WD 36GB Raptor $79.99 at Outpost.com / Frys

[mparr1708]

Linkage

Seems warm to me. I've seen this as low as $69 but not in a while.

 

[tr1kstanc3]

thanks for the link, pretty good deal

 

[Ronnie]

how fast are frys rebates.

 

[dafatha00]

Originally posted by: Ronnie
how fast are frys rebates.

It took me 5 months to get mine. By the 4th month, I had forgotten all about it and when it came, it was a complete surprise.

 

[bluslice]

these babies usually cost around $100 on newegg.com so it is definitely a good deal especially if you want to buy two of these and raid em together. thanks for the deal!

 

[Bad Dude]

Fry's rebates are just as bad as Tiger Direct.

 

[mparr1708]

Originally posted by: bluslice
these babies usually cost around $100 on newegg.com so it is definitely a good deal especially if you want to buy two of these and raid em together. thanks for the deal!

Newegg 110.50
zzf 108.00
Monarch 108.00
mwave 115.00
tigerdirect 129.00
Pricewatch (bestbargains.com - Jigga who??) 101.00

So yeah pretty good price. I originally noticed this because I already have one. I'm considering getting a 2nd to raid. You can get the 74 GB at newegg.com for $155 after rebate though. RAID isn?t always as fast as people think (especially when it?s only two disks. RAID speed goes up exponentially with more disk in place) so I haven?t decided if I want to pick this up yet or not. Even more doubtful now that people are saying fry?s rebates are slow. I hate rebates to begin with much less slow ones.

 

[Trikat]

Originally posted by: Bad_Dude
Fry's rebates are just as bad as Tiger Direct.

But these are WD rebates.

 

[mparr1708]

Originally posted by: Trikat

Originally posted by: Bad_Dude
Fry's rebates are just as bad as Tiger Direct.

But these are WD rebates.

i do think you are correct now that I look again. It has one Fry's logo but looks like a WD rebate to me. That makes this more attractive! Anyone had bad luck with there rebates?

 

[intogamer]

where do u have to send it to?

 

[mparr1708]

Originally posted by: intogamer
where do u have to send it to?

WD Promo #843-086
P.O. Box 400120
El Paso, TX 88540-0120

 

[bluslice]

the price of getting a single 74gb raptor vs. 2 x 36.7gb raptors is going to be about the same, true. but if i raid the two 36.7gb ones and decide the add one on later, then my hard drive performance will increase exponentially. so i think getting two 36.7gb raptors might be the better route for me.

 

[bluslice]

i just bought the two 36gb raptors. it put a big dent in my wallet and in my heart. but since these are WD rebates, it puts me more at ease. there's even a number they provide for you if you don't receive your rebates in two months. so it seems reliable to me.

 

[carpenter]

Just bought one from NE, wish I could use this rebate on it.

 

[cyberfuzz]

Originally posted by: bluslice
the price of getting a single 74gb raptor vs. 2 x 36.7gb raptors is going to be about the same, true. but if i raid the two 36.7gb ones and decide the add one on later, then my hard drive performance will increase exponentially. so i think getting two 36.7gb raptors might be the better route for me.

acutaully i think anandtech showed that the 74 gig is faster than 2 36 raided? or was it 2 74 are tthe same as a single 74. eh might want to look it up if you cant choose.

 

[realdeals]

Nice find. OP

 

[chilifries]

Some helpful info.

 

[bluslice]

Originally posted by: chilifries
Some helpful info.

the 74gb one does outperform the 36gb one according to this article. my question is do the two 36gb raided together outperform one 74gb raptor? In the sysmark benchmarks it seems that the 36gb one only trails a few points at most. Although the 74gb does outscore the 36gb in terms of raw disk performance, two 36gb raided together could potentially be just as good or better than a 74gb. In the future if you wanted to raid three 36gb raptors then you would have even greater performance.

 

[Devistater]

Originally posted by: chilifries
Some helpful info.

http://www.storagereview.com/a...20040126WD740GD_2.html
This is a better link. Shows transfer rates and access times. 🙂
The 74gig raptor hits up to 72MB/sec sustained. The 36gig up to 57MB/sec (outer zone numbers, inner zone are slower since the rotational speed is a constant RPM). So there is a significant differance.
Assuming a raid could max out the rate of each drive (which would never happen), two 36gig drives could hit up to 104MB/sec and two 74gigs could hit 144MB.

Here's the numbers on a 8 drive array of some 8 meg cache drives (dont think its raptor, not sure):
http://www6.tomshardware.com/s...31114/raidcore-24.html
Hits max of 250meg/sec transfer rate with the RAIDcore pci-express raid controller.

The's another article where they grouped up 4 of those controllers and did 32 SATA drives together and got something over 1.1 GIGABYTE/second transfer rates. OMG lol. They had issues with windows limiting them to 2 TB per partition in NTFS ahhaahhahah.

So looks like you are right, the SATA raid scales pretty good. Probably cause I think each SATA drive has its own bus connection. So with 8 of them raided on a PCI-X card you are limited to the bus speed of PCI-X for 8 of them.

