fkittred@bbn.com (Fletcher Kittredge) (03/29/91)
In article <7401@titcce.cc.titech.ac.jp> mohta@necom830.cc.titech.ac.jp (Masataka Ohta) writes: > >I have a copy of a paper on PA-RISC 1.1 published in "Spring '91 IEEE >COMPCON Digest of Papers". > >According to the paper, the performance of PA-RISC 1.1 Workstation is: > >Clock Cache Dhrystone1.1 LINPACK SPECmarks >(MHz) (I/D) VAX/MIPS MFLOPS Overall Integer Float >50 128K/256K 59.0 13.7 42.8 38.1 46.2 >66 128K/256K 76.0 18.1 55.3 49.9 60.1 Over the years I have a wide variety of the current "fastest" workstations on my desk from a variety of vendors. I have become cynical about the supposed speed of these boxes. Frequently, the CPU speed is overshadowed by bottlenecks elsewhere in the system. My impression is that "fast" workstations have slow network, memory and disk throughput. After six weeks, the things that impress me about the 9000/720 (and I am impressed) are: 1) It really is as fast as they claim. I have used it heavily as a development box for porting and general development of a large statisical analysis application. Both the normal development process and our application benchmarks show that it really is 42 SPEC marks. 2) It seems to be a balanced system with faster net and disk i/o to match the CPU speed. 3) The speed increase is such that a new style of computing is possible. 2 years ago, 4 to 7 SPEC marks wasn't bad for a workstation, this beast is an order of magnitude faster. I describe it has the first workstation I have seen capable of really supporting X. HP has this thing called VUE which is supposed to be a GUI/WM for Unix of higher quality than what is available on the Mac. On the Series 300 and 800, VUE is a dog and we don't run it. On the Series 700, VUE flies and is a large step forward in GUIs. 4) HP was able to get this speed increase without abandoning the previous architecture. It is possible to run our old software with the new systems. As a software developer, it has opened my mind to the potential styles of computing a cheap workstation (~12,000$) can support. regards, fletcher Fletcher Kittredge Platforms and Tools Group, BBN Software Products 10 Fawcett Street, Cambridge, MA. 02138 617-873-3465 / fkittred@bbn.com / fkittred@das.harvard.edu
mohta@necom830.cc.titech.ac.jp (Masataka Ohta) (03/30/91)
In article <63455@bbn.BBN.COM> fkittred@spca.bbn.com (Fletcher Kittredge) writes: >>I have a copy of a paper on PA-RISC 1.1 published in "Spring '91 IEEE >>COMPCON Digest of Papers". >> >>According to the paper, the performance of PA-RISC 1.1 Workstation is: Sorry about that the performance figure was not the newest. I couldn't have post the newest one because of NDA. But it was, I think, fast enough. >2) It seems to be a balanced system with faster net and disk i/o to match the >CPU speed. I don't think so. HP 700's network interface (ethernet) is already saturated. To be a balanced system, it should have a faster network interface such as FDDI. But, it is not a serious problem because a EISA FDDI card seems to be already available (though not from HP). Masataka Ohta
campbelr@hpcuhe.cup.hp.com (Bob Campbell) (04/02/91)
> I don't think so. HP 700's network interface (ethernet) is already saturated. > To be a balanced system, it should have a faster network interface such as > FDDI. But, it is not a serious problem because a EISA FDDI card seems to be > already available (though not from HP). An EISA card for ethernet is availible that has better performance than the standard one. Also, EISA FDDI has been announced as being available later this year. For official information and/or big glossy info sheets, contact your local HP sales office (in U.S. call 1-800-752-0900). --------------------------------------------------------------------------- Bob Campbell Some times I wish that I could stop you from campbelr@hpda.cup.hp.com talking, when I hear the silly things you say. Hewlett Packard - Elvis Costello