New WD VelociRaptor 300GB On it,s way
View Single Post
July 4, 2008, 07:59 PM
Join Date: Jul 2008
Sorry, I know this thread is kind of old, but I stumbled across it and wanted to comment on some statements that I find are fairly inaccurate.
"Didn't impress me all that much. Almost all tests the WD Caviar 640 and the Samsung F1 640 were very close in terms of performance. I don't think I'd be buying this drive."
Do you have any links to these tests? All the tests I've seen show the VR has substantial gains over both the WD6400AAKS and the F1. Those two drives you mentioned are superb 7200rpm drives, but their performance is not near the level of a Raptor. Access times don't come close.
"However, every rational person wants to see value in the money they spend, especially when they are paying a sizeable premium."
It's all in how you perceive value. To a regular user, the VR increase in performance may not look attractive, but for someone who maximum performance out of their PC, even what may be thought of as
slight increases are a good "value". Think about people who buy top end video cards for $800. They're probably only gaining an average of 10fps and only at very high resolutions, but that 10fps is gold. To another user, that 10fps may seem completely impractical for the cost. You can't say someone is
irrational just because they have a different concept of value. I completely agree with you on "bang for your buck", but I don't think the two are always be interchangable.
"*The access time is very high because the 320GB model is tuned for quiet operation over performance. The 640GB & above models have access times that are 6.5-7ms lower."
How is it tuned for quiet operation? Do you have benchmarks or specs for the higher model showing access times of 6.5ms? Very interesting if this is true. The transfer rates are certainly impressive especially if it's not in a raid configuration since many raid setups will give similar STRs.
"Sorry to say but a hard drive plays absolutely NO role in a overclock."
Correct, but I don't think he was refering to his "actual overclock." Overclockers, especially extreme enthusiasts, will do what they can do obtain the best possible overall system performance. If you buy a QX9770, a GTX 280, and 4GB of RAM. Why wouldn't they go for the fastest drive possible?
"Raptor drives have always been beneficial to those who use large files because thats where they are simply gonna excel at compared to the regular drives."
Not really. One of the main attractions to Raptors are their low access times. Yes, Raptors are beneficial to those that use large files, but the main benefit of a Raptor over a Raid 0 setup is the access times. This is where you will really see a huge difference over a regular drive or even a Raid 0 setup. Load times, CS3 applications, compiling videos, installing programs. Anything that deals with a large number of SMALL files will see great improvements in speed. Those wanting improved speeds only for LARGE files would go for a Raid 0 setup since the STR would likely be greater than a single Raptor. Raid 0 will not help with your access times, only transfer rates. For the mainstream user, yes, the Raptor makes no sense, but the mainstream user would probably still notice more performance from a Raptor than going Raid 0. Most mainstream users will be dealing with hundreds of small files as opposed to many large
"VR are overpriced for their performance and only gullible people will buy them to gain a whole extra 1 second load difference in anything you do."
This statement isn't entirely false because sure, there are going to be instances where the VR will only give you 1s improvement, but it is far to general. It all depends on what you're doing and the size/number of files you are dealing with. If it takes 10s to load a game on a 7200rpm drive and only 6s to load on a VR, that's a huge improvement. You can think of it as "who cares, it's only 4s, it's not worth it", but that's a 40% improvement. Some warrant this price premium for that kind of performance. Magnetize this example 100 or even 1000 times and you're seeing huge differences in performance. Think about people who compile thousands of files.
I think a lot of the time people equate performance with how fast you can load a program or run a game. Some people are using their PCs to compile massive ammounts of data, data that can sometimes take days to compute, no I'm not exaggerating. Think about medical reasearch or space age technology. Not all of this is always done at work on supercomputers.
I see the arguements of both sides, but lets be fair and acknowledge where Tomvill is coming from.
My System Specs
View Public Profile
Send a private message to new_guy
Find all posts by new_guy