Ibm Assembler Placement Papers - Ibm Assembler Interview Questions and Answers updated on 24.May.2024

START is an assembler directive which has an optional operand, (0 in your example). This operand acts as the starting address of the program e.g. PROGNAME START X'3E0' tells the linkage editor that this program is to be loaded into main storage at address '3E0' in hex. 

Lenght feild is 8 bits so the data is stored in binary format so if we have all 8 bits set to 1 then the length will be 128+64+32+16+8+4+2+1 = 255.or 2^7+ 2^6+2^5...2^0 = 255.

so the lenght is 256 .

In MVS assembler data type X denotes hexadecimal data type  which unsigned pack. suppose you define VAR1 as "VAR1 DC X'01'".

It will occupy 1 byte in the memory and stored as: 

  • 0 in the zoned nibble and 1 in the numeric nibble.
  • P denotes the packed data type, similar to COMP-3 in COBOL. 
  • if you declare any variable with this data type then it  must have a sign byte at last nibble. 

See following example:

VAR2 DC P'1'

it will occupy one byte in the memory and stored as '1C'.

VSAM file can be accessed through an Assembler application program by using assembler macros RPL, ACB

Registers are part of the CPU logic and should not be confused with memory real or virtual. The question is about as sensical as asking where the uterus resides in a man. Also, general purpose registers 0--16 are an entirely different breed that the floating point registers. As for the numbering, a decision made by the manufacturer's design people.

Neither DROP nor USING affect the register contents. Once a register has been loaded with the address of a piece of storage, the using statement can be used to 'map' that storage against a set of labels defining the layout of that storage e.g. a DSECT. Then whenever one of those labels is referenced in the code, in moves etc, the assembler resolves the relative address as a particular displacement from the absolute address in the register. The DROP instruction removes the relationship between the labels and the register and renders subsequent references to those labels as unresolvable, giving rise to errors at assembly (compile)time. 

Typically the DROP instruction will be used to allow use of the register for another purpose, e.g. address a different bit of storage via a using staement on second DSECT without the risk of corrupting that data via moves referencing the original DSECT.

By convention, general purpose register 1 will have the address of the parameter list. The list will be a list of pointers (addresses) to individual parameters. The CALL macro does this, but you can bypass the CALL macro and be creative on how you set up the parameter list. Don't forget the parameter must be on a fullword boundary.

It will give an error at the time of assembly, if there are some labels defined in the program(provided they are being referenced in your program). This is because the assembler resolves the displacement of that variable from the location where your base register is pointing to.

If you write an MVC statement with an equated value as the sending operand, then the assembler will try to resolve that operand value as a base and displacement, it will not necessarily throw an error at assembly, but the results at execution will be unpredictable and may well give rise to a protection exception. 

The point of the MVI statement is that the single byte sending operand value is assembled as part of the instruction itself and does not have to be 'fetched' at execution time, therefore if you are only moving a single byte of fixed value, then an MVI will be marginally more efficient than an MVC

S0C7 occurs when the data exception occurs.They are many ways to produce S0C7

  1. Move non numeric data to a numeric feild 
  2. Compare junk data with a numeric feild
  3. Use a non numeric data in COMPUTE statement
  4. Refer to the occurence in a table beyond the occurs time with SSRANGE not checked in complier options

  1. If there is any updates in the Macros regarding the structure change or addition/removal of byte, then you need to Promote the macro to Live first.
  2. Reassemble your codes against the Live Macros.
  3. Raise ELAS for the Codes to go to production.
  4. Promote the codes to production.

Use Accept in procedure division.

Example : 
01 empno. Pic x(05).
01 empname pic x(15).
01 empsal pic 9(10).


In JCL :

//sysin dd *

A base register is any general purpose register chosen by programmer. It is assigned at the beginning of the program as part of housekeeping with the USING assembler keyword, and it's purpose is to maintain addressibility within a page (4k) of code or data.

There are two possible scenarios: 

  • This is a classroom quesiton where you are expected to  do some research on your own, and not ask someone in the  cloud for the exact wer.
  • The question is too vague, and would require a full  essay to wer. Please be more specific.

XR Rx,Rx:

  • This is the best way and most efficient way of initializing
  • the register values or Label values to '0'.
  • Its because the time required to execute a Logical
  • Instruction is always less than the Arithmetic Instructions.
    • e.g SR Rx,Rx consumes more execution time than XR Rx,Rx.
  • So XR Rx,Rx is a better way to do obtain our target.

House keeping is used to store the contents of the base register from one register to another for using that register. These are house keeping instructions where contents of the registers are stored

  1. content of the register 15 is stored in the register 12 
  2. address in reg 12 is mapped to the module type
  3. save area is stored in the register 13

There are 16 registers, and ALL can be used as a base EXCEPT for register 0, so the wer is AT LEAST @AT LEAST is specified here because in any section of code, you can "re-use" a previous base register once you are no longer within the original address range ("scope") of that particular base register. By re-using the registers, you can have base registers that will cover ALL of the memory in the machine - but not all at once - you have to "bite off" 15 base reg-at-a-time chunks of memory (all addressable memory does not have to be contiguous - it can be scattered around memory in 4K pieces).

We can certainly use the MVC instruction to move the pack data to another pack field.

it all depends upon the length of your program , if it is more than 4095 then we need 2 register else only one can do needful. basically we can even use savearea register as a base register, so as to accomodate the length of the program

R1 contains a fullword that contains a address pointing to the parm data. In pgm before accessing the parm data use L  Rn,0(,R1) where n=3,..11 , Rn contains the address that points to parm data. Create DSECT that contains a halfword and the length of the data.

For example we want to test if given variable is numeric we can do it by following code.

We can define a table with all FF execpt from 0 to 9 00

Table1 DC 256c'ff'
org table1
dc 10c'00'
Now we can give TRT var1,table1 
BZ numeric

As this is mainframe assembler section, this is a trick question - there is no JMPNZ opcode for mainframe (recently added JNZ with relative addressing in the z/800 and later,but no JMPNZ) and the mainframe has no RET instruction (there is a PR to return from a cross address space or PC "call" statement) and the mainframe has no accumulator, ADB opcode, or B register.

This is a vague two-part question.

  1. You can round by adding .5
  2. Adds 5 to WKUR, but it had better be a valid packed field, and not just a label.

ICM me INsert character under mask. its a movement of character under mask bits.


ICM R8,B'0110',SRC

Before execution: 

R8 ==> 12 34 56 78
SRC ==> AB CD EF 10

After Excecution:

R8 ==> 12 CD EF 78

Assuming displacement is referring to x(d,b) addressing and not something like displacement from the start of a CSECT.

You will need to open de file SYSIN via a DCB and read the data via GET commands. The data in the PARM field is passed via register @Any dataset you want to use in Assembler you will need to open. There are no automatic allocation.

since the machine-code for MVC moves up to 256 bytes, you would need to do a series of MVCs to initialize 20000 bytes. this requires that you maintain a register or two to keep track of how far you've progressed through initialization. 
possibly, you could get MVCL to do it; i've never tried...

MVCL uses 2 sets of even-odd pairs of registers to do the move.

you specify source address, destination address, length of source, length of destination, and fill character in the registers. the fill-character goes into the high order byte of (I THINK...) the destination length register (in this case r4)

l r4,=f'20000'
l r6,=f'20000'
la r7,source_field
la r5,dest_field
mvcl r4,r6