IBM A22-6825-0 360 Instr Timing free download

File name A22-6825-0_360_Instr_Timing.pdf

File 8360-01 Form A22-6825-0 Systems Reference Library System/360 Instruction Timing Information This bulletin contains instruction timing in- formation for all models of the System/360. For full description of each instruction, refer to IBM 8ystem/360 Principles of Operation, Form A22-6821. SySTEM/360 INSTRUCTION TIMING INFORMATION These instruction time tables provide average in- field) is assumed to be greater than, or equal struction times, in microseconds, for all models of to, the length of the second operand (i. e. , the the IBM System/360. All times for instructionE:1 source field). that reference storage include the time required to perform single indexing by one general register as 2. For decimal-add and decimal-subtract instruc- referenced in the instruction (i. e., B 1- 0). For tions, recomplementation tini.e is not included those instructions in which indexing may be per- in the time s given. formed two times (i. e. , all RX format instructions), the time required for the second indexing must be 3. In the EDIT AND MARK (EDMK) instruction, added to the basic time given in the table, as follows: an address is stored once. That is, this in- struction is used with a single field, or a line MODEL ADDmONAL TIME (in microseconds) with only one numeric field rather than a com- plete print line. 30 6.0 40 1.25 4. In the TRANSLATE AND TEST (TRT) instruc- SO ~ 0-5 60 0.125 tion, it is assumed that a non -zero byte from 62 0.125 the Translate and Test table is found. 70 0.0 The instructions that may be indexed a second time are highlighted by an asterisk in the table. FLOATING POINT INSTRUCTIONS The follOwing assumptions were used in the development of these instruction times tables. The instruction times for the floating-point instruc- tions depend upon the number of hexadecimal digits that are preshifted and postshifted, as well as the VARIABLE FIELD LENGTH INSTRUCTIONS number of times recomplementation -of a result occurs. The times given in the instruction time 1. For decimal-add and decimal-subtract instruc- tables for floating point instructions are a weighted tions, the first operand (i. e., th