mip@massormetrix.ida.liu.se (Mikael Patel) (12/12/89)
I'am looking for some appropriate benchmarks for a C-based Forth I'am
working on. I have run the sieve benchmark (both versions) and the
fibonacci function and now looking around for some more.
Do you have any others? Please post them!
Also how should we deal with comparison of threading methods, processor,
cache, clock rate, bit-width, etc? Some suggestions? For instance, a
poor threading method (as using C as the implementation language) can
be "compensated" for by using a faster processor e.g. workstation.
Last, how should we construct Forth benchmarks? What properties do we
want to measure? And how can we measure them accuratly?
Mikael R.K. Patel
Researcher and Lecturer
Computer Aided Design Laboratory (CADLAB)
Department of Computer and Information Science
Linkoping University, S-581 83 LINKOPING, SWEDEN
Phone: +46 13281821
Telex: 8155076 LIUIDA S Telefax: +46 13142231
Internet: mip@ida.liu.se UUCP: {uunet,mcsun,...}!liuida!mip
Bitnet: MIP@SELIUIDA SUNET: LIUIDA::MIPkoopman@a.gp.cs.cmu.edu (Philip Koopman) (12/12/89)
In article <1475@massormetrix.ida.liu.se>, mip@massormetrix.ida.liu.se (Mikael Patel) writes: > I'am looking for some appropriate benchmarks for a C-based Forth I'am > working on. I have run the sieve benchmark (both versions) and the > fibonacci function and now looking around for some more. > > Do you have any others? Please post them! > > Also how should we deal with comparison of threading methods, processor, > cache, clock rate, bit-width, etc? Some suggestions? For instance, a > poor threading method (as using C as the implementation language) can > be "compensated" for by using a faster processor e.g. workstation. > > Last, how should we construct Forth benchmarks? What properties do we > want to measure? And how can we measure them accuratly? Benchmarks are really tough in any language on any processor. Strictly speaking, sieve and fib are more toys than benchmarks. They fit completely into cache on almost any machine having cache, and have a very limited mix of instructions. Synthetic benchmarks such as Dhrystone aren't very useful, because they are tuned to conventional languages such as C and don't make sense as Forth code. I don't have an answer for you, but you should know that you are embarking on a long and arduous journey if you are looking for real, meaningful benchmarks. This is a problem that even the conventional computing community has been struggling with for a long time. Phil Koopman koopman@greyhound.ece.cmu.edu Arpanet 2525A Wexford Run Rd. Wexford, PA 15090 Senior Scientist at Harris Semiconductor. I don't speak for them, and they don't speak for me.
u803535@lanl.gov (Wayne A. Vieira) (12/14/89)
In article <1475@massormetrix.ida.liu.se> mip@massormetrix.ida.liu.se (Mikael Patel) writes: >I'am looking for some appropriate benchmarks for a C-based Forth I'am >working on. I have run the sieve benchmark (both versions) and the I have been looking for some benchmarks for C-Forth also. I have gotten the thing running on Cray's, and wanted to be able to check it against other languages available on UNICOS (Cray Unix) Thanks in advance, Wayne A. Vieira Cray Research Inc. waynev@craywr.cray.com -- Wayne A. Vieira Cray Research Inc. waynev@craywr.cray.com