[email protected] (Parwez Hamid) writes:
> z14 Key H/W high-lights:
>
> Up to 170 Customer PUs @ 5.2 GHz each on a 14 nm 10 core chip
> Up to 32 TB Memory
> Uni = 1832 'mips', 170-way = 146462 'mips'

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, 141 processors, 100BIPS (710MIPS/proc), Jan2015
...
z14, 170 processor, 146.5 BIPS, (862MIPS/proc - half uni), Aug2017

z196 documentation claims that half the per processor performance
improvement (compared to z10), is the introduction of out-of-order
execution (compared to being used for decades in other processors) ...
i.e. half of 156MIPS increase from 469MIPS to 625MIPS. out-of-order
helps to mask huge latency in memory access ... potentially allowing
execution of other instructions while waiting on cache miss.

added to out-of-order execution are branch prediction and speculative
execution ... 360/195 had just out-of-order execution ... but
conditional branches drained the pipeline ... most codes ran at only
half the 195 rated mip-rate (5mips rather than 10mips).

Current latency to memory, when measured in number of processor cycles
... is comparable to 60s disk access latency, when measured in number of
60s processor cycles.

almost 18yrs, the number of processors increase by factor of ten times,
while per processor performance increase by 5.5 times ... overall
increase 58.6 times.

-- 
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