ATA 133: is it really faster?

[hellrazrblade]

i just built my new rig (in sig) and i put a 40gig Maxtor D740X ATA133 hard drive in it (primary master). The drive is partitioned into a 10gig C and a 27gig F drive. After running HD Tach 2.60 i got the following results:
11.9ms random access time
84.7mb/s burst speed
max read speed of 43811kps
avg. read speed of 35342kps
cpu utilization of 20.5%
i put my 40gig Western Digital Caviar ATA100 hd in as the primary slave and got the following results:
14.7ms random access time
81.9mb/s burst speed
max read speed of 33308kps
avg. read speed of 24290kps
cpu utilization of 17.5%

the access and read times are notably worse on the WD, but the burst speed is almost exactly the same. It seems to me that the "extra" 33mb/s bandwidth isnt even used. Does anyone have any comments or results to share?

-im sub-10 on posts so lemme know if somethings wrong

[turd]

everthing ive read so far seems to say that the only dif is burst and reading long contiguous files. I think burst wise ATA 133 tops at about 100MB/s while ATA at about 60MB/s under actual conditions when you take in bus congestion and and error correction.

[hellrazrblade]

so the 80ish burst i got on the ata100 isn't accurate? know any better proggies for testing?

[nihili]

Everything I've read is that there is almost no discernible difference between ATA66 and ATA100, and that the difference bettween 100 and 133 is even less. At this point it seems to be a physical limitation on the drives themselves. So it may be worth it to spend the bucks on a 133 mobo in the hopes that the drives will eventually use the extra bandwidth. But I wouldn't pay much extra for a 133 drive over a 100.

nihili

[FireMogle]

I agree, raising the burst rate isnt going to increase perfromance much at all.

[Johnny Knoxville]

at the moment the only drives that support ATA 133 are the maxtors and IBM 120GXP and Western digital special edition (both ATA 100) are faster, so at the moment its not faster but in the future it will be.

[OldGuy]

I have one of the Maxtor 133 drives and it gives the same benchmarks on my onboard ATA100 controller as it does on an add-on ATA133 controller.

[hellrazrblade]

good, i thought i had something wrong (which i'm sure i do, just not about this ;) )

[nikhsub1]

Originally posted by Yodums2
at the moment the only drives that support ATA 133 are the maxtors and IBM 120GXP and Western digital special edition (both ATA 100) are faster, so at the moment its not faster but in the future it will be.
In the future there will be SERIAL ATA, making all this obsolete. I don't think any other drive manufacturer will make ATA133 drives.

[supergenius74]

current drives max read and write speeds (including scsi) barely use ATA66 let alone 100 and 133. Serial ATA is coming and waiting for a mobo that supports it will be worth it. but until HD manf. figure out how to make faster drives, Im afraid we are at a hardware "bottleneck" for the time being. Serial ATA will not do anything for current drives because current drives don't fully use the current ata interface.

[hellrazrblade]

so SERIAL ata is going to use a different connector or still the same? does anybody have a good link to some info on it?

[supergenius74]

http://www.dell.com/us/en/gen/topics/vectors_2000-sata.htm is a good link, serial ata is basically ready, im not sure what the hold up is but we should have had it by now, i'm ready to upgrade but im waiting for motherboards to support it before i buy and im getting very impatient. its completely compatable with everything we have today, drives, software, everything, it uses a differnt connector, a cable, much smaller like the size of your CD audio cable, and the first version (v 1.0) transfers at 150 mbps. you only plug one device per cable (floppy, CD, CDRW,HD) and there are no jumpers to set any more. There are supposed to be adapters out to allow old parts to work with new parts and this link http://www.theinquirer.net/30050210.htm shows a new drive with serial ata so it should be soon I hope.

[Johnny Knoxville]

You can always get SCSI instead of waiting for serial ATA

[Sprockett]

I'm building a new system right now using the new D740X 80Gb Maxtor drive, I'm running it off a KR7A MB which is supposed to fully support ATA 133. As soon as I get all my parts in next week I'll finish everything and run some tests. I did read somewhere on the Mac sites (I own both PC and Macs running OSX) that Maxtor had put in a "write verify" check in there drives and that after XX amount of reboots the software disabled it'self and the drive sped up when writing.

