Do you notice a difference between the Raptor and a normal drive?

[Rickster]

Im starting to think that Ive wasted quite a fair bit of money on a raptor. The raptor isn't noisy as everyone says but it is exceptionally hot. And I don't see much difference in windows startups. Ive played games however I don't notice much difference, maybe Im playing games that aren't memory hoggers (playing winning eleven 9 now).
Ive run the diagnostics, no errors.

Is it just me or have I really just wasted my money. Maybe I should time the windows startups with a stopwatch.

[nd4spdbh2]

by n e chance do you have the old hard drive still in the system.... go to google and download hd tach... numbers dont lie....

I notice a big difference between my Hitachi ATA-100 drive:

50MBps average sequential
11ms random seek
94MBps burst


and my seagate 2x80gb raid 0 setup:

88.9MBps average sequential
12MS random seek
140MBps burst...

Theres a big difference for me, in start up... 1.5-2 passes of the windows loading bar with the Raid0 setup and 5+ with my hitachi... also stuff just loads way faster. You should notice a big difference if you have one of the newer raptors cus they have higher density platers that make everything faster.

[harlam357]

Raptors are worth every penny IMO. :thup: Where they win out is access time. If you check things with HD Tach you'll see that your Raptor is much quicker than your older PATA drive or even your (my) SATA 3.0Gb/s 2x320GB Raid-0 array (for comparison). :D

http://img.photobucket.com/albums/v663/harlam357/HD-Tach.jpg

[boris_37]

RAID raptors and you will notice the difference. We get a lot of people on here talking about their raptors not being far supurior to a regular drive.

If you had a good previous drive and are only running a single raptor i garuntee that you will not notice any different between the two. However if you buy another and run RAID im sure you will notice the different... i have. The only problem i have is noise because i have an open front case and my rubber mounts are all used up so i get a fair bit of vibration. :) hope it helps

[Okagi]

I just bought a ADFD coming from a single Hitachi 7K80 and I didn't notice any difference. I know it's there, but I can't say that I "feel" it. I'm sure if I went about with a stopwatch I would really notice it in certain tasks. I plan on pairing it with another when I get more cash.

[dicecca112]

I did notice a huge difference in linux, not really in windows, the windows load times were slightly better. Mine was a tad bit loud but it was because it was dying

[custom90gt]

I noticed a huge difference in load times for my engineering software, and windows seems more responsive.

[Snugglebear]

It makes the Dell Precisions at the office pleasant to use. The previous Compaqs we were using, even with the newest 7200s available at the time, were pathetic. While not as fast as the 15k SCSI units at home, it is rather nice to be able to multitask properly, which, given our staffing levels, is a necessity.

[mortimer]

Im starting to think that Ive wasted quite a fair bit of money on a raptor. The raptor isn't noisy as everyone says but it is exceptionally hot. And I don't see much difference in windows startups. Ive played games however I don't notice much difference, maybe Im playing games that aren't memory hoggers (playing winning eleven 9 now).
Ive run the diagnostics, no errors.

Is it just me or have I really just wasted my money. Maybe I should time the windows startups with a stopwatch.

They are faster from a seek perspective.

I have two 80GB (ish) Caviars in a stripe. I just ran HDTach. Got 244 MB/s. About $120 on an nforce2 board. A *lot* less than one 150GB Raptor.

Compare price performance ratio with the 283MB/s above.

[Rickster]

Raptors are worth every penny IMO. :thup: Where they win out is access time. If you check things with HD Tach you'll see that your Raptor is much quicker than your older PATA drive or even your (my) SATA 3.0Gb/s 2x320GB Raid-0 array (for comparison). :D

http://img.photobucket.com/albums/v663/harlam357/HD-Tach.jpg


darn, i dont get the hdtach concept. it says higher is better. but you pointed out that the raptor is better than your 2X230GB, but the bandwidth shows its less than the 2X230GB

edit: ok i kinda get it.


less cpu utilisation is better
lower random access is better
higher average read is better (but then isn't the raptor suppose to beat the sh*t out of your 2X230GB cause its known for its fast read times)


the thing i dont get is what is:
burst speed
sequential read speed

[Snugglebear]

First and foremost, HDTach is a linear benchmark.

Burst speed is the rate at which the drive can transfer data out of its cache to the disk controller. It is a test of the drive's firmware, onboard controller, the available channel bandwidth (SATA/PATA/SCSI), and bus bottlenecks (e.g. PCI vs. PCI-X, vs. PCIe, vs. dedicated channel off the southbridge). These numbers will be incredibly high yet have almost no influence on real-world performance.

