Hi!
My machine is home-built and Vista's rating of it surprised me...
I am curious about the reasons for the scoring, and hoping that
somebody can explain.
Mobo - Abit IP35 Pro
RAM: 4 GB DDR2
CPU: Intel Dual Core E4500 (2.2 Ghz, overclocked to 3.3Ghz)
Graphics: Nvidia BFG 8600 GT
Hard drives; fairly new ones; three Samsung Spinpoints and one
Hitachi; all internal 7200rpm and Sata2.
Vista scores it like this (6 is highest score) :
Memory (RAM) 4.00 GB ----------5.9
Graphics NVIDIA GeForce 8600 GT -----------------------5.9
Gaming graphics 1663 MB Total available graphics memory
----------------5.5
Primary hard disk 158GB Free (233GB Total) --------------------5.5
BASE AVERAGE SCORE IS -------5.5---- (determined by lowest subscore)
I understand that the 3D gaming score is low-ish in this system.
But I don't get the hard drive!
What is wrong with keeping the OS on a 250GB Samsung Spinpoint SATA2
drive?
What hard drives are rated highly by Vista in it's scoring?
Looking forward to your feedback!
J.
I found an article here, and a few links at the bottom of the article.
http://en.wikipedia.org/wiki/Windows_System_Assessment_Tool
I don't see a formula for data transfer rate versus WEI, but a
check of a site that keeps sample results, showed all the users
were getting 5.9 for their disks. And they cannot all have
Raptors. That tells me, your disk sucks
I'd check the transfer mode it is in. Either benchmark the
thing. Or check the transfer mode.
HDTune can indicate the mode the disk is in. Mode for SATA
disks is relatively meaningless (or should be). (A mode may
be reported, but may not actually be affecting the results.)
My IDE drive indicates UDMA Mode 5 UltraATA 100, and that
is a function of my chipset (known to be limited to 100MB/sec).
My drive may support 133MB/sec, but the chipset prevents it
from being used. (No Intel Southbridge supports more than
100MB/sec on IDE interfaces.) The sustained transfer rate is
normally less than the cable rate, which is why the difference
between 100 and 133MB/sec doesn't bother me in the least.
http://www.hdtune.com/HDTune_Info.gif
When benchmarking, the results should be a curve, with
fastest transfer occurring at the beginning of the disk,
and slower transfer near the end of the disk. Such a
benchmark tests sustained transfer rate, and should be
media limited (data rate at the heads). If the transfer
is cable limited, then the curve will have a flat section.
The worst, is when you're in PIO mode (polled transfer),
where the CPU transfers each and every byte (no DMA is used),
and the performance dips to 4MB/sec, instead of about 70MB/sec
for a media limited transfer.
Another tool is HDTach, but I don't know which of these tools
runs in Vista.
http://www.simplisoftware.com/Public/index.php?request=HdTach
You may have to resort to Sandra, to find something that runs
in Vista. It probably wouldn't be a good idea to use the
Lite version, if you already own a licensed version of the
product. You never know how licensed and unlicensed versions
may interact. If you've never used it before, give the Lite
version a try.
http://www.sisoftware.co.uk/index.html?dir=&location=downandbuy&langx=en&a=
Sandra seems a bit pessimistic. Reports 52.7MB/sec at beginning of disk,
and 27MB/sec at the end of the disk. I also ran HDTach right afterwards,
and it reports about 58MB/sec at the beginning and 30MB/sec at the end.
Such benchmark tests are independent of file system, as they sequentially
read the raw disk to come up with the sustained transfer rate numbers.
The tests do take into account bad sectors, in the sense that if you
hit a patch of spared-out sectors, that can cause a downward glitch
in the results. For example, I tested a brand new Seagate drive against
a year old Seagate drive, using the "detailed" test in HDTach, and
one thing that impressed me, was fewer downward spikes in the transfer
rate curve. (A downward spike doesn't *guarantee* a bad sector, but it is
one explanation for a performance change from the norm. A marginal sector
may be re-read a number of times, until a read op with good CRC results.)
HTH,
Paul
