| From: | Shane Ponting |
| Date: | 27 Aug 2001 at 19:51:48 |
| Subject: | Re: Reasons for benchmark difference - expected or not ?? |
> the problem with benchmarks is that they belong to statistic (meaning there
> is two kinds of lies
> ordinary lies and statistic and the biggest is statistics)
Running dnet also reduces GLQuake performance by at least 1FPS - real world enough for ya ?
I just want to know if thats to be expected or can some adjustments be made (Like turning all other WOS tasks 'off' when running an app that really needs every scrap of bandwidth it can get? )
Setup:
A1200:BlizzardPPC:603p@266Mhz/68040@30Mhz
192mb of 60ns RAM (Set to 66.6~mhz)
Fast SCSI-II Controller (Overclocked)
-Quantum Atlas II 4.3gig HD 7200RPM
-Mustek 12000CX 1200DPI Scanner
-32x NEC SCSI CDROM Drive
-2xWrite 4xRead HP 4020i CD-R Drive
8mb BVision GFX Card (Permedia 2)(Overclocked)