I'll have to dig on their site and see what the deal is, I also was checking into the difference between a standard ata 100 and an ata 133 but for the life of me all I can figure out is it's a marketing ploy :)

Cheers

-Paul-

[hellrazrblade]

except for the fact that scsi is expensive with minimal performance gains in real life

[ShiFtY2001]

You do know that win xp doesnt support ATA133 yet?
Win 2k only supports ATA33 unless you install the ATA66 registry key. Dont know about 98/ME

That could be a reason for your benchies...

Also read speed is only about 30Mb/s which is the main limiting factor, ATA66 is unlikely to get completely "filled" even.

[FTC]

ShifFY2001,


Don't know 'bout XP, but my win2K+SP2+via IDE drivers partition is using ATA66 all the way. Id did not install any special registry key for this, and am 100% certain about ATA66 because :

- I can query the disk directly through the S.M.A.R.T tool
- When I had the wrong (40 wire) cable connected, it reported UDMA2 (ATA33), and benchies confirm the difference.

Note that it is NOT at ATA100 because my mobo does only support ATA66.

... So at least for win2k, if SP2 is installed and VIA IDE drivers, ATA66 ... is supported.

Regards
FTC

[Hammerhead]

Can you please explain how to use this S.M.A.R.T. tool?

[FTC]

Sure,

Download the program DTEMP (http://private.peterlink.ru/tochinov/DTemp/DTemp.zip) , install it, and you can then run it and then see the currently running temperature in the taskbar, and by right clicking its icon, access drives info and SMART attributes. .... among those you find the currently running DMA mode.

Regards
FTC

[jstutman]

i feel that the difference in ata100 and 133 is not physically noticable ( time wise) but benchmark wise. yes there is a BIG change in performace .

[nihili]

Originally posted by jstutman
i feel that the difference in ata100 and 133 is not physically noticable ( time wise) but benchmark wise. yes there is a BIG change in performace .

Specifically what benchmark and under what sort of setup are you talking about? Every benchmark I've seen reports at most marginal gains. There are greater differences among ATA 100 drives than there are between ATA100 and ATA133. If you have data that shows otherwise, I'd be very interested in seeing it.

nihili

[TRANCER24]

Originally posted by nihili


Specifically what benchmark and under what sort of setup are you talking about? Every benchmark I've seen reports at most marginal gains. There are greater differences among ATA 100 drives than there are between ATA100 and ATA133. If you have data that shows otherwise, I'd be very interested in seeing it.

nihili

Even if there is a slight marginal gain it is worth it ATA 133 all the way besides the maxtor drives are better anyways there the only ones that make there drives with fluid dynamic bearing motors.

[nihili]

Originally posted by TRANCER24


Even if there is a slight marginal gain it is worth it ATA 133 all the way besides the maxtor drives are better anyways there the only ones that make there drives with fluid dynamic bearing motors.

:rolleyes:

I think we've had this discussion before. Look for the drive that has the best actual performance and reliability in your price range. If that happens to be an ATA133, great. If it happens to be ATA33 great. The ATA designation just gives a theoretical maximum. The real performance statistics are out there. Find them and make your choice based on them.

nihili

[Sonny]

Originally posted by hellrazrblade
i just built my new rig (in sig) and i put a 40gig Maxtor D740X ATA133 hard drive in it (primary master). The drive is partitioned into a 10gig C and a 27gig F drive. After running HD Tach 2.60 i got the following results:
11.9ms random access time
84.7mb/s burst speed
max read speed of 43811kps
avg. read speed of 35342kps
cpu utilization of 20.5%
i put my 40gig Western Digital Caviar ATA100 hd in as the primary slave and got the following results:
14.7ms random access time
81.9mb/s burst speed
max read speed of 33308kps
avg. read speed of 24290kps
cpu utilization of 17.5%

the access and read times are notably worse on the WD, but the burst speed is almost exactly the same. It seems to me that the "extra" 33mb/s bandwidth isnt even used. Does anyone have any comments or results to share?

-im sub-10 on posts so lemme know if somethings wrong Could it be that when you ran the tests the HDD were heavily fragmented? Which 4in1s are you using? Did you install them properly for your OS? Here are my scores running a Seagate Barracuda III 20G;