PCBUILD Archives

Personal Computer Hardware discussion List

PCBUILD@LISTSERV.ICORS.ORG

Options: Use Forum View

Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
From:
John Sproule <[log in to unmask]>
Reply To:
PCBUILD - Personal Computer Hardware discussion List <[log in to unmask]>
Date:
Sun, 11 Dec 2005 09:07:05 -0500
Content-Type:
text/plain
Parts/Attachments:
text/plain (61 lines)
By "disk read bandwidth", I believe that David is referring to how quickly
data can be read from the hard disk and made accessible to the memory and
processor to work on.  Although there a number of factors that will
determine this, the primary hard drive specification for determining this is
the platter speed.  Besides the reviews of hard drives found on your usual
hardware review sites, there are also sites dedicated to the review of hard
drives, such as http://www.storagereview.com/.

With reference to David's comments about the Celeron's amount of cache,
limiting the CPU's cache has always been the primary way that Intel has
hobbled their Celeron versions of their Pentium processors.  The previous
version of the Celeron, based on the Northwood Pentium 4 processor were
extremely hobbled in this regarded.  While the amount of L1 cache on the
original P4 based Celeron was the same as the Northwood P4, that is 8KB, the
amount of L2 cache was cut from 512KB on the Pentium 4 to only 128KB on the
Celeron.  With the move to Prescott Pentium 4s and the Celeron D, based on
this core, there is some improvement.  The Prescott Pentium 4 now has 16KB
of L1 cache and 1MB of L2 cache, and the Celeron D has the same 16KB of L1
cache and 256KB of L2 cache.  While this still hobbles the Celeron D,
compared to a full Pentium 4, these increases in cache size made bigger
improvement to the Celeron D's performance than they did to the Pentium 4
(comparing the Northwood based processors to the Prescott based processors.)

A couple of sources for more information.

Anandtech's CPU "cheatsheet":
http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2178&p=3

Anandtech's review of the Celeron D (with some comparison numbers for the
previous version of the Celeron and for the Pentium 4):
http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2093&p=1

With regard to the above comments, I guess I should mention that the
previous results, which I found running MD5checksum on a large file, were
done with a Athlon XP mobile processor, which has 512KB of L2 cache onboard,
and the hard drive I'm using is a 7200rpm, EIDE model (with 8mb of cache).

It may be that the second run of the checksum on the same file, which I
mentioned doing, went from taking about 24 seconds to only 4 or 5 seconds,
because the 700mb file was now stored in RAM (since I have 1.5GB of RAM),
and the need to read the file from the disk was by-passed.

John Sproule

----- Original Message -----
From: "Venkat Viswanathan" <[log in to unmask]>
To: <[log in to unmask]>
Sent: Friday, December 09, 2005 9:25 AM
Subject: Re: [PCBUILD] Processor speed question


> Many thanks to David Gillette and John Sproule for their posts. it was
> remiss of me not to mention the system memory in my original post. it is
> just 512 MB DDR if it matters at all.
>
> on a side note a few lines on "disk read bandwidth" will be appreciated.

        The NOSPIN Group has added a new feature on our website,
           web based bulletinboard for questions and answers:
              Visit our sister website at http://nospin.com

ATOM RSS1 RSS2