Path: utzoo!utgpu!news-server.csri.toronto.edu!bonnie.concordia.ca!nstn.ns.ca!news.cs.indiana.edu!know!zaphod.mps.ohio-state.edu!wuarchive!emory!gatech!usenet.ins.cwru.edu!eagle!gonzo!tohanson From: tohanson@gonzo.lerc.nasa.gov (Jeff Hanson) Newsgroups: comp.benchmarks Subject: Re: VGX benchmark redux Keywords: VGX, GPC, PLB Message-ID: <1991Apr10.123750.4232@eagle.lerc.nasa.gov> Date: 10 Apr 91 12:37:50 GMT References: <1991Mar28.213128.9355@hellgate.utah.edu> <1991Apr1.154902.17858@odin.corp.sgi.com> <1991Apr9.154616.1976@hellgate.utah.edu> Sender: news@eagle.lerc.nasa.gov Reply-To: tohanson@gonzo.lerc.nasa.gov (Jeff Hanson) Organization: NASA/Lewis Research Center, Cleveland Lines: 88 Rich Thomson writes (and makes some good points, too.) [ ... stuff deleted ... ] > Sadly the graphics community does not yet have the > equivalent of the specmark rating on which to intelligently compare > different platforms. Just look at the claims made when comparing X > implementations. The customer gets left in the lurch unless they > undertake analyzing the voluminous output of x11perf to find out the > real story. Any interested in x11perf benchmarking and/or information on PLB benchmark should get the following publication. HP Apollo 9000 Series 700 - Performance Brief (5091-1137E 3/91). In it you will find x11perf organized into 4 groups as proposed by Digital Review. (I wrote to DR urging them to make their programs available that organize the data and draw the Kiviat graph, no reply so far. Perhaps HP could make this available.) You will also find the preliminary PLB numbers that were published in the January issue of the Anderson Report. These numbers were also published in Unix Today. I urge anyone involved in graphics and benchmarking to get more information about PLB because you will be able to create PLB benchmarks and run them in the very near future (say 6 months, max). A brief synopsis is below. The Picture-Level Benchmark - The Industry's Solution for Measuring Graphics Display Performance. What is the PLB - The PLB is a software package that provides a standard method of measuring graphics display performance for different hardware platforms. It consists of three elements: The Benchmark Interface Format (BIF), a standardized file structure that allows users to port application geometry and actions the geometry will perform to the PLB program. The Benchmark Timing Methodology (BTM), which provides a consistent method of measuring the time it takes for hardware to display and perform actions on a user's application geometry. The Benchmark Reporting Format (BRF), which provides a standardized report that allows "apple-to-apple" comparisons of graphics display performance for different hardware platforms. How do you use the PLB? - The first step is to translate your data sets from a typical application into the standard BIF. Once your data set has been translated, you are ready to run performance test. At the vendor's site or your own, you can view your data set as it runs on the vendor's system. The viewing is important, since the PLB does not measure image quality -- it is up to you to make these visual comparisons among the different systems you test. For more information contact: NCGA Technical Services and Standards 2722 Merrilee Drive, Suite 200 Fairfax, VA 22031 Phone: 703-698-9600, ext. 318 Fax: 703-560-2752 [ ... stuff deleted ... ] > Also, at a recent VGX demonstration at > the U, the rep couldn't tell me details about the figure, nor could he > show me a program with a high polygon rate. He also didn't have any > models with several hundred thousand (say, 40% of the peak figure, > or 300K - 400K polygons) polygons, although he's a sharp enough man > that I imagine he WILL have them next time in case I'm there. ;-} The powerflip program accepts several models so you can load up a few thousand polygons. It also gives the polygons/second. > Hopefully, when the Graphics Performance Committee releases its > Picture Level Benchmark program (& numbers come forth from vendors) > this situation will be alleviated. For now, we are stuck with > comparing performance numbers from each different vendor and > attempting to infer useful comparisons from widely differing measures. Beat on your vendor of choice for PLB numbers. User demands shall be heard! [ ... stuff deleted ... ] -- *-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-* \ / \ / \ / \ / \ / \ / Jeff Hanson \ / \ / \ / \ / \ / \ / * ViSC: Better * tohanson@gonzo.lerc.nasa.gov * * * * * * / \ / \ Science / \ / \ NASA Lewis Research Center / \ / \ Through / \ / \ * * * * * * * Cleveland, Ohio 44135 * * * Pictures * * \ / \ / \ / \ Telephone - (216) 433-2284 Fax - (216) 433-2182 \ / \ / \ / *-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*