Mistake in caculating Benchmark Average


 
Thread Tools Search this Thread
UNIX Standards and Benchmarks UNIX & LINUX Benchmarks (Version 3.11) UNIX Benchmarks Mistake in caculating Benchmark Average
# 1  
Old 02-27-2006
Mistake in caculating Benchmark Average

There is a basic principal that you cannot average averages. The overall result for the benchmark should be calculated by dviding the sum of the target results by the sum of the baseline results. This is not the same as dividing the sum of the index results by 6. Changing the way this is calculated will give a more mathematically valid way of comparing performance from one host to the other.
Login or Register to Ask a Question

Previous Thread | Next Thread

8 More Discussions You Might Find Interesting

1. UNIX for Dummies Questions & Answers

Can anyone help me to spot my mistake?

Hi there can anyone help me to spot my mistake and please explain why it appears My code : #!/usr/bin/gawk -f BEGIN { bytes =0} { temp=$(grep "datafeed\.php" | cut -d" " -f8) bytes += temp} END { printf "Number of bytes: %d\n", bytes } when I am running ./q411 an411 an411: ... (6 Replies)
Discussion started by: FUTURE_EINSTEIN
6 Replies

2. UNIX Benchmarks

unix benchmark thread?

Type: UltraSPARC IIIi 1,593 Mhz x2 Ram: 16G Disk: 2*70G fw scsi drives Load: db application kernel: Sunos5.10 pgms: compiled Sun cc -O2 ============================================================== BYTE UNIX Benchmarks (Version 3.11) System -- SunOS sun.spmbox.com 5.10... (2 Replies)
Discussion started by: mr_manny
2 Replies

3. AIX

Did a Mistake with HACMP

Hi, I needed space on a FS, and when I've added the space on the filesystem, I did it trough the regular smitty fs inteface and not with smitty cl_lvm. Can someone help me to repair the situat before a faileover happen ? Thanks for your help,:mad: (13 Replies)
Discussion started by: azzed27
13 Replies

4. UNIX Benchmarks

HP hardware benchmark

CPU: 1 x PA8600, 440MHz RAM: 1GB Hardware model: 9000/800/N4000-44 BYTE UNIX Benchmarks (Version 3.11) System -- HP-UX xxx B.11.11 U 9000/800 615379343 unlimited-user license Start Benchmark Run: Tue Apr 4 05:43:42 IST 2006 1 interactive users. Dhrystone 2 without register... (0 Replies)
Discussion started by: blowtorch
0 Replies

5. UNIX Benchmarks

Precompiled Benchmark software

Hi, Is there a precompiled binary for Solaris 8 available? I need to bench mark our Oracle server, as we are upgrading from SFv880 to SFv890. Both are fully loaded. I can't find a sun machine that I can compile the software on. Tks JohnHo (0 Replies)
Discussion started by: John Ho
0 Replies

6. Linux

Benchmark for Linux

I work in a computer company which sells computer configurations and parts of them. And I want to give a choice to customers. If they want to buy a PC with Linux installed, not Windows. But I find difficult to test the Graphic Cards in Linux OS. I have searched the web and I didn't found any... (2 Replies)
Discussion started by: vlatkop
2 Replies

7. UNIX Benchmarks

AMD-K5 Benchmark

AMD-K5 Processor at 133Mhz 32MB RAM 5 GB Hard Drive 10MB NIC 1MB ARC Graphics Card PS/1 Keyboard CD-ROM Floppy Drive Kickin' Fast BABY! hehe BYTE UNIX Benchmarks (Version 3.11) System -- FreeBSD evil-linux.net 5.0-RELEASE FreeBSD 5.0-RELEASE #0: Thu Jan 16 22:16:53 GMT 2003... (0 Replies)
Discussion started by: Phobos
0 Replies

8. UNIX Benchmarks

Ultra 10 benchmark

CPU/Speed: Ultrasparc IIi / 300MHz Ram: 128MB (Not enough) Motherboard: Sun Ultra 5/10 Bus: 4 PCI Cache: 512K Ecache Controller: Onboard IDE ATA/33 Disk: 40GB IBM ATA/100 Load: Low, 1 user, apache, samba, ipf, dtlogin disabled. Kernel: Solaris 5.10 b72 Kernel ELF?: yes pgms: I used the... (0 Replies)
Discussion started by: 98_1LE
0 Replies
Login or Register to Ask a Question
MINTEGRATE(1)							   User Commands						     MINTEGRATE(1)

NAME
mintegrate - evaluate average/sum/integral/derivative of 1-d numerical data SYNOPSIS
mintegrate [OPTION]... [FILE] DESCRIPTION
mintegrate is a program to compute averages, sums, integrals or derivatives of numerical 1-d data in situations where ultimate numerical precision is not needed. OPTIONS
-a compute mean value (arithmetic average) and standard deviation -c compute integral on closed x-data interval; In case that dx is not specified by the '-d' flag, the data are supposed to be from an irregular x-grid, and dx is computed separately for every x-interval. The integral is computed by the trapezoidal rule. -d <float> compute integral on open x-data interval with the specified dx; Can be used also in combination with '-D' and '-c'. -D compute difference btw. numbers or derivative of the y-data; In the default scenario where x- and y-data column are same, the dif- ference btw. the current and the previous data value will be output. In this case when '-d' is defined as 0, the x-data value will be print out in front of the calculated difference. If x-and the y-column are different and if the x-data resolution is not defined or it is !=0, then the derivative of the y-data is calculated. When the x-data resolution is constant, specify it explicitly by '-d' to achieve a higher numerical precision by a 'leapfrog' algorithm. -x <int> x-data column (default is 1). If 0, the x-range is an index; -y <int> y-data column, where y=f(x) (default is 1) -r x_0:x_1 x-data range to consider -s print out accumulated y_i sums: x_i versus accumulated f(x_i); In the case of a closed integral you have to specify also the x-data resolution dx (see '-d' above). -S compute the accumulated y_i-sums and add it to the output -p <str> print format of the result ("%.10g" is default) -t <str> output text in front of the result (invalid with '-s' or '-S'); A blank can be printed by using a double underscore character '__'. -T run a self-test that the program is working correctly -V print version number --version output version and license message --help|-H display help -h display short help (options summary) If none of the options '-a', '-D', '-d', or '-c' is used, then the sum of the provided data will be computed. Empty lines or lines starting with '#' are skipped. This program is perfectly suitable as a basic tool for initial data analysis and will meet the expected accuracy of a numerical solution for the most demanding computer users and professionals. Yet be aware that, although the computations are carried with double floating pre- cision, the computational techniques used for evaluating an integral or a standard deviation are analytically low-order approximations, and thus not intended to be used for numerical computations in engineering or mathematical sciences for cases where an ultimate numerical pre- cision is a must. For deeper understanding of the topic see http://en.wikipedia.org/wiki/Numerical_analysis. COPYRIGHT
Copyright (C) 1997, 2001, 2006-2007, 2009, 2011-2012 Dimitar Ivanov License: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html> This is free software: you are free to change and redistribute it. There is NO WARRANTY, to the extent permitted by law. mintegrate 2.2.1 February 2012 MINTEGRATE(1)