RE: Disk vs Tape scenario

Art,

With the storage available for me to test it is hard to eliminate cache
affects for writing because the whole volume can fit in cache.

Theoretically a volume using RAID-10 on 8x10K RPM disks (4D+4D) can handle
about 950 writes/second, but with just a single volume active it would be a
much higher rate because the seek would be minimal.

A write SSCH rate of 475/sec would push the RAID-10 parity group to 950
IOPS, so using typical QSAM BUFNO of 5 and half tracks BLKSIZE you would
need to sustain higher than 63MB/sec to a single volume to get Deferred
Write affects.

For Sequential writes to RAID-5 with 8x10K RPM Disks (7D+P) you would need
to push higher than 111MB/sec (where the disk vendor uses gathered writes
for RAID-5).

RAID-5 using 3D+P (4x10K RPM) would need to exceed 48MB/sec, and old
fashioned RAID-1 (1D+1D) must exceed 16MB/sec to realise deferred write
affects.

Anyway, I just tested with two IEBDG jobs writing 20 million 80 byte records
each to the same 3390 volume with one alias. That's about 3052MB at
73MB/sec. I was using DCB=BUFNO=16, so the QSAM chain was 8 which matches an
SSCH rate of 340 IOPS.

Ron

-----Original Message-----
From: IBM Mainframe Discussion List [mailto:IBM-MAIN@xxxxxxxxxxxx
Behalf Of Art Celestini
Sent: Monday, 13 February 2006 11:40 PM
To: IBM-MAIN@xxxxxxxxxxx
Subject: Re: Disk vs Tape scenario

Bruce, Ron:

I was referring to the "write side" of the operation. Reading back
the intermediate files, although potentially benefiting from speed
enhancements like read-ahead and PAV, is going to be throttled by
the network bandwidth.

I'll admit that it's been a while since I did any serious DASD throughput
tests with decent hardware. Best I've seen (informally) with the hardware
I have available to me at the moment is about 15MB/sec (on a z800). What
kind of numbers do you see when *writing*? (File must be large enough to
remove the effects of any write cache.)



----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to listserv@xxxxxxxxxxx with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
.

Relevant Pages

  • Re: Cache questions
    ... > sets and the MMX and XMM registers where possible. ... > I'm writing a routine which will need to read from a source bitmap ... especially when it causes cache misses. ... > some stack space, and I could write one of my local variables with, ...
    (comp.lang.asm.x86)
  • Re: need fastest way to write 2gig array to disk file
    ... Eric Taylor wrote: ... > writes seem to go directly to the cache and so a 2 gig output ... > On another similar system, with scsi disks, once the program ... (when writing to the cache, ...
    (comp.os.linux.development.apps)
  • Re: [opensuse] Hard disc questions Slight OT
    ... speed than outer radius. ... but they pack more data onto the outer tracks than ... Since you are writing to a file, you don't have anyway to say what ... Kernel cache can be even bigger. ...
    (SuSE)
  • Cache questions
    ... sets and the MMX and XMM registers where possible. ... I'm writing a routine which will need to read from a source bitmap ... But actually my knowledge of how the L1/L2 cache works is a little ... my local variables to memory, what's the best way to do that? ...
    (comp.lang.asm.x86)