 

[bluslice]

1.1 gb/s ROFL

 

[user1234]

Originally posted by: Devistater

Originally posted by: chilifries
Some helpful info.

http://www.storagereview.com/a...20040126WD740GD_2.html
This is a better link. Shows transfer rates and access times. 🙂
The 74gig raptor hits up to 72MB/sec sustained. The 36gig up to 57MB/sec (outer zone numbers, inner zone are slower since the rotational speed is a constant RPM). So there is a significant differance.
Assuming a raid could max out the rate of each drive (which would never happen), two 36gig drives could hit up to 104MB/sec and two 74gigs could hit 144MB.

Here's the numbers on a 8 drive array of some 8 meg cache drives (dont think its raptor, not sure):
http://www6.tomshardware.com/s...31114/raidcore-24.html
Hits max of 250meg/sec transfer rate with the RAIDcore pci-express raid controller.

The's another article where they grouped up 4 of those controllers and did 32 SATA drives together and got something over 1.1 GIGABYTE/second transfer rates. OMG lol. They had issues with windows limiting them to 2 TB per partition in NTFS ahhaahhahah.

So looks like you are right, the SATA raid scales pretty good. Probably cause I think each SATA drive has its own bus connection. So with 8 of them raided on a PCI-X card you are limited to the bus speed of PCI-X for 8 of them.

PCI-X - do you mean pci-express ? that's called pci-e, not pci-x which is something else. Anyway when using pci-e, you'd be limited by the bandwidth of the pci-e channel (x1 is about 250 MB/sec), so I guess you would have to use pci x4 at least, to have enough bandwidth for a raid array of 8 raptors. Now where would you find a mobo with such a connector, or for that matter such a pci-e x4 card anayway ?

Of course the workaround is to connect the sata drives directly to the motherboard, which then gives each one the full 150MB/s sata bandwidth. But that just means all your mumbo-jumbo about pci-x cards is bs.

Also to the people here that talk about raid performance vaguely, it might help you to know that raid 0 does not improve the access time, in fact the access time of a raid-o array equals the acces time of the slowest drive in the array, because data have to sent/received to all the drives simultaneously. Raid-0 does increase the transfer rate, theoretically by a factor equal to the number of drives, but in reality it's a decreasing curve - that is, two drive almost double the transfer rate, but four drives barely triple the transfer rate.

btw, the main advantage of 10k drives is the seek time (access time), and the transfer rate improvement is actually smaller. So it depends on the workload/benchmark where you'll se more performance increase( e.g. sequential read vs random access pattern). Definetly for gaming, the transfer rate hardly matters as the level loading time involves cpu tasks such as decompression, so it's cpu bound. So raid doesn't help at all for level loading time. But the fast seek time of 10k drives does help in gaming a little.

So to summarize, as far as bang for the buck - if you buy two 7200 drives and do raid-0 with them, you'll get a faster transfer rate (2 x 50 MB/s = 100 MB/s) than a single 10K drive (72 MB/s for raptor 2), and it will be much cheaper (2 x $50 = $100) and you'll get a lot more storage (2 x 120GB = 240 GB). BUT - the access time will still be the same relatively slow access time of 7200 drives (because as I mentioned, raid-0 doesn't improve the access time) - so the single 10K drive would still be better for gaming (or any other workload which performs a lot of random access), because the only time when waiting on the disk in gaming is when waiting to load small files which depends on the access time (and as I also mentioned - level loading time, which loads huge files sequentially, is actially cpu bound in gaming - so a fast transfer rate does not help here).

Kapish ?

 

[user1234]

Just to prove the point above about gaming - if you look at level files in far cry fro example, you'll see no level is larger the 30MB (average size is 20MB). So it would take less then one second for any drive to read the raw data, but it actually takes a minute for far cry to "load" the level. That's because the raw data from the disk is processed by the cpu, which builds all the relevant in-memory represenation, and this is what takes most of the "level loading" time. So level loading is not helped at all by increasing the disk transfer rate (which is what raid-0 does).

 

[Devistater]

The tomshardware article I mentioned was on PCI-X which was on an intel motherboard, not the PCI express that most of the motherboards have now. As for the specific terminology for that interface, PCI-X is both the term that tomshardware uses and the term that the official manufacture's page says here: http://www.broadcom.com/produc....php?product_id=BC4852
Also according to the official standard site:http://www.pcisig.com/home
The official terminology is: "PCI Express?, PCI-X 2.0 and PCI 2.3"

Yes, RAIDing drives together does not in general make for better access time. Mostly nothing will improve access time except better hardware for example better hard drives.

