ATA 100 vs SATA II

[Dennis]

- I have a WinXP system running on a 250gig ATA 100 drive.

- I got a new SATA II 250gig drive today. I need more storage.

- I'm trying to decide if I should reinstall WinXP onto the SATA drive
and start from scratch (alot of work).. or just throw the SATA drive in
the comp and leave XP running on the ATA.....

....but how much faster is it? 5%? or like 50%? Is it worth the effort
reinstalling XP?

...Can anyone give me a some facts or point me to some 'straight
forward' links as to how much of a speed difference I would see?


Thanks

 

[Clint]

Personally, I doubt you'd see much difference between identical SATA II and
ATA drives. However, you probably will see a performance increase by doing
a fresh install, and if had an older ATA drive, your new one might have
improved specs (bigger cache primarily, things like NCQ as well depending on
the model you bought). What type of applications are you running that
you're hopping to see an improvement in? Database? Games?

http://www.harddrivereport.com/pata_vs_ata_vs_sata_vs_ide.html

http://www.storagereview.com/

Clint

 

[Dennis]

Personally, I doubt you'd see much difference between identical SATA II and
ATA drives. However, you probably will see a performance increase by doing
a fresh install, and if had an older ATA drive, your new one might have
improved specs (bigger cache primarily, things like NCQ as well depending on
the model you bought). What type of applications are you running that
you're hopping to see an improvement in? Database? Games?

http://www.harddrivereport.com/pata_vs_ata_vs_sata_vs_ide.html

http://www.storagereview.com/

Clint

I do alot of multitasking... downloading torrents and gaming and
playing tunes... and theres alot of filesharing through the Lan... etc
etc... I'm thinking if its 5% its not worth it. If its 20% it is. The
standard links you find with a quick google search dont give a definate
answer and are confusing. IE The 'harddrivereport.com' link you
provided says:

"When you hear ATA vs SATA you now know that the difference between the
latest ATA hard drives and the newer SATA hard drives is a performance
boost of about 5%. (Considerably more for SATA II hard drives but
you'll also need several other changes within your computer to take
advantage of them). "

.....considerably more for SATA II ???? How much more is my question.

 

[sbb78247]

I do alot of multitasking... downloading torrents and gaming and
playing tunes... and theres alot of filesharing through the Lan...
etc etc... I'm thinking if its 5% its not worth it. If its 20% it is.
The standard links you find with a quick google search dont give a
definate answer and are confusing. IE The 'harddrivereport.com' link
you provided says:

"When you hear ATA vs SATA you now know that the difference between
the latest ATA hard drives and the newer SATA hard drives is a
performance boost of about 5%. (Considerably more for SATA II hard
drives but you'll also need several other changes within your
computer to take advantage of them). "

....considerably more for SATA II ???? How much more is my question.

ok consider this - real world use 2x120g hitachi ide drives in a raid 0
stipe versus a single sata2 80g on the kids computer. the sata2 gives the
same performace as the raid setup on most tasks.

i know someone will call bullshit, but give it a try IF your motherboard
supports NCQ and such. if not you are probably wasting your time.

 

[Clint]

Maybe this will help then. I have two HD's in my system. An ATA 200GB
Seagate ST3200822A drive
(http://www.seagate.com/support/disc/specs/ata/st3200822a.html) and a SATA
II 320GB ST3320620AS
(http://www.seagate.com/cda/products/discsales/marketing/detail/0,1081,759,00.html).
I just ran HD Tach on both drives.

Details:
Ave. Read Burst Speed Seek
200GB 55MB/s 91.2 MB/s 15ms
320GB 65MB/s 134.8MB/s 13ms
320GB* 65MB/s 254.4MB/s 13ms

*Seagate ships the drives with a jumper set to SATA I (1.5Gb/s). On systems
that support it, the jumper can be pulled off. Didn't know about this till
I did some searching to reply to you. Thanks! However, as you can see,
that didn't affect the average throughput of the drive.

Keep in mind that the SATA drive has a 16MB cache, while the ATA one has an
8MB cache. Really, I don't notice a difference between the drives.
Unfortunately (for you), I changed too many things on my system from the
time I was using the ATA drive as my OS drive to using the SATA one as my OS
drive. Big step up in processor, video card, even memory. But if there's
any other benchmarks you'd like me to run on both drives, send me a link,
and I'll give them a whirl.

