Date: Sat, 31 Oct 1998 11:37:48 -0800 From: Mike Smith <mike@smith.net.au> To: Narvi <narvi@haldjas.folklore.ee> Cc: Mike Smith <mike@smith.net.au>, Tom <tom@uniserve.com>, "Milliken, Scott" <MillikS@salestech.com>, "'stable@freebsd.org'" <stable@FreeBSD.ORG> Subject: Re: RAID support in FBSD? Message-ID: <199810311937.LAA00838@dingo.cdrom.com> In-Reply-To: Your message of "Sat, 31 Oct 1998 16:08:09 %2B0200." <Pine.BSF.3.96.981031155813.13946I-100000@haldjas.folklore.ee>
next in thread | previous in thread | raw e-mail | index | archive | help
> > On Fri, 30 Oct 1998, Mike Smith wrote: > > > > This is wrong. Max sustained bandwidth for 33Mhz 32bit PCI bus is 80MB/s. > > > 132MB/s is the theoretical peak bandwidth on an infinite burst transfer. > > > For higher bandwidth, you need either wider (64 bit) or faster (66Mhz) PCI > > > bus. > > > > You're presuming on the latency timer here, right? I don't recall > > there being a cap on the burst length. > > In the spec, probably yes. Unless you can say "definitely" yes, I'll contend that the only cap on the burst length is the latency timer. At a typical value around 80 cycles and with an overhead of ~8 cycles to set up/shut down the transfer around the timer, that only loses you 10% of your bus bandwidth. Unfortunately I left my copy of the PCI spec at my last job, so I can't quote chapter and verse. > But obviously there is a cap on the burst length. The cap is due to the > reason that PCI is not AFAIK full-duplex. Once the device has transmitted > all data you asked for, it is going to have to pause until you ask for > more. FreeBSD does transfers in amounts of max. max_phys_io, right? > > And ordinary machines have just one PCI<->host bus on which several more > things are going on (somebody moves the mouse, a network packet arrives, > something is logged to the screen, etc.). Now you're trying to cloud the issue; you originally said "Max sustained bandwidth for 33Mhz 32bit PCI bus is 80MB/s." Now you are trying to say "Maximum bandwidth in some hypothetical typical use is 80MB/s". I suggest you need to do some more math. It's certainly possible to sustain a lot more than 80MB/s, even with multiple masters on the bus. -- \\ Sometimes you're ahead, \\ Mike Smith \\ sometimes you're behind. \\ mike@smith.net.au \\ The race is long, and in the \\ msmith@freebsd.org \\ end it's only with yourself. \\ msmith@cdrom.com To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199810311937.LAA00838>