Transfer rates do scale fairly well with SATA RAID setups according to the benchmarks. But as you noted, thats not the only thing involved. Thats one of the reasons why Raptors are so popular, because they are improved in pretty much every area, transfer rates, access times, etc. As for level loading being "not helped at all by increasing disk transfer rate" I disagree. It is helped to some degree, even if the level itself is 30 megs. I'm pretty sure there's a lot more than 30 megs worth of transfers that happens during a level load. Like sounds and textures, and other things are loaded as well as just the level. And I'm sure that level loading DOES depend greatly on the CPU as well. To what degree each one is important and which is more important, you'd have to run some tests to find out. All I know is that when I used to do a mirror raid setup on some new fast hard drives when my CPU was starting to be middle aged, I would always be the first to be loaded with a new level when the CS map changed. Now granted this was a while back when 45 gig hard drives were the thing. But it DID help improve my loading time. I had only changed the hard drive subsystem at that time when I spotted a good sale thanks to AT.

 

[user1234]

Originally posted by: Devistater
The tomshardware article I mentioned was on PCI-X which was on an intel motherboard, not the PCI express that most of the motherboards have now. As for the specific terminology for that interface, PCI-X is both the term that tomshardware uses and the term that the official manufacture's page says here: http://www.broadcom.com/produc....php?product_id=BC4852
Also according to the official standard site:http://www.pcisig.com/home
The official terminology is: "PCI Express?, PCI-X 2.0 and PCI 2.3"

Yes, RAIDing drives together does not in general make for better access time. Mostly nothing will improve access time except better hardware for example better hard drives.

Transfer rates do scale fairly well with SATA RAID setups according to the benchmarks. But as you noted, thats not the only thing involved. Thats one of the reasons why Raptors are so popular, because they are improved in pretty much every area, transfer rates, access times, etc. As for level loading being "not helped at all by increasing disk transfer rate" I disagree. It is helped to some degree, even if the level itself is 30 megs. I'm pretty sure there's a lot more than 30 megs worth of transfers that happens during a level load. Like sounds and textures, and other things are loaded as well as just the level. And I'm sure that level loading DOES depend greatly on the CPU as well. To what degree each one is important and which is more important, you'd have to run some tests to find out. All I know is that when I used to do a mirror raid setup on some new fast hard drives when my CPU was starting to be middle aged, I would always be the first to be loaded with a new level when the CS map changed. Now granted this was a while back when 45 gig hard drives were the thing. But it DID help improve my loading time. I had only changed the hard drive subsystem at that time when I spotted a good sale thanks to AT.

It's a common misconception that raid-0 improves level loading time - according anandtech article AND maximum pc - raid does not help noticeably. They conducted tests with a bunch of real games (far cry, doom 3, etc) and compared the performance of single drive to dual raid-0 (using the same drive and system), and actual level loading time did not change much (less then 2% difference). Search for it if you don't believe it. With all due respect to your "scientific" "objective" impressions, the real tests show no difference. Also regarding level data size, there are very very few games whose levels are more than 100MB, and that would take only 2 seconds for a single average drive. So level loading time mostly depends on the cpu.

 

[Devistater]

Originally posted by: user1234
It's a common misconception that raid-0 improves level loading time - according anandtech article AND maximum pc - raid does not help noticeably. They conducted tests with a bunch of real games (far cry, doom 3, etc) and compared the performance of single drive to dual raid-0 (using the same drive and system), and actual level loading time did not change much (less then 2% difference). Search for it if you don't believe it. With all due respect to your "scientific" "objective" impressions, the real tests show no difference. Also regarding level data size, there are very very few games whose levels are more than 100MB, and that would take only 2 seconds for a single average drive. So level loading time mostly depends on the cpu.

In my case it was RAID 1. Notice how I said "mirror raid"? That means RAID 1, not RAID 0. Also, as I said, there's more data than just the level in a game, there's textures and sounds and other stuff as well. Now it could very well be that my experiance was just from getting a new up to date hard drive and not because of any type of RAID. Which was half my point, that buying a faster hard drive CAN increase your level loading performance. The question really is by how much. How much is CPU and how much is hard drive? That is the real question. Thats what I want to see benchmarks on.

As for science, nowhere did I claim that my anecdotal experiance was scientific or objective. Scientific methods rely on repeatable experiments and observations of such. A lot of things people think are scientifically based on the scientific method, are in fact not.
For instance, weather prediction. Turns out we actually cannot really predict weather because we dont understand it. The system is far too complex. What generally goes on is that we look at what has happened in the past with similar weather and what the season is and what it has been before and the particular geographical location, and then says well since it happened like this in the past under these conditions, there's a good chance it will happen like this in the future. No one experiements with weather, there's no controls in the experiment. We just observe and say hey it happened like this it might happen again like this. Thats where you get the 75% chance of rain for. In 3/4's of the past weather conditions they looked at, rain happened on ones that were similar. So weather prediction is based more on educated guesses seeing what happened in the past under similar conditions, than it is based upon the scientific method.
Another example is a lot of the cold fusion claims. They sound really good, only it turns out that pretty much all of them cannot be repeated elsewhere by other scientists.
Now gravity, that can be repeated and observed by anyone. Drop a weight and see how long it takes to fall. Thus we have various scientific "laws" relating to gravity.