Re: Compressed Disks

    For reading compressed data, unless you
 are short on CPU cycles, I believe that it
 is faster to do less I/O.  Hercules is not
 an exception.  Flex would not only not be
 an exception, but any MIPS limits imposed
 by a zOS license might leave extra Intel
 CPU cycles for other work?

    A test that I just ran on a 9672-RB5 with
 RAMAC disks and a data set in both normal and
 one in SMS compressed PS-E format. ( same data ) 
 Data set:
 RECFM=FB
 LRECL=245
 Records=981159
 Tracks-compressed=2243 Excp-Count=4,485
 Tracks-uncompressed=4304 Excp-count=8,608
 Tests read the compressed, then uncompressed, then
 compressed, then uncompressed with IDCAMS REPRO.
 Output was DD DUMMY.  No BUFNO or other parameters
 specified.  i.e. Just DISP=SHR and DSN=specified

 Step #1
 ELAP.TIME 00:00:26.41
 STEP CPU 00:00:19.71
 SRB TIME 00:00:00.29

 Step #2
 ELAP.TIME 00:00:34.25
 STEP CPU 00:00:04.63
 SRB TIME 00:00:00.30

 Step #3
 ELAP.TIME 00:00:27.05
 STEP CPU 00:00:19.83
 SRB TIME 00:00:00.22

 Step #4
 ELAP.TIME 00:00:33.65
 STEP CPU 00:00:04.67
 SRB TIME 00:00:00.30

    I could also try copying 2400 HET files to null
 and 2400 AWS files to null, but since I know that
 I'm getting about 10 to one compression and have
 enough Mhz to decompress, I also know that the AWS
 copy would take much longer reading from my 5400
 RPM disk drive.  ( All of that was one sentence? )

--- In [email protected], "Peter D. Ward" <pdw@m...> wrote:
>I've edited your reply for logical consistency.  PDW
>edamjr wrote:
>>I (and others) have found **** under Hercules **** a
>>noticable performance improvement with compressed DASD
>> - I guess with current CPU performance, it's faster
>>to decompress on the fly than read more data from disk.
> [Non-text portions of this message have been removed]


------------------------ Yahoo! Groups Sponsor ---------------------~-->
Get 128 Bit SSL Encryption!
http://us.click.yahoo.com/JjlUgA/vN2EAA/kG8FAA/dpFolB/TM
---------------------------------------------------------------------~->

To unsubscribe from this group, send an email to:
hercules-advocacy-unsubscribe-hHKSG33TihhbjbujkaE4pw@public.gmane.org

 

Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/