Performance of 4.x vs 5.x (nbench results)

[Kris Kennaway]

On Tue, May 24, 2005 at 09:31:17PM +0200, Bohdan Horst wrote:
> 
> 
> I have 4 identical PC with FreeBSD (2x4.11R and 2x5.4R)
> 
> Results (/usr/ports/net/benchmarks/nbench):
> 
> (CPUTYPE=p3 in /etc/make.conf; portupgrade -vf nbench)
> 
> tests:
> NUMERIC SORT:  4.11= 499.52 494.31  5.4= 451.26 449.7  (4.11 faster) 
> STRING SORT:   4.11= 24.13  24.11   5.4= 25.286 25.294 (5.4  faster)
> BITFIELD(e+8)  4.11= 1.133  1.1319  5.4= 1.1239 1.1239 (4.11 faster)
> FP EMULATION   4.11= 18.951 18.755  5.4= 25.044 25.029 (5.4  faster)
> FOURIER        4.11= 5361   5356.3  5.4= 5076.3 5076.3 (4.11 faster)
> ASSIGNMENT     4.11= 5.4795 5.4469  5.4= 5.8986 5.8968 (5.4  faster)
> IDEA           4.11= 595.23 592.13  5.4= 660.65 660.65 (5.4  faster)
> HUFFMAN        4.11= 332.47 330.91  5.4= 409.06 408.95 (5.4  faster)
> NEURAL NET     4.11= 7.0374 6.9882  5.4= 7.2976 7.2955 (5.4  faster)
> LU DECOMP.     4.11= 225.94 226.76  5.4= 211.56 210.84 (4.11 faster)
> original:
> INTEGER INDEX  4.11= 12.305 12.238  5.4= 13.409 13.400 (5.4  faster)
> FLOATING-POINT 4.11= 9.309  9.296   5.4= 9.052  9.041  (4.11 faster)
> new:
> MEMORY INDEX   4.11= 3.321  3.313   5.4= 3.448  3.448  (5.4  faster)
> INTEGER INDEX  4.11= 2.895  2.873   5.4= 3.271  3.268  (5.4  faster)
> FLOATING-POINT 4.11= 5.163  5.156   5.4= 5.020  5.014  (4.11 faster)
> 
> 
> 
> all 4 hosts have important services (w3cache, firewall, ircd, etc) and i
> cannot turn them off to perform good UFS1/UFS2 tests :(

Thanks, but if the machines are not completely idle then it's
impossible to tell whether these numbers are meaningful :-(

Kris
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 187 bytes
Desc: not available
Url : http://lists.freebsd.org/pipermail/freebsd-stable/attachments/20050524/df161e3e/attachment.bin