Sequential read speed is how fast the drive can transfer data off the physical disk itself. HDtach simply starts reading data from various sectors and transfers the data to memory. While still potentially bottlenecked by channel or bus issues, usually the limit is the rate at which the drive can move bits under the read heads. Since the operations are contiguous, no random accesses are required.

Random access time is how fast the drive can get the heads from one cylinder to another and perform the requested operations. This gives a better indication of real-world performance since many random I/Os are required every second (even NTFS is a B-tree and thus requires random I/O even before the data nodes are accessed).

The reason even old Raptors still do well is because of the lower random access times. While they have higher spindle speeds, the platters are far less dense than those of modern 7.2k units and the platters are also smaller than that of a 7.2k unit. As such there are fewer bits moving under the heads per unit time as compared to a modern 7.2k unit or a RAID0 stripeset of them. The higher spindle speed of the Raptor does make up some of the difference, but not all. However, the faster spindle speed also decreases access times since rotational latency is significantly reduced. Upgrading to a modern Raptor will bring faster linear speeds due to increased platter density. It will also have decreased access times due to improved onboard intelligence and a more robust head actuator system.

[Rickster]

so basically theres only 2 things which really make a harddisk perform well in real-world performance. sequential read speed and random access time.

why is a denser platter better than a not so dense platter?

thanks snugglebear

[Seinken]

Denser platter = less movement for the read head from what I understand.

less movement = faster response = generally faster experience.

[Snugglebear]

The main advantage of a denser platter is that for a given drive size it takes fewer platters to build it. Thus manufacturers can offer obscenely large drives with only two or three platters. For performance, the main advantage for a denser platter is that for a unit of area there are more bits packed in there. This translates to increased contiguous read/write speeds as more bits are moving under the head per unit time. If seeks are required there's an increased chance that the data required will be on a nearby cylinder as the cylinders are smaller and more densely packed with data. Do not be mistaken, the main reason manufacturers are in a density race is due to cost and marketability, that this also helps performance some is a tertiary consideration.

In real-world performance testing there are more factors than that. However, HDTach only presents a couple of them. Write performance is oft-overlooked since the numbers tend to be slower than read performance over the same access patterns. Random access times are also deceiving by themselves, so it helps to see benchmarks where there is a good deal of random I/O. If you go over to storagereview you'll see a variety of different benchmarks used for every drive, and most use random patterns to try and replicate different types of use (e.g. home user, vs. business user, vs. developer, vs. server). By looking at those it becomes clear how well all the different pieces come together, contiguous read & write speeds, access times, onboard controller intelligence/firmware optimizations, caching, and bus speeds.

[Recursion]

Raptors are amazing and 100% worth it.

The only people you will see say that they are bad are the ones who dont own then already/

[Rickster]

The main advantage of a denser platter is that for a given drive size it takes fewer platters to build it. Thus manufacturers can offer obscenely large drives with only two or three platters. For performance, the main advantage for a denser platter is that for a unit of area there are more bits packed in there. This translates to increased contiguous read/write speeds as more bits are moving under the head per unit time. If seeks are required there's an increased chance that the data required will be on a nearby cylinder as the cylinders are smaller and more densely packed with data. Do not be mistaken, the main reason manufacturers are in a density race is due to cost and marketability, that this also helps performance some is a tertiary consideration.

In real-world performance testing there are more factors than that. However, HDTach only presents a couple of them. Write performance is oft-overlooked since the numbers tend to be slower than read performance over the same access patterns. Random access times are also deceiving by themselves, so it helps to see benchmarks where there is a good deal of random I/O. If you go over to storagereview you'll see a variety of different benchmarks used for every drive, and most use random patterns to try and replicate different types of use (e.g. home user, vs. business user, vs. developer, vs. server). By looking at those it becomes clear how well all the different pieces come together, contiguous read & write speeds, access times, onboard controller intelligence/firmware optimizations, caching, and bus speeds.

good info :thup:

darn it i missed out on the newer raptor by a month :(... oh well

[JCLW]

As always, it depends on how you use your drive. As a few people have already pointed out Raptors really come into their own when performing seeks.

If you're just loading big chunks of data you probably won't notice much of a difference between a 10k Raptor and a newer generation 7.2k drive.

[Snugglebear]

Yeah, but if you thrash your computer the enterprise drives stand out. The 3rd gen raptors are tempting, but I'm still feeling the call of U320/SAS. If only I had an integrated SAS controller on this board...

[Rickster]

epppppps theres a 3rd gen of raptors?