PL/SQL to Java Interface - Overwhelming Overhead?

My company is running Oracle 10g R2. I am currently exploring the use of java classes from PL/SQL. This is because we have multiple code bases, including Java, PowerBuilder (which can interface with Java), and PL/SQL (which can interface with Java).
Doing some performance testing, I found some rather alarming results. Running a simple java function 1 million times, I tested performance of PL./SQL to java and PowerBuilder to Java.
In PL/SQL, the 1 mil calls executed in ~120 s.
In PowerBuilder, the 1 mil calls executed in ~54 s.
Is it possible that our database is incorrectly configured for best java performance? Not only does it appear that PowerBuilder, a much less pervasive language than PL/SQL, is over twice as fast, but the PB test was run on my laptop, which is no where near as powerful as the database server that the PL/SQL procedure ran against, leading me to believe that the performance gap is actually much wider.
Anyone have any tips to improve the performance? Or is this simply the nature of the beast? I can't seem to find any posts on the net about people experiencing huge PL/SQL to Java overhead, so I'm hoping it is just something I am doing wrong.
Thanks for your help!
-Brett Birschbach

Below is my test code. Any suggestions are appreciated.
Thanks!
-Brett
==================================================
Java Class:
public class TestConvert {
    // For running straight from Java
    public static void main(String[] args){
        System.out.println(callAndTime());
    // For running 1 mil calls with only one call from PowerBuilder or PL/SQL
    public static long callAndTime(){
        TestConvert.convertEsnDecToHex("13005454488");
        long start = System.currentTimeMillis();
        for (int i = 0; i < 1000000; i++){
            TestConvert.convertEsnDecToHex("13005454488");
        long end = System.currentTimeMillis();
        return end-start;
    // For performing the actual work
    public static String convertEsnDecToHex(String esn) {
        String result = null;
        result = Integer.toHexString(new Integer(esn.substring(0, 3)).intValue());
        result += Integer.toHexString(new Integer(esn.substring(3)).intValue());
        return result.toUpperCase();
    // Non-static wrapper necessary for PowerBuilder
    public String callAndTimeNS(){
        return String.valueOf(callAndTime());
    // Non-static wrapper necessary for PowerBuilder
    public String convertEsnDecToHexNS(String esn){
        return convertEsnDecToHex(esn);
}PL/SQL Script for 1 Million Java Calls:
DECLARE
var       VARCHAR2( 20 );
t_begin   NUMBER;
t_end     NUMBER;
BEGIN
var := testconvert.convertesndectohex( '13005454488' );
t_begin := DBMS_UTILITY.get_time;
FOR i IN 1 .. 1000000
LOOP
    var := testconvert.convertesndectohex( '13005454488' );
END LOOP;
t_end := DBMS_UTILITY.get_time;
DBMS_OUTPUT.put_line( var );
DBMS_OUTPUT.put_line( ( t_end - t_begin ) / 100 || ' seconds' );
END;
/PowerBuilder for 1 Million Java Calls:
EJBConnection lEJBConn
TestConvert lnv_test
String ls_ret
Long i
Time t_begin, t_end
lEJBConn = CREATE EJBConnection
lEJBConn.CreateJavaInstance( lnv_test, "TestConvert")
ls_ret = lnv_test.convertesndectohexns("13005454488")
t_begin = now()
FOR i = 1 TO 1000000
     ls_ret = lnv_test.convertesndectohexns("13005454488")
NEXT
t_end = now()
MessageBox("Hex Version",ls_ret + " - " + String(SecondsAfter(t_begin, t_end)) + " seconds")
DESTROY lEJBConn

Overheads not calculated

Hello Experts,
After launching the KKS5, error messages for no calculation of the overheads on some materials.
When launching KKF6N for June the Indirect Costs are there but it's not the case for the month of July.
The costing variant planned for the materials is PREM.
I would like to have more explanations about this issue.
Thanks in advance,
Edited by: Ferdaws ABID on Sep 19, 2011 3:57 PM

