Originally posted by: ValsalvaYourHeartOut
Originally posted by:
GetSome681
You are just flat out wrong. Take for example OS X. OS X has a metric ton of files associated with the OS. On booting (aside from many other operations), access times are more important that transfer speeds considering the size of these files are no where near as large are raw dv files/etc that you mention. When you raid, you do not increase access times, and if anything you raise them.
http://www.tomshardware.com/storage/20040123/wd740-08.html Have a look for yourself, although the increase in access times in a raid configuration could be considered negligible, it is still present.
Sorry, YOU are the one who is flat out WRONG, and the way you reason the system burden at boot-time is erroneous.
"SR Bootup DriveMark 2002. This particular test features higher-than-normal queue depths for a single-user machine, reflecting optimizations that Microsoft has made to the Windows XP bootup process. This high load allows an array's two or more independent actuators to service queued requests and improves overall performance. . .this individual test reflects the transfer rate advantage that one achieves through RAID 0. Therefore, if the primary purpose of one's machine is to start Windows XP, RAID 0 offers overwhelming performance benefits" -- Storagereview.com
Furthermore, who the HELL cares about OS X.
Sure, if you're transfering large files, raid systems make sense, and hence their large presence in the market, but you are just wrong if you think that a power user can't benefit from extremely low access times.
For the billionth time, if you are just running the "typical desktop environment," upgrade your HD to a Raptor will only BARELY improve your
overall system performance because the tasks in the typical desktop are cpu/memory limited...NOT HD-bottlenecked...therefore, even if you dropped your access times down, it wouldn't matter that much because most of your tasks are not heavily dependent on HD activity to begin with.
The only time a faster HD-system makes a HUGE difference in overall system performance is when you're manipulating large files (like in multimedia editing)...in which case STR is far more more important than access-time...this is why RAID0 make a HUGE difference, even though the access time of RAID'd Maxtors isn't quite as good as a single Raptor.
Read it and weep:
It is true that for many aspects of computing, the hard disk's performance level is not much of an issue. If you are recalculating a massive spreadsheet, or doing complex rendering of 3D objects, the amount of sheer processing power in the system is of paramount concern; the hard disk will only come into play periodically . . . The applications where hard disk performance issues are most important are obviously those that do a lot of reading and writing to the hard disk. . . Multimedia editing applications, especially those dealing with large audio and video files, are probably the ones most affected by the speed of the storage subsystem. Also up there are applications that process files, including compilers and many disk utilities. Initially starting your PC is also a very I/O-intensive application, as the operating system loads literally hundreds of files. Improving hard disk performance can shave time off the boot process in a very palpable way. The need for improved performance is a major driving factor behind the rise in popularity of RAID -- Storage Review
You get 63.3 MB/s w/ the 74 raptor, and 84.9 MB/s with a raid0 seagate 7200 setup. Sure this is a big difference, but in comparison to most single 7200 setup (~40 MB/s), the raptor is still extremely fast, not to mention the raptor access time is 40% faster than the raid0 setup.
Who the hell is comparing a single Raptor to a single Maxtor? NS, the Raptor would be faster. The more valid comparison would be 4 Maxtor's in RAID0...That would have a theoretical STR of ~160MB/s (although the relationship is not completely linear)...160MB/s vs. 63.3MB/s???? Yeah, you do the math. Imagine manipulating that 4GB DVD rip with a single Raptor vs. RAID0 x 4 Maxtors. I laugh. bwhaahha.
OH WAIT...
I didn't mention that the above is solely for read speeds...let's take a look at write speeds, shall we?
http://www.tomshardware.com/storage/20040123/wd740-09.html
A SINGLE 74 RAPTOR BEATS A 7200 RAID0 SETUP IN WRITE SPEED, 63 MB/s to 53 MB/s.
Bottom line is this: The raptor boasts 40% faster access times, 26% slower read rates than a 7200 raid0, 58% faster read rates than a single 7200, and 19% faster write rates than the raid0.
Access times = I have already addressed this. Only makes a marginal difference in overall system performance for typical desktop, makes virtually NO difference for HD-intensive applications (i.e. multimedia/large file editing).
Read-rate compared to a single 7200 = irrelevant. NS. I'm not making this comparison, you are.
Write-rate = you may have a weak point here...I'll have to examine the benchmarks myself. I'm not sure if I trust Tom's results, because according to that review, the write-speed actually SLOWS DOWN from Raptor x 1 to RAPTOR RAID0 x2...this goes against every other benchmark results we've seen prior to this. I'm worried that there was something wrong with his setup/config...or he did something different on his RAID0 setup...He also doesn't list his RAID0 array strip-size which can make a HUGE difference on performance.
Bottom line is this: the read-performance on the RAID0 x2 array is clearly faster than that of a single Raptor and would be MUCH MUCH faster if using 4 drives. This is the bottom line, my friend.
You saying that the 74 raptor isn't a power user's drive = you lose.
I never said that. Quote me exactly where I wrote that; otherwise, learn not to make up things to add to the fairy-tail world that you live in.
Edit: oh yeah, did i mention support for command queuing once sata controllers support it. can't get that w/ normal 7200 drive.
Oh, I'm sorry...the last time I checked, the Raptor does NOT support tagged command queuing...Oh, yeah...and thanks for the irrelevant point. appreciate it.
Valsalva