|
Steve, There is a certain amount of overhead consumed in managing multiprocessor systems ... the total CPW rating reflects that overhead. Doing the division by 32 gives a per processor total that assumes the overhead and that each processor is being used at 100% of capacity. As my Motor Vehicle clients say ... "Your results may vary ...". A 32 processor system will not be 32 times as fast as a one processor system ... in a commercial workload, different processors will operate at different percentages of utilization all the time. Under OS/400, a job gets dispatched to one processor, which runs it. Unless OS/400 has changed significantly (and I'm very sure that it hasn't), a 32 processor system will be no faster than a one processor system if only one job is running in the system. HTH ... let me know if you are looking for more detail on this. John At 02:47 PM 4/29/2002, you wrote: >The cpw of the new 890 looks to be about 1170 per cpu. ( 37,400 / 32 ). >The cpw per processor of the other iSeries models is about 1070 ( cpw of >the 270/2432 ). The 890 runs on a 1.3 GHz Power4 processor. The processor >of the 270 runs at less than half that speed ? > >Why is the cpw per cpu of the new 890 not higher ? > >Does the power4 processor of the 890 actually run at 1.3 GHz? > >Steve Richter > > > >i890 REGATTA SERVER > >The 32-way i890 nearly doubles the processing power of the i840 to > >37,400 CPW (see Figure 1 in the online version of this article). The > >1.3 GHz Power4 chips cram 174 million transistors onto two > >processors and enable the i890 to crank out transactions so fast > >that it consumes energy at a breakneck pace, requiring 3-phase power > >and dissipating up to 25,000 kBTUs of heat per hour. > > >The i890 triples the I/O capacity of last year's i840, supporting up > >to 72 TB of disk, 2,047 disk arms, and 256 GB memory. The machine > >also connects up to 32 Integrated xSeries Adapters (IXAs) and offers > >up to 534 MB of L2/L3 cache. > > >-- > >_______________________________________________ >This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing list >To post a message email: MIDRANGE-L@midrange.com >To subscribe, unsubscribe, or change list options, >visit: http://lists.midrange.com/cgi-bin/listinfo/midrange-l >or email: MIDRANGE-L-request@midrange.com >Before posting, please take a moment to review the archives >at http://archive.midrange.com/midrange-l.
As an Amazon Associate we earn from qualifying purchases.
This mailing list archive is Copyright 1997-2024 by midrange.com and David Gibbs as a compilation work. Use of the archive is restricted to research of a business or technical nature. Any other uses are prohibited. Full details are available on our policy page. If you have questions about this, please contact [javascript protected email address].
Operating expenses for this site are earned using the Amazon Associate program and Google Adsense.