Access times question And esata speeds WD vs Seagate

[ds650]

How important are access times? I have a WD My Book ES & a Seagate FreeAgent Pro. (both usb2/esata models) No matter if I use the Esata, usb2 or firewire (On the Seagate only) The access times on the HD Tach are around 13.7ms for the WD, but around 20.3 ms for the Seagate. How big of deal is that? The reason I ask is I only plan on keeping one of them.

When doing the HD tach test the WD thru usb2 records:
Burst=16.5MBs
access=13.5ms
avg read=15.4 MBs
cpu=36%

The Seagate thru usb2 records:
Burst=17.2MBs
access=20.5ms
avg read=16.1MBs
cpu=29%

Pretty close I'd say. However when using esata its different. My esata is an old 1.5g card that is no longer supported. In fact both WD & Seagate fail their respective diagnostic tests when connected to the card. (tho both HDDs seem to work fine) When connected to usb2 they both pass their software tests btw.

Anyway the esata results for the WD are:
Burst=66.5 MBs
access =13.7ms
avg read= 52.9MBs
cpu=2%

The Seagate comes in at:
Burst=69.3MBs
access=21.2ms
avg read= 32.7MBs
cpu= 4%

The access time is about the same as the usb2.0 marks but the avg read is way lower on the Seagate thru esata. Any ideas why? That really bothers me because otherwise I like the Seagate better. (design, warranty & QUIET). Does anyone think the old card maybe just works with the WD better?

fwiw I record these numbers thru the fw400 on the Seagate.
Burst= 41.8MBs
access= 20.1ms
avg read= 32.9MBs (higher than esata!)
cpu=7%


Thanks for all the help guys. Sorry to sound dumb but kind of a newbie here.

[just brew it!]

Without knowing specific model numbers it's hard to say for sure what's going on here. But off the top of my head, it looks like maybe the WD drive has a faster spindle speed. That would explain both the lower access times, and the higher noise levels.

Personally I'd take pick the WD; access times can make a pretty big difference in real-world performance.

[ds650]

Without knowing specific model numbers it's hard to say for sure what's going on here. But off the top of my head, it looks like maybe the WD drive has a faster spindle speed. That would explain both the lower access times, and the higher noise levels.

Personally I'd take pick the WD; access times can make a pretty big difference in real-world performance.

They are both 7200rpm w/ 16mb buffers. Its weird how the usb2 results are close, the fw400 on the Seagate seems OK too, but the esata seems so much better on the WD for the avg read speed. I guess its got me confused. I like the WD, except it actually vibrates my desk!!! So I bought the Seagate to try. I do like it too. Mostly I just use either to backup & store data. I dont even use them everyday. Just want the most for my dollar I guess. Heck fw400 is actually fast enough if I really think about my needs! I just like esata's potential.

[just brew it!]

That's pretty weird. The access times on the Seagate look rather high for a 7200 RPM drive, FWIW. I wonder if maybe it is mislabeled?

Edit: Changed "low" to "high".

[UberGerbil]

Access time is a fundamental property of the drive. It is the combination of two values: seek time, which is the time it takes to move the heads to the correct track, and rotational latency, which is the time required to wait for the requested sector to rotate under the head. Obviously these will vary somewhat (depending on where the requested track is relative to where the head currently is at, and where the requested sector happens to be on the disk when the head gets there) which is why it is listed as an average value (and why you'll see slight variations from test to test). It isn't going to be affected, at all, by the interface used to connect the drive -- as your test results show.

Burst speed is the throughput of data read out of the drive's cache (all drives have some RAM inside them used by the drive's firmware to cache data that may be requested in the near future). That will be limited by the interface speed, since RAM is much faster than any HD interface, but it has only a limited impact on overall performace of the drive.

Transfer rate is the bandwidth the drive can deliver when streaming data from contiguous sectors and tracks on the drive (if they're not contiguous, seek comes into play again). This is purely a function of the rotational speed of the drive and the density of data on the platters, and it is lowest on the inside tracks of the drive and increases as you move towards the outermost tracks. Of course it can be limited by interface speed if the interface is extremely slow.

In your case, something is limiting the USB speeds. Hi-speed USB 2.0 should be capable of reaching close to 40MB/s (the theoretical speed of 60MB/s isn't really achievable in practice), but it's clear from your burst speed results that something -- the controller, the chipset in the drive housing, a hub that might be in between them, or the drivers -- isn't capable of that.

But that's really not important since you clearly have other connection options at your disposal.

When the interface is not limiting performance, the WD is clearly superior in access times and transfer rate, so it's really no contest. As JBI says, that looks like the difference between 7200rpm and 5400rpm drives.

[ds650]

Not sure why my usb2 speeds are low. I have an older computer & a pci usb2.0 card & a belkin hub. But my pci card is the esata 1.5g/usb2.0 combo. Its an old Adaptec card that is no longer supported. Wondering if a new 3g card would make a big diff?

I saw a test with a Vantec 3.0g esata pci card & Seagate esata HD. (It was the boxy grey one, not the Freeagent). They were getting burst rates over 100MBs & avg reads of 53MBs. My WD gets 53MBs now on the 1.5 card but only gets burst of about 65MBs. I'm wondering if the WD isn't maxed on the 1.5 or if a 3.0 would almost double the WD and Seagate both.

One other thing, they wrote in that test that the Seagate had a high access time. They said its to make it quiet I think. Might be, the WD is LOUD during the HD tach test.

Here is another test I found
http://www.cluboverclocker.com/reviews/hard_drives/seagate/Seagate_eSATA/eSATA_500gb/Page5.htm