[email protected] (David Crayford) writes:
> On a z13 that's stretching plausibility to breaking point!

last several mainframe generations haven't even bothered to give rate,
just relative to previous generation:

z900, 16 processors, 2.5BIPS (156MIPS/proc), Dec2000
z990, 32 processors, 9BIPS, (281MIPS/proc), 2003
z9, 54 processors, 18BIPS (333MIPS/proc), July2005
z10, 64 processors, 30BIPS (469MIPS/proc), Feb2008
z196, 80 processors, 50BIPS (625MIPS/proc), Jul2010
EC12, 101 processors, 75BIPS (743MIPS/proc), Aug2012

z13 is 30% increase (system) throughput (over EC12) or about 100BIPS
with 40% increase in no. of processors or about 710MIPS/proc.

23,000 TIPS would be around 230,000 max. configured z13s

a cloud megadatacenter staple has been e5-2600 blade, current e5-2600v4
blade is somewhere around 1.5TIPS ... so that would be 15333 systems (a
cloud megadatacenter tends towards half million systems), in
high-density configurations might be around 300-400 racks. There have
been past articles about organizations spinning up supercomputer for a
couple hrs (that would rank in the top 50 in the world) from cloud
operator (automated, w/o manual intervention), using online credit card
transactions.

HPC configurations have been juiced with graphic co-processors ... lots
of discussion, depending on application can be 100-400 times speedup
https://en.wikipedia.org/wiki/General-purpose_computing_on_graphics_processing_units
some benchmarks
http://wccftech.com/ultimate-cpu-gpu-floating-point-performance-battle-amd-intel/
http://www.tomshardware.com/reviews/geforce-gtx-750-ti-review,3750-16.html

-- 
virtualization experience starting Jan1968, online at home since Mar1970

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to