New WD Raptor Drive on the horizon...
- Thread starter MasterBaiter
- Start date
karlotta said:AND with my usage of raid, and the raptor, i found no real noticable difference in loadtimes in games.. just a snappy feel. Both are not as good as a gig of ram or two. 600$ for snappy is cool tho......
I timed the level load times in Quake4, and the time it took (20 s) is sufficient to read over a Gigabyte from the disk. Decompression and setup is more likely what takes the time. People who are into large image editing (I typically handle 200+MB images) or video would make a hell of a lot more sense as a target audience than gamers. The fact that they market a "Gamer Edition" is rather an indication that they believe that there is a group who are easily parted from their money, than that their product has any substantial benefit for gameplay. Not that WD is alone in this. Just look around, see what kind of products are specifically marketed towards gamers, and they all imply that their audience has more dollars than sense, and are easily swayed by package designs featuring large tits.
The raptors have fast access times and good transfer rates. Servers on Gigabit LAN are the true market for a product such as this, but there are a lot of well heeled consumers around for fleecing. Good thing my credit card is suffering from seasonal depression, or I would have been sorely tempted.
Cartoon Corpse
Veteran
what does the 10K rpm buy you? improved swapping? why not just buy more ram?
Sounds like quite a stretch to me. In the long run I think things would even out, as statistically you'd get the data you DON'T need first as often as the other way around.pcchen said:
The 840/EE gives me about 50% decrease in build times for my target application.
Guden Oden said:
Why? You are reading data from two identical disks with identical data. And when you don't need that data (say, one disk get it earlier than another disk), you can always cancel the I/O.
Of course, you get nothing from writing data. However, writing is generally not latency sensitive.
ANova said:
I spent quite a bit of time benchmarking different machines. In our compile test bench, the 840EE was a considerable winner, over the 840 non/EE (by about 15-25%), the FX/55 (by about 50%), and over a 640 (by about 30%).
Think what you want. When you're talking about shaving an hour of an engineers time a day (or even a week), that $470 is worth it. It pays for itself in a few weeks/months easily.
Chalnoth said:
No, you don't need to do that. Hard disks already have CRC and ECC for data integrity. I think there's no RAID 1 system which really compares the data from two disks to make sure the data is correct. Well, maybe there are really such system, but they surely are not the mainstream. And you can always have more data integrity check in software if you really want. For example, you can make your OS to do a SHA-1 or MD5 for each sector.
The main point of RAID 1 is, well, on-line backup. You have data on both (or more) hard disks, so if anyone of them crashed, you still have your data.
For gaming (and mostly single-threaded workloads) I would recommend the Hitachi Deskstar 7k500. It delivers similar performance to the 74 GB Raptor in gaming and single threaded applications (but has slower access times), offers 500 GB and costs about 300-330€. Tip: www.storagereview.com.
I am sure the "Gamer" edition and the other will just be firmware differences, but I think that the difference could be quite substantial from their previous offereings. The "gamer" edition would actually be quite good for desktop use as well if it follows their previous trend. I think they will simply give up the much better multiple I/O performance it has in comparison to drives such as the 7k500 and try to acheive better performance for the regular desktop user. I certainly did notice a decent difference in my HDD speeds when I got a raptor, and I had 2x10k RPM scsi drives previously, but they were not in raid, just a fujitsu and a seagate I think.
We don't do either.ANova said:
As software developers, we compile our codebase over and over.
The 840/EE and associated intel platform smoked the pants off of its nearest competitors.
Not that I own one at home, because you're right--it makes very little difference in home applications, and games it makes even less (where performance is mostly dominated by graphics cards).
We compared with a 4000+suryad said:
I think the big tweak for us was the multiple execution units paired with the stellar IO subsystem of the intel chipsets
On a different, single threaded, tool suite, the AMD product performed quite admirably and beat the 840/EE by a small margin. Since our new tool suite was multi-threaded and IO bound, we went with 840/EEs and it truly outperformed the 4000+ by a 2:1 margin. Even the 840 non-ee outperformed the 4000+.
Several reasons.Sxotty said:
1: Dell doesn't make any.
2: The single core 4000+ was well behind the single core 640 in performance, and extrapolation would say that the 4800+ or Opteron 175 would probably continue that relation with respect to the dual core P4s. 4 execution units do a better job keeping the I/O pipe stuffed full and dealing with the bursty nature of compilation.
I originally pushed to test Opterons, but after testing the 4000+, I'm pretty confident that the Opteron systems would perform worse, and cost more than the equivalent Dell machines with 840/EEs we use for the same level of service we get from Dell.