FWIW, I use my system for writing software (Visual Studio, SQL Server, etc),
getting torrents, playing games like Oblivion.

Basically, everything I've read says that a single drive can't saturate an
ATA100 channel, much less a SATA I or II. So unless you're putting together
a RAID system, you won't see much real-world difference. One thing that is
nice about the SATA drives is the cabling. But you knew that already, since
you have one.

Clint

 

[Clint]

Can you define "real world" performance? When you load an application, it
takes as long on either one from the time you double-click the shortcut?
Benchmarks are the same? What about the rest of the system specs?

Clint

 

[Dennis]

Maybe this will help then. I have two HD's in my system. An ATA 200GB
Seagate ST3200822A drive
(http://www.seagate.com/support/disc/specs/ata/st3200822a.html) and a SATA
II 320GB ST3320620AS
(http://www.seagate.com/cda/products/discsales/marketing/detail/0,1081,759,00.html).
I just ran HD Tach on both drives.

Details:
Ave. Read Burst Speed Seek
200GB 55MB/s 91.2 MB/s 15ms
320GB 65MB/s 134.8MB/s 13ms
320GB* 65MB/s 254.4MB/s 13ms

Well I ended up reinstalling windows and swapping everything I could to
the new SATA drive just to be sure. Now that I'm back up and running
I've also done some benchmarks, and they are not terribly impressive...
How important is burst speed? It seems my SATA II is scoring about the
same as your 320GB on SATA I? Is it possible I need to remove a jumper
too? Mine's a Western Digital WD2500KS.


HD Tune: ATA-100 Benchmark

Transfer Rate Minimum : 33.8 MB/sec
Transfer Rate Maximum : 68.1 MB/sec
Transfer Rate Average : 54.8 MB/sec
Access Time : 15.3 ms
Burst Rate : 82.4 MB/sec
CPU Usage : 3.7%



HD Tune: SATA II WD2500KS Benchmark

Transfer Rate Minimum : 32.5 MB/sec
Transfer Rate Maximum : 58.6 MB/sec
Transfer Rate Average : 50.3 MB/sec
Access Time : 13.6 ms
Burst Rate : 112.1 MB/sec
CPU Usage : 2.5%



WinXP Device Manager Speed Test: ATA-100 DRIVE

Burst Speed: 86.2
Sustained Speed: 69.3



WinXP Device Manager Speed Test: SATA II WD2500KS

Burst Speed: 128.0
Sustained Speed: 60.4




.....thoughts?

 

[Clint]

The only difference between my drive running at SATA I vs II was the burst
speed. And AFAIK, that's only relevant when what the drive is sending back
to the computer is in the buffer.

Take a look at this page:
http://wdc.custhelp.com/cgi-bin/wdc...GV4dD1qdW1wZXIgc2V0dGluZ3M*&p_li=&p_topview=1

It shows that there's a similar jumper on your drive as on mine. Don't know
if it defaults on or off. You'll want to check that your motherboard
supports SATA II as well before you bother taking things apart. I really
doubt I'll see any real benefit of switching it over, but so long as it
doesn't hurt anything, I'll be happy.

What kind of drive did you upgrade from? You're getting a little better
performance on your old drive than on your new one. The
www.storagereview.com website has a pretty good database of drive reviews.
I know they've got your new drive in there; they may have your old one as
well that may give you a better comparision.

Clint

 

[jaster]

On Wed, 07 Jun 2006 20:45:42 -0700, Dennis thoughtfully wrote:

Tomshardware.com has a new harddrive interactive comparison chart. You
can compare 2 hds on several criteria. Not all drives are represented
but you can compare ATA to SATA from most hd vendors.

 

[sbb78247]

Can you define "real world" performance? When you load an
application, it takes as long on either one from the time you
double-click the shortcut? Benchmarks are the same? What about the
rest of the system specs?
Clint

very unscientific on my part, but the system was an opteron 144 cpu 1g ram
and 9800 video. and benchmarks can be manipulated to give the results you
want. take nvidia and ati for example.

overall operation whether it is loading an application, burning discs, or
what ever is about the same with either storage option. i am just telling
you that for day in and day out use on the kids system, the speed stayed the
same, i just lost the danger of the stripe set. now that you have a single
ide drive or a single sata2 i would take the sata2 for speed and the fact
the drive is newer and hopefully and i say hopefully not any where near a
failing condition. meaning a new motor versus a used motor, which would you
install for reliability?????