Hello Tarek,
I have checked the configuration of the overhead in the transaction KZZ2 as you mentionned but everything seems to be OK.
The validity period is OK and the overhead rate is maintained correctly.
Do you have other suggestions ?
Regards,
Ferdaws

Overheads not calculated in CK11N cost estimation

Hi all,
I have two issues
1. When i am executing CK11N,i am not able to get the overhead calculation in itemization,but its showing some ridiculess firgure in cost component structure.I am using dependency D020.
2 . when i am using dependency D050 the system not calculating overheads.
We have single plant with two process Repetative and Process order.I want to calculate the overhead on the Order type basis but the system is not calculating the overhead.
So i changed the dependency to D050,now its showing the overhead figure in cost component structure but whic is not matching with the overhead rate i have given..
as my go live is 20 of this month .....its bit emergency
Revert
Regards
Pran

Hi all,
Thnx for your reply.
But i have maintained all the entries in the respective area,even though i am geeting the issue.
It was working fine for my previous repetative manufacturing ...i am using the single cost sheet for both repetative and process order.
And one more thing is that i have a Semi finished goods which is the output of REIM order and is included in process order of the finished good.
The system is calculating the overheads for the semi finished good too once again.
Eg,
Raw Material A and B produces Semi finish material C overheads and the utilities are calculated correctly in Repetative manufacturing.
The semi finished material C is used in producing the finished material D ....in which the system is not caculating the utilities.The big problem is that the figure(value) of itemzation and the Cost component structure is showing two different values.
I am not sure and able to know how and why the system calculates and from where it gets these values ...... totally confussed and under pressure (may not be able to to explain properly).
Revert
Pran

Costing Sheet Overhead Rate Change Mid Month

Does SAP Support a mid month overhead rate change in a costing sheet? I have an overhead rate change effective 5/12/2007, but my PM Order is not recognizing the change unless the Basic Start Date of the order is 6/1/2007.

hi
it will not support mid month overhead. it should be either period basis or yearly basis.
regards,
siva

HT1338 I'm trying to attach my MacBook Pro to an overhead projector and can't get it to work. If I go to System Prefs/Displays, I see the box on the screen. But when I collapse the Prefs the desktop is not seen. Can you help?

I'm trying to get an overhead projector to identify my macbookpro. If I go to Prefs and select screens that dialogue box comes up on screen. So I know there's nothing wrong with the VGA cable. But when the screen box is collapsed nothing happens. My computer is not recognized by the projector. Can anyone help?
On a PC I used to hit "control F5. But I don't believe that works on a MacBook. Any suggestions?

About the only thing you can try is using the Online Internet Recovery system to reinstall Lion on your Mac over the top of the current install. that will reset all system files and should not touch your personal files or the apps you have installed.
Or iof you have a recent Time Machine backup use that to restore your Mac to that point in time.

Cost center planning for overhead calculation

Gurus,
I have a doubt regarding the cost center planning for overhead calculation.
My scenario is:
There is a service cost center from which a certain percentage I want to allocate to the products as overhead. Now while doing so I have created a cost center and a cost element of type 43 (internal activity allocation). This cost element is the base for overhead calculation. Also I have created one cost element of type 41 (overhead rates). I have defined the credit (of cost sheet) with the cost element of type 41 and the cost center as mentioned.
Now I am facing a problem while planning the cost center with cost element type 43 through KP06.The system is throwing the error message: 'Cost element xxxx can only be used for activity input'. My understanding is we must carry out the planning for the cost center to allocate a certain percentage to the product cost estimate. If I am not allowed to use the cost element of type 43 to plan, what should I use for the planning?
Request your suggestion.
Thanks in advance!
Snigdho.

