ATA133 vs ATA100

[Ricedaddy]

I'm looking at getting a larger hard drive, and I would like to see if there is a significant difference between ATA100 and 133.

I currently have a 40gb Maxtor ATA133 on my raid controller.

Would there be a big difference in performance with ATA100?

Thanks.

 

[djnes]

I'm looking at getting a larger hard drive, and I would like to see if there is a significant difference between ATA100 and 133.

I currently have a 40gb Maxtor ATA133 on my raid controller.

Would there be a big difference in performance with ATA100?

Thanks.

ATA100 and ATA133 are basically the same thing. It's more of a marketing term that Maxtor created, than an actual standard.

 

[aug1516]

There will be no performance difference between the two. There is a reason that ATA133 has not been adopted by most of the hard drive manufacturers, the performance just isn't there. The fastest transfer rate I have seen from a single drive is almost 80 MB/sec on a top of the line 15k SCSI drive. That 80 MB/sec is still short of the 100MB/sec allowed with ATA100 and even further from the 133MB/sec allowed with ATA133, and that is with a 15k SCSI drive. Your Maxtor drive would be even lower than that and will not come close to saturating the 100MB/sec bandwidth allowed with ATA100. The ATA133 interface is just a marketing tool that some companies use to make their hard drives sound faster to the un-informed.

 

[Ice Czar]

^ good explaination

for more info see > http://www.hardforum.com/showthread.php?s=&threadid=699246
specifically > I: Internal Drive Performance > Effective Internal Transfer Rate (TxD)

but the whole series is highly recommended
you can find a link to it in the FAQ


under extremely limited circumstances, employing RAID you might see a small advantage in ATA133, but its unlikely the internal transfer rate of just 2 IDE HDD still not maxing out the buswidth, there would be a greater advantage with a dedicated RAID engine and wider bus (64bit vs 32bit) and more drives

not to mention an application that actually requires that type of disk access
(read realtime video editing)

both ATA100 and ATA133 are the burst rate for transfer and as mentioned, saturating the bus is alot easier said than done
the other metrics are equally if not more important (and are covered in that series)

try this and see
http://www.bustrace.com/products/ide_monitor.htm
http://www.bustrace.com/downloads/IDEUtil.exe

 

[Doctor X]

once again... more fud. I really wish people would not make generalizations and propagate misinformation without thinking what the effects would be. I do all my own experiments and prove and disprove what I read. Just like raid0 vs single drive discussions that have been going around, someone screws up a test and it gets propagated.

The real answer is it depends. Saying ata100 and ata133 doesnt make a difference is to over generalize. It depends on your hardware and how you are set up. I have a rocketraid404 and during experiements, I can lose 5-10MB/s sustained from setting my 4 drive array from ata133 to ata100. I use maxtors for the raid set. The same effect is noticed with single drives. One mistake that may be happening is, while technically you can mix ata100 and ata133 drives on the same cable and set the speeds up independently, as the utilization of the channel increases, ata133 will drop down to ata100 if timing collisions are detected. the controller and drive would do this automatically.

The moral of the story is ... if you have all ata133 hardware (including just one device), you WILL notice a difference in both feel and in your benchmarks.

one more thing... in the above post, there is talk of ata100 and ata133 are burst rates... that is true... what is not true the statement that you cannot saturate the bus. If you are running a raid the you can... especially if you have 3 or 4 drives in a raid0. the rocketraid404 supports 8 drives...


Doctor X

 

[Ice Czar]

That qualification was made

but its unlikely the internal transfer rate of just 2 IDE HDD still not maxing out the buswidth, there would be a greater advantage with a dedicated RAID engine and wider bus (64bit vs 32bit) and more drives

however I noticed an error, I meant to link > III: Effective Host Transfer rate, not Effective Internal Transfer Rate

though they are related as a determining factor

For better scalability, we are using a RAID Level0 configuration on a SiliconImage SiI 3112 onboard controller. At small data transfer blocks, the effective host transfer rates doubles as the data size doubles while the CMD overhead remains constant. The limiting factor is the total host transfer rate, which in this case is ~ 105 MB/s. Whether or not we are looking at a RAID configuration does not matter since the small transfers are falling within one stripe block anyway, that is, they are "pulled" from a single drive, even if those drives are alternating within the array. As soon as the data size reaches the point where the total data transfer match the effective internal drive performance, the transfers will reach a plateau phase. In summary, transfers are caught between two limitaing factors, one of which is the effective host transfer rate, the other one is the effective internal media transfer rate.

also review speed matching which heavily impacts multi-channel RAID arrays on a 32bit bus