Return to BSD News archive
Path: euryale.cc.adfa.oz.au!newshost.anu.edu.au!newshost.telstra.net!act.news.telstra.net!psgrain!news.uoregon.edu!hpg30a.csc.cuhk.hk!news.cuhk.edu.hk!agate!howland.reston.ans.net!nntp.coast.net!zombie.ncsc.mil!news.mathworks.com!tank.news.pipex.net!pipex!dispatch.news.demon.net!demon!heist.demon.co.uk!not-for-mail From: iain@heist.demon.co.uk (Iain Baird) Newsgroups: comp.unix.bsd.freebsd.misc,alt.comp.periphs.mainboard.asus,comp.sys.ibm.pc.hardware.systems,comp.os.linux.hardware,comp.sys.intel Subject: Re: PCI bus freq with CPU freq at (X * 40)Mhz? Followup-To: comp.unix.bsd.freebsd.misc,alt.comp.periphs.mainboard.asus,comp.sys.ibm.pc.hardware.systems,comp.os.linux.hardware,comp.sys.intel Date: 5 Mar 1996 13:54:04 -0000 Organization: Home Lines: 62 Message-ID: <4hhh1s$pn@heist.demon.co.uk> References: <DnpJqF.82n@info.elvisti.kiev.ua> <4hg8sa$1jn@heist.demon.co.uk> <Dns7BK.DBI@ritz.mordor.com> NNTP-Posting-Host: localhost X-NNTP-Posting-Host: heist.demon.co.uk X-Newsreader: TIN [version 1.2 PL2] Xref: euryale.cc.adfa.oz.au comp.unix.bsd.freebsd.misc:14924 alt.comp.periphs.mainboard.asus:11979 comp.sys.ibm.pc.hardware.systems:28596 comp.os.linux.hardware:32287 comp.sys.intel:66608 Chris Mauritz (ritz@ritz.mordor.com) wrote: : Iain Baird (iain@heist.demon.co.uk) wrote: : : PCI is specified to run at up to 33MHz. To accomodate faster CPU : : clocks motherboards can provide the facility for PCI to run at : : some fraction of the CPU. My motherboard (GA-486AMS) allows a : : CPU:PCI ratio of 1:1 or 2:1, configured in the BIOS setup. In : : principle other ratios could be used (e.g. 4:3), but I'm not aware : : of any motherboards which support this. : : : : Having said that, I have run PCI at 40MHz with a DX4-120, and it : : worked (with AHA-2940 and S3-968). This made little or no difference : : to kernel build times compared to 20MHz, although this is a crude : : benchmark. Doing this could reduce the life expectancy of PCI : : cards. : Reduce the life expectancy? How? My experience has been that : overclocked cpu's/cards either work or don't work from square : one...they don't work and then fail. If you have different : experience please share. I'd hate to think I'm risking expensive : cpu's and #9 graphics cards... I haven't experienced any such failures either. The risk was mentioned in a followup to an article I posted some time ago. IIRC, the gist was that if you clock something faster, it runs hotter. This *could* increase the risk of failure. I suppose it depends on the tolerances of the components used; hopefully "good quality" cards have sufficiently high tolerances. : : With 15ns cache and 70ns SIMMs I was able to run a DX4-120 with : : 0WS and 2-1-1-1 cache burst read, the fastest settings supported. : : This is probably motherboard dependent. I should have mentioned that I have both cache banks filled, so cache accesses are interleaved. : : This is the 5x86-133, it's only 160MHz if you overclock it, and then you're : : back where you started... : Really....According to AMD's web page, it's almost identical in : performance to a 486/120... : : : Now I'm seriously considering AMD 133 part (this one is 33MHz, 4x). : : : Guys have a very nice opinions on it. : : : : For PCI, the AMD 5x86-133 makes more sense. I'm running one now, : : and very happy with it. : I might try one for grins, but AMD's own benchmarks put it almost : exactly where the 486/120 is...I'm not sure I understand their : marketing... Apart from the faster clock, the AMD 5x86 has 16KB L1 cache compared to 8KB for the DX4-120. OTOH, the DX4 accesses memory at 40MHz instead of 33MHz. Which performs better probably depends on what you're using it for. The clock increase from 120 to 133 is about 11%. My kernel build times improved by about 15%. This suggests that, for this application, the larger cache is a modest win. iain