RE: Cyrix 6x86

Mike Jagdis (mike@roan.co.uk)
Mon, 25 Nov 1996 12:35:08 +0000 (GMT/BST)

> I have a rev 2.7 chip. These changes seem to work ok, but there seems to be
> NO difference in performance. In fact, even with the "6x86opt.zip" which
> turns on lots of 6x86 features, there seems to be no difference. Kernel
> compile (which is mostly CPU bound) takes just as long with as without.

Most 6x86 BIOS' seem to enable most of the performance features
anyway (not that is easy to benchmark the difference without
using somewhat contrived code). The main difference comes from
using VSPM to map the kernel and that only helps things that
bounce around in kernel space pretty regularly. Larry McVoy's
pipe bandwidth benchmark picks it up well (which also suggests
that we could boost performance in the non-VSPM case by doing
a code locality analysis of the kernel...)

Mike

-- 
.----------------------------------------------------------------------.
|  Mike Jagdis                  |  Internet:  mailto:mike@roan.co.uk   |
|  Roan Technology Ltd.         |                                      |
|  54A Peach Street, Wokingham  |  Telephone:  +44 118 989 0403        |
|  RG40 1XG, ENGLAND            |  Fax:        +44 118 989 1195        |
`----------------------------------------------------------------------'