 

[Clint]

Well, I can't say I'm surprised by your observations. Unless you're running
operations that take a measurable length of time, you'd need to have a
considerable performance increase to notice the difference. Take, for
example, the opening of Word. Say it takes 10 seconds to open on the
"slower" system, and takes 9 seconds on the faster one. Would you notice
that? I doubt it, unless you actually try to measure it. But that's a 10%
difference, which probably translates to a 25% difference in actual HD
performance (since loading an application is more than just sucking data off
a drive).

On the other hand, you don't provide enough other information to provide any
meaningful comparision. Model number of drives involved in your casual
study, which would give things like RPM's, disk cache, etc. Given a RAID 0
array of two drives versus a single drive, all the same model and tested
under identical conditions, should yield appreciable performance increases.
There's plenty of published benchmark results out there to support that
theory. Whether those performance increase are noticable to the average
consumer is another thing entirely. Things like CD burning, downloading
torrents, playing music, surfing the net, chatting with friends are NOT
typically constrained by hard drive performance. So if the HD's aren't the
bottleneck, then you could have a 15K SCSI drive in your system, and you
wouldn't notice much of an increase in speed. The Internet connection and
your processor don't get any faster just because you put a faster HD in.

Clint

 

[Dennis]

On Wed, 07 Jun 2006 20:45:42 -0700, Dennis thoughtfully wrote:

Tomshardware.com has a new harddrive interactive comparison chart. You
can compare 2 hds on several criteria. Not all drives are represented
but you can compare ATA to SATA from most hd vendors.

That is exactly the kind of link I was looking for, I wish I had
stumbled across it before. Thanks!


http://www23.tomshardware.com/storage.html

 

[Dennis]

Ave. Read Burst Speed Seek
200GB 55MB/s 91.2 MB/s 15ms
320GB 65MB/s 134.8MB/s 13ms
320GB* 65MB/s 254.4MB/s 13ms

Clint: Thanks for the link... I checked my jumpers, there are none,
meaning its set on the factory default, which should be SATA 3gb. How
do I prove this though...? I still dont understand how your burst
speed is 254mb/s... Is that normal or is mine at 130? What is the
equivalent term for 'burst speed' on Tom's chart?

http://www23.tomshardware.com/storage.html

 

[Geoff]

I built a p2b-ds, dual pentium 600's back in 1998, 1 gig of ram, running
win2000. I have been very happy with it and changed it last year for an
a8v, amd 3200, running xp.

When I ran spinrite on my 2 80 gig HD's, each disk took 12 hours. These
disks are ata-100's, western digital. I'm using the same disks with the
a8v, spinrite now takes 3 hours for each disk. A pretty big performance
boost.

-g

 

[Clint]

What motherboard do you have? Does it support SATA II?

I believe the Burst Speed that I gave would be equivalent to the Interface
Performance in the benchmark. I don't believe they had a SATA II compatible
MB when they did their testing. If you check out the more detailed review
they did (http://www.tomshardware.com/2005/09/27/round/), the ASUS board
used doesn't support SATA II.

Also check out http://www.anandtech.com/storage/showdoc.aspx?i=2454 and
finally http://www.behardware.com/art/imprimer/624/. They did some testing
on the SATA 3Gb/s rating. I really don't think it matters much, since the
cache would be emptied out so quickly anyway.

Clint

 

[Craig Palme]

you can't just use something like ghost to clone the ata onto the sata the
set the sata as the boot drive?

 

[John Weiss]

Well I ended up reinstalling windows and swapping everything I could to
the new SATA drive just to be sure. Now that I'm back up and running
I've also done some benchmarks, and they are not terribly impressive...
How important is burst speed? It seems my SATA II is scoring about the
same as your 320GB on SATA I? Is it possible I need to remove a jumper
too? Mine's a Western Digital WD2500KS.

