Date: Fri, 20 Feb 2004 14:35:25 -0500
To: Charles Swiger <cswiger@mac.com>
Here's some knobs to try, ie;
RAID5: Write Back Cache, Normal Read Ahead, and Direct I/O
RAID1: Write Through Cache, Normal Read Ahead, and Direct I/O
That picked up my write speeds just enough so it wasn't nearly as bad
writing as reading.
At 02:27 PM 2/20/2004, you wrote:
>On Feb 20, 2004, at 11:47 AM, mi+mx@aldan.algebra.com wrote:
>>Also, is the RAID5 writing speed supposed to be _so much_ worse, than
>>that of a single disk?
>
>it's normal for RAID-5 write performance to be slower than that of a bare
>drive. RAID filesystems involve tradeoffs between cost, performance, and
>reliability. RAID-5 maximizes cost and reliability at the expense of
>performance....
>
>--
>-Chuck
>
>_______________________________________________
>freebsd-performance@freebsd.org mailing list
>http://lists.freebsd.org/mailman/listinfo/freebsd-performance >To unsubscribe, send any mail to "freebsd-performance-unsubscribe@freebsd.org"
Re: On-chip, 7-way associative ... worse performance than EV68 with 16 MB of fast board-level L2 at the same clock rate in SPECint: ... the CPUs can access the whole system's cache at full speed. ... I question whether overall small-system board costs would have been any lower using EV68 than EV7: in fact, I suspect the opposite - if the components were priced reflecting HP's cost rather than reflecting what HP thought it could get for them. ... (comp.os.vms)
Re: -mmX 4G patches feedback [numbers: how much performance impact] ... i assumed a 1MB cache.... > note that it's not only the tlb flush having the cost,... this is what i called the 'secondary' cost of the TLB flush.... the pagetables, necessiating 128 cachemisses. ... (Linux-Kernel)
Re: How to develop a random number generation device ... embedded systems - I know when reliability is important, ... Looks like the number of cores per chip is at least ... by avoiding context switches, except in the sense of reliably being able ... The problem with so many cores accessing a shared cache is that you have ... (sci.electronics.design)
Re: Part 1 of 3: Micro economics 101 (was: Code density ...) ...Decoded Instruction Cache ~ Intel "trace" cache ... are reasonable things to do about the former (i.e., decoded instruction... lot of gates", or "likley to incur extra gate delays" NOT as ...cost and complexity. ... (comp.arch)
Re: How to develop a random number generation device ... Until you can come up with some sort of justification, however vague, as to why you think one cpu per process is more reliable than context switches, this whole discussion is useless. ... I'm happy to accept that doing things in hardware is often more reliable than doing things in software (I work with small embedded systems - I know when reliability is important, and I know about achieving it in practical systems). ... You have repeatedly said that current OS's (software OS's running on one or a few cores) is inherently unreliable, while your idea of a massively multi-core cpu running a task per core would be totally reliable. ... I don't imagine we'll see so very many more cores on a device, because the interconnections and cache scaling get too difficult, and the whole device is too limited in memory bandwidth. ... (sci.electronics.design)