Hi
Since you mentioned catg 43 cost element, I thought you are using Activuty Type
To apply overhead on the product during CK11N, you need not plan any value on the cost center... The values planned in the cost center can not be used as a BASE in your costing sheet....
It goes like this
1. Your cost estimate has BOM + Routing
2. So, the cost elements that you would see during CK11N, can be broadly classified into 3 types
a. Primary cost ele - Material Consumption accounts for BOM components
b. Secondary cost ele (Catg 43) - If you use any actvity type in the routing
c. Secondary cost ele (Catg 41) - These cost elements are parasites... They need some basis to calculate the overhead value... That basis can be either the (a) or (b) above.....
I think, what you are planning to do is Plan Primary cost ele in a cost center and trying to use the same as the Basis.. That wont work!
What you need to do is
a. PLan the values in the cost center, say, 100,000 USD
b. Absorb it on the basis of 2.a or 2.b
If 2.b is not applicable in your case, since you are not using activity types, then 2.a is the Only option left for you
Hope it is clear....
Regards
Ajay M

Overhead costs on raw materials

Our requirement is to have some percentage overhead on raw materials. So I created a costing sheet and applied a percentage to them and did a std cost estimate. Now when GR is received it valuates against the std cost (which includes OH) and posts the difference to PPV. Client wants it to valuate against the orginal price of material (without OH)
Ex; ROH Mat A $1.00 in planned price 1
Costing sheet : 10% OH
Standard cost of this ROH A is $1.10
Purchase order will be at $1 so GR will post Dr. Inv $1.10 Cr. Gr/Ir $1.00 PPV $ 0.10
But clients want that PO price $1 to be valuated against $1 (true ROH cost) and post $0.10 to absorption account
Dr. inv $1.10, Cr Gr/IR $1 Cr abs $0.10
if P.O is $1.10 then Dr. Inv $1.10 Cr. Gr/IR $1.10 Dr.PPV $0.10 Cr ABS $0.10
I am not sure if it is even possible if not waht options do I have to have and overhead for Raw materials with out updating standard cost estimate with $1.10.
Thanks

Hi
Yes, it is possible
What you want is to Debit the inventory with 1.1 i.e. inclusive of Overhead.. But, Credit the GR/IR with 1 $ and post the 0.10 to absorption account instead of PPV... So, that PPV reflects true PPV
You need to create an accrual key for the same..
1. Create an accrual condition type and accrual account key in MM pricing procedure. In this way, there is a credit entry to the GL account represented by accrual account key. Hence the PPV represents the true value.
2. For the accrual condition type, you can create a condition record such that its value is same as your overhead. You can use the user exit to automatically populate its value from KEKO and KEPH table.
Refer this post where it was discussed a few days back
Product Costing-Overhead inclusion in Inventory standard cost without PPV
Regards
Ajay M

Overhead Calculation not display

Hi,
Please tell me while I do Overhead Calculation through KGI2, inspite of successful posting I cannot find overhead calculation figure which has posted to GL accounts. There is no errors, update successful and I click the icon to get the result but it does not shows line items. As per our scenario we use sales order costing. In sales order level costing is calculated and overhead shows properly as Material overhead, Admin overhead, production overhead ans Sales and Distribution overhead as configured in Cost componenet layout and cost sheet. In costing sheet Material cost is maintained under base as Total Cost. So overhead should calculated on Total cost of Material cost but it calculates on Variable cost. In KP26 I have tested both either fixed and variable cost price. Still it calculate on Variable cost inspite of cost sheet base is maintained on Total Cost. We use % Overhead calculation.
After sales order we prepared production order and final confirmation is made on production order's quantity issued. After that while I run 4 cycle which are Overhad calculation, WIP calculation, variance calculaton and settlment of production order then in first step I found Overhead calculation output does not appear. Even I cannot find any error also. What is the reason ?
Best Regards,
Samrat Roy

Dear,
Check your origin group & material unit of mesaurement is same.
some time in costing sheet origin group is maintain in different unit & for materail it's maitain in other unit of mesaurement.
You can see unit of measure for material in Additional data - unit of measure.
Check BOM component material unit also.
Check same unit of measure is maintain in KZS2
I hope above will useful.
GOPAN

Overhead of cputrack on UltraSPARC T1

Similar Messages

Maybe you are looking for