HD Tune: ATA-100 Benchmark

Transfer Rate Minimum : 33.8 MB/sec
Transfer Rate Maximum : 68.1 MB/sec
Transfer Rate Average : 54.8 MB/sec
Access Time : 15.3 ms
Burst Rate : 82.4 MB/sec
CPU Usage : 3.7%

HD Tune: SATA II WD2500KS Benchmark

Transfer Rate Minimum : 32.5 MB/sec
Transfer Rate Maximum : 58.6 MB/sec
Transfer Rate Average : 50.3 MB/sec
Access Time : 13.6 ms
Burst Rate : 112.1 MB/sec
CPU Usage : 2.5%

WinXP Device Manager Speed Test: ATA-100 DRIVE

Burst Speed: 86.2
Sustained Speed: 69.3

WinXP Device Manager Speed Test: SATA II WD2500KS

Burst Speed: 128.0
Sustained Speed: 60.4

....thoughts?

The physical limitations of a single 7200 RPM HD will override any bus speed
improvements, unless your SATA bus is "free" and your ATA bus is saturated.
Bottom line is that you are not likely to see any perceivable improvements
with a 7200 RPM SATA drive -- other than those inherent to the data density
on a newer HD -- unless you go to a multiple-drive RAID configuration.

If you need the performance, go with a WD Raptor (or a pair of them).

 

[sbb78247]

Well, I can't say I'm surprised by your observations. Unless you're
running operations that take a measurable length of time, you'd need
to have a considerable performance increase to notice the difference.
Take, for example, the opening of Word. Say it takes 10 seconds to
open on the "slower" system, and takes 9 seconds on the faster one. Would
you notice that? I doubt it, unless you actually try to
measure it. But that's a 10% difference, which probably translates
to a 25% difference in actual HD performance (since loading an
application is more than just sucking data off a drive).

On the other hand, you don't provide enough other information to
provide any meaningful comparision. Model number of drives involved
in your casual study, which would give things like RPM's, disk cache,
etc. Given a RAID 0 array of two drives versus a single drive, all
the same model and tested under identical conditions, should yield
appreciable performance increases. There's plenty of published
benchmark results out there to support that theory. Whether those
performance increase are noticable to the average consumer is another
thing entirely. Things like CD burning, downloading torrents,
playing music, surfing the net, chatting with friends are NOT
typically constrained by hard drive performance. So if the HD's
aren't the bottleneck, then you could have a 15K SCSI drive in your
system, and you wouldn't notice much of an increase in speed. The
Internet connection and your processor don't get any faster just
because you put a faster HD in.
Clint

no shit sherlock. but loading games and other utilities which actually use
the drive system are of equal speed. the raid stripe was in the opteron
system with 2 120g 7200 hitachis and were replaced for the boot drive with
80g sata2 western digital.

like it was stated ****wit - very unscientific but observation of real world
use. try and follow the bouncing ball next time.

hth, hand, and gfia

 

[John Doe]

A foulmouthed active troll

Path: newssvr13.news.prodigy.com!newsdbm04.news.prodigy.com!newsdst02.news.prodigy.com!prodigy.com!newscon02.news.prodigy.com!prodigy.net!pd7cy1no!shaw.ca!news.alt.net!news.alt.net!bnewspoutqueer00.burp.opseuuu.nut!$393e1fea!257.2.2.193.Mi5M4tcH
From: "sbb78247" <sbb78247 ****off.com>
Newsgroups: alt.comp.hardware.pc-homebuilt
Subject: Re: ATA 100 vs SATA II
Date: Sat, 10 Jun 2006 03:38:13 +0545
Organization: piss off
Lines: 46
Message-ID: <e6dtp9.n0.1 257.2.2.193.Mi5M4tcH>
References: <1149710772.053458.21990 f6g2000cwb.googlegroups.com> <MBGhg.33419$JX1.11001 edtnps82> <1149712863.357525.196930 f6g2000cwb.googlegroups.com> <20060608012519.200783E3A9 smtp4.wanadoo.nl> <_2Mhg.22077$I61.4469 clgrps13> <e6ab44.3p8.1 257.2.2.193.Mi5M4tcH> <YaWhg.24393$771.18444 edtnps89>
Xref: prodigy.net alt.comp.hardware.pc-homebuilt:467822

 

[sbb78247]

John Doe wrote:

nothing of significance

come on crybaby, the questions above still stand. or are you afraid?