_P_e_ar_lALegend said:
That is not true. It improve performance a lot. Best sustained transfer,
best throughput, a feel of improved performance in every day use.
Sustained transfer is not affected. (Name a drive that can approach a
sustained transfer of 100MB/s, let alone exceed 133MB/s). 70MB/s is the
absolute max I've seen from a drive.
In fact, you'd be hard pushed to consistently get a noticeable difference in
seek, transfer or any other benchmark. Unless you're only testing the cache
to host performance, which is rarely a limiting factor. Average cache size
on a quality new drive is 8 MB, with a transfer rate of around 60MB/s you
can empty it in 133ms, assuming that it's completely full with the data you
require. Hardly noticeable.
Take a 20Mb file, assume half of the cache holds that particular file, and
the drive has the following specs:
Sustained drive transfer speed of 50MB/s
Cache of 8Meg
I'll give you the benefit of doubt and assume 100% efficiency of the
interface, i.e., 133MB/s and 150MB/s transfer between drive and host, and
then a faster link to the memory controller. That accesses from the cache
are instantaneous, and that the cached and data read from the platter are
not simultaneous. I will overlap the 10ms seek with the 100ms or so of
emptying the cache.
Total transfer time for an ATA133 drive:
(20-4)/50 + 4/133 = 0.32 + 0.03 = 0.35s
Total transfer time for a SATA150 drive:
(20-4)/50 + 4/150 = 0.32 + 0.026667 ~ 0.347s
Wow, thats an improvement of a MASSIVE 1%.
If anybody can be bothered to consider the fact that it can start caching
the file whilst it is emptying the cache then go ahead and do the calcs
again. It'll be a similar number.
If u say so, u really never worked with SATA.
Pah. I have 2 SATA drives. A WD360GD and a WD2500JD
The difference between IDE and SATA is like the difference between the day
and the night.
The difference between ATA133 and SATA150 is not that great. The protocols
are mostly the same.
BTW, the interface is totally different and improved. Not
just simple drive swapping between interface ::-(((
The physical interface is of course, quite different. It uses a
differential pair, not a parallel cable. (In this respect it is like USB or
IEEE1394) Connectors and cabling are therefore much different (and better).
Other advantages include support for TCQ (like SCSI) which may help
performance under high loading as it allows the drive to re-order queued
commands to reap best performance (essentially it can reduce the seeks
required, amongst other things).
Also, the commands have error checking, something that normal ATA doesn't
have. Both of course have error checking on the data.
Thats all I can think of off the top of my head.
Perhaps you should try reading, rather than making up what you say.
Oh, and if you're comparing a 3 year old ATA drive with a brand new SATA
drive, then you're not really playing fair, are you?
Check a benchmark for say, a WD2500JB and a WD2500JD. They are the same
drive, one has ATA100 or 133, the other is SATA. I think you'll find that
the SATA has a marginally reduced performance due to the bridge required -
not noticeable without accurate benchmarks though (and thats tricky on a
drive). Native SATA drives should not differ noticeably, but I wouldn't
expect them to be slower than ATA.
The difference between night and day? Do you perchance live close to a
pole, such that you have ~6 months of light and 6 of dark (very similar
night and days)?
I look forward to your comments.
Ben
