MAXBUFFER

Synopsis

HELP:   Maximum buffer size for FLAM matrix [65536]
TYPE:   NUMBER
SYNTAX: MAXBUFFER/MAXB=num

Description

During compression with modes cx7, cx8, and vr8 FLAM reserves the matrix buffer for the temporary storage of the records from the original file. Please note that a matrix buffer of the same size has to be available during decompression, too. This is also the case when the decompression is done on other systems.

Parameter MAXBUFFER specifies the maximum size of used matrix buffer. The information about MAXBUFFER is stored in the FLAMFILE, so it is well known during decompression.

During compression FLAM reserves the matrix buffer for the temporary storage of the records from the original file. Please note that a matrix buffer of the same size has to be available during decompression, too. This is also the case when the decompression is done on other systems.

Syntax restrictions for FLAMv4 or older:

• z/OS
  • MAXBUFFER=n
  • MAXB=n
    • Note: Used for CX7, CX8 and VR8 only.
      Regarding ADC, NDC and AES this parameter is not used. In this case selection is always a buffer of 64KB.
• UNIX
  • -maxbuffer=n
    • Note: Used for modes CX7, CX8 and VR8 only.
      With mode adc, this parameter has no effect.
• WINDOWS
  • maxbuffer=n
    • Note: Used for modes CX7, CX8 and VR8 only.
      Regarding ADC, NDC and AES this parameter is not used. In this case selection is always a buffer of 64KB.

Values

• z/OS
  • There are three groups of values with differing meaning
    • Values between 0 and 7 map to these fixed sizes:
      • 0 = 32 KiB
      • 1 = 16 KiB
      • 2 = 32 KiB
      • 3 = 64 KiB
      • 4 = 128 KiB
      • 5 = 256 KiB
      • 6 = 512 KiB
      • 7 = 1024 KiB
    • Values between 8 and 2047 are in KiB and are rounded upwards to a value of the following list:
      • 2, 4, 6, 8, 10, 12, 14, 16, 32, 48, 64, 80, 96
      • 112, 128, 144, 176, 224, 256, 288, 320, 352
      • 384, 416, 512, 640, 768, 896, 1024, 1536, 2048
    • Values from 2048 are in bytes.
      This value is rounded upwards by KiB according to the following list or is rounded downwards to 2560 KiB:
      • 8, 10, 12, 14, 16, 32, 48, 64, 80, 96
      • 112, 128, 144, 176, 224, 256, 288, 320, 352
      • 384, 416, 512, 640, 768, 896, 1024, 1536, 2048
• UNIX / WINDOWS
  • Integer number between 1 and 2,621,440.
    Values n, where 0<n<2,560 are interpreted as a number of KiB (1 KiB = 1,024 bytes), while higher values are interpreted as a number of bytes. FLAM chooses the nearest equal or value value (in KiB) from the list below:
    • 2, 4, 6, 8, 10, 12, 14, 16, 32, 48, 64, 80, 96
    • 112, 128, 144, 176, 224, 256, 288, 320, 352, 384
    • 416, 512, 640, 768, 896, 1024, 1536, 2048, 2560

Default

• z/OS
  • 64 kByte
• UNIX
  • none
• WINDOWS
  • for CX7, CX8 and VR8: none
  • for ADC, NDC and AES: Parameter is not used. Selection is always a buffer of 64KB.

Valid for

• compression, mode=cx7,cx8,vr8