Date: Thu, 9 Jan 1997 22:38:16 +1030 (CST) From: Michael Smith <msmith@atrad.adelaide.edu.au> To: mark@salford.ac.uk (Mark Powell) Cc: lcs@jumpnet.com, freebsd-scsi@freebsd.org Subject: Re: Can the P/I-P55T2P4 be overclocked? In other words does it support 75 and 83MHz bus speeds? Message-ID: <199701091208.WAA26952@genesis.atrad.adelaide.edu.au> In-Reply-To: <Pine.BSF.3.95.970109095942.18913A-100000@plato.salford.ac.uk> from Mark Powell at "Jan 9, 97 10:27:22 am"
next in thread | previous in thread | raw e-mail | index | archive | help
Mark Powell stands accused of saying: > > > Correct, only the plain P55T2P4 rev. 3.0/3.1 have everything needed to do 75 & 83Mhz, > > but I seriously doubt that a heavily loaded 2940 will perform well above 33Mhz PCI bus > > speeds.... > > Why's that? Also, what d'you mean by "well"? A good question. Providing the bus interface on the aic7xxx part is up to it, it might run faster. Bear in mind that it's only designed for 33MHz though. > > The SC-200 and the SC-875 stand a much better chance of running well at higher PCI > > bus speeds when drives are attached that fully use their capabilities. > > Not too sure on the quality of the FreeBSD drivers for the NCR though. And The FreeBSD NCR drivers are excellent. Note that experimental evidence tends to indicate that the 2940 is better under multi-drive loads than the NCR, but by a fairly small margin. Also note that many so-called "onboard 2940" controllers are actually lower-spec parts that won't perform anywhere near as well. Basically, we have the following Adaptec parts : aic7880 : AHC398xU, AHC3940U, AHC2940U aic7870 : AHC398x, AHC3940, AHC2940 aic7860 : AHC2940AU, "onboard 2940" aic7850, aic7855 : "onboard 2940" If at all possible, demand to know which chip is actually used. The comment "when drives are attached that fully use their capabilities" is a bit of a curious one though. Drives don't use a controller's capabilities; the driver does, in order to use the drive's capabilities. > I don't think the NCR 53c875 is supported at all, just 810 & 825 :( The FreeBSD/NetBSD NCR driver supports the 810, 815, 820, 825, 860 and 875. > Suppose, I could just use narrow drives instead of the wides I > currently have? Hmmm, I really wanted to overclock a Cyrix 6x86 > P166+ to a P200+, and this seemed to be the way to do it from > other's expericence. Why? Unless you are going to be doing things that are totally CPU bound, you should be more worried about memory and I/O throughput than plain CPU cycle time. Going to narrow drives will just defeat this. > Do you have any idea whether the DFI G586VPS Pro (using VLSI Lunx chipset) > or the or the MTech R534 (http://www.mtiusa.com/r534.htm), using the > Sis5571 chipset which I think I read somewhere "allows the PCI bus to be > locked at 32MHz [sic?] whatever the external clock speed", would be better > for this purpose? I would be demanding datasheets on the chipsets in question and studying the timing values programmed by the BIOS for these boards in comparison with the Intel chipsets and their recommended timing. The 430FX chipsets push things pretty hard already; I'm skeptical that these other newcomers are likely to be more than marginally better. > Mark Powell - Unix Information Officer - Clifford Whitworth Building -- ]] Mike Smith, Software Engineer msmith@gsoft.com.au [[ ]] Genesis Software genesis@gsoft.com.au [[ ]] High-speed data acquisition and (GSM mobile) 0411-222-496 [[ ]] realtime instrument control. (ph) +61-8-8267-3493 [[ ]] Unix hardware collector. "Where are your PEZ?" The Tick [[
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199701091208.WAA26952>