lkml.org 
[lkml]   [1997]   [Jul]   [11]   [last100]   RSS Feed
Views: [more markup]  [less markup]  [headers]  [forward] 
 
Messages in this thread
/
DateFri, 11 Jul 1997 14:43:35 +0200 (CEST)
FromAndrea Arcangeli <>
SubjectRe: Linux on AMD K6
On Thu, 10 Jul 1997, linux kernel account wrote:

>
>
>On Thu, 10 Jul 1997, Christoph Lameter wrote:
>
>> In article <19970710123233.38080@test.legislate.com> you wrote:
>> : Vadim E. Kogan <vadim@vadim.biz-net.net> wrote:
>> : > 2. SCSI doesn't make big difference if you have enougth RAM
>> 
>> : The impact of IDE on a web server with enough RAM isn't latency
>> : or transfer rate -- it's cpu lossage.  With a typical IDE setup,
>> : you can drop packets from this.
>> 
>> And that only happens if you have not configured the harddrive to either
>> 
>> A. Use DMA
>> 
>> B. allow interrupts during transfer (hdparm command)
>> 
>> IDE is faster and more stable than SCSI especially with modern motherboards.
>> 
> Try not to use absolutes, it indicates ignorance. Yes, in certian places
>IDE can be better.. But if you enable DMA on many current IDE drives you
>will risk corruption.. [..]

You not risk corruptions using DMA on IDE, the kernel use DMA as default
at boot (compiling the kernel with the Tritron Chipset support). You risk
HD corruption using -u flag of hdparm (but for me and all my friends works
fine :). 

> [..]

Andrea Arcangeli


\
 
 \ /
  Last update: 2005-03-22 13:39    [from the cache]
©2003-2008