raid 0 game loading times

[choirbass]

just thought i would post this... the posted benefits of raid 0 seem to fall in line with whats 'generally' experienced in game loading times, for the reasons listed in the article (contrasting whats stated by most people who think otherwise about raid 0)...

if anyone cares to verify this, using two or more different cpus, but identical setups otherwise, both in raid 0, and single drive configurations, theyre welcome to...

http://faqs.ign.com/articles/606/606669p1.html


EDIT: hm... apparently this is very true... theres a much easier way to test though than switching out cpus... ...the easiest way is if you have dual core cpu, and a game that supports multithreading, such as quake 4 (it tends to use upwards of 70% of both cores when theyre in use, sometimes as high as 100%, though less often), so, its a good game to test... havent checked other games though... but, setting your cpu affinity from both cores, to only 1 core, will drop the total processing capability up to half... ...i also have 4*36GB raptors (3 of which are in raid 0, for hosting my OS and various applications, and a 4th non raided which is being used for the pagefile, and to keep a few files as an extra backup, in addition to other backups)... i copied quake 4 over to the 4th raptor though

i dont have a watch to time the loading with, so i just had to go by how long it seemed to take to load each level... where the pauses in loading happened, guaging the difference between just one core, and both cores in use, and the difference between 3 raptors in raid 0, and the single raptor... ...apparently the only real difference in loading times, was when i switched the game over to just 1 core, instead of using both cores... ...the 3 raptors in raid 0 had very little impact (compared to the single raptor), or not that i noticed really anyhow...

but, thats an inexpensive way to test; if you already have a dual core cpu on hand, and a game that can take advantage of the second core (which is why i used quake 4, though oblivion is also multithreaded too, as are quite a few other games)... ...and, goes to show, that i guess if you want to load games faster, instead of investing in raid 0, just get a faster cpu/memory configuration (because memory speed also affects overall cpu processing speed)... ...overclocking your existing cpu and memory should help here also... ...thats not to say either, that you should just forget about hdd speed entirely, hdds are slow enough as it is, compared to every other major speed based component lol... so, speed and capacity is definetly the way to go for general use when it comes to hdds...

as far as the raptors i have, yes, i did purchase them a few years ago to benefit game loading times, among other things, such as just helping my computer to be more responsive, lol... but, i never saw much change for games, even when i had all 4 in raid 0...

now, for older games, i suppose this might not quite be the case, because the game levels might not take that much processing and decompressing to load, and raid 0 may benefit better there... but for current games, a faster processor/memory configuration definetly seems to help speed things up, such as this...

anyhow, yeah... this definety is the case so far... have to test some other games though too sometime, just to see if its not just isolated to quake 4 (i wouldnt think it would be, but just to see)

 

[Uber_sven]

i agree...while it does speed up data read/write times to/from the HDD, in my experience with RAID 0, I have found no improvement in game load times. I have seen, however games load faster once i've upgraded my CPU, ram and MB chipset.

However, while upgrading CPU & memory for faster game loading times, this is a relatively expensive option (especially given the current prices for ram). For overall system performance, I feel that in terms of "bang for buck", system performance can be improved very cheaply by adding extra HDDs in RAID 0. while games will not load any faster, i have found that overall system responsiveness improves, as do load times (marginally), and program installations.

lets face it, as the recent THWG article on HDDs pointed out, while capacity has gone up at an incredible rate (and costs gone down), HDD performance has not increased as nearly the same rate. I suppose RAID is just a way around this.

 

[SuperFly03]

This keeps getting asked time after time, and time after time blind fools who use RAID 0 keep touting its benefits without realizing it really doesn't help in load times.

I remember the MaxPC issue, it was about 2 years ago and it was a side column article. If you search the net you will find numerous instances of RAID 0 being debunked. Glad to hear more than 1 person realizes it is useless in gaming.

 

[choirbass]

yeah, the most inexpensive way to get a reduction in game loading times, is to just overclock what you already have... seeing as how most people on THG do that already, lol

and yep, there are quite a few articles out that debunk any benefit raid 0 offers for gaming... ...raid 0 is just gonna offer strictly raw data read/write performance boosts for audio, video, other media editing, and the like

 

[Uber_sven]

Mind you, the one benefit i have found for RAID 0 in gaming is to run the page file off a stand-alone array of RAID 0 disks. I started playing BF2 with 1Gb of ram and a decent video card (7900GT) and found that while graphics performance was excellent, memory lag was a major issue. I baulked when i saw the price of another 1Gb of ram, and realised that RAID would be a cheaper option. Once i have set up a RAID 0 array of 2 X 7200.9s, with only the page file (2Gb) running off it, lag in BF2 became much less of an issue.

I've seen people on the forums state that running the page file off a RAID 0 disk is unsatisfactory because the latency of the drives increase (worsens) with RAID, but in my experience, (at least with BF2) this has not been the case.

 

[killmess]

...i also have 4*36GB raptors (3 of which are in raid 0,

i dont have a watch to time the loading with, so i just had to go by how long it seemed to take to load each level... where the pauses in loading happened, guaging the difference between just one core, and both cores in use, and the difference between 3 raptors in raid 0, and the single raptor... ...apparently the only real difference in loading times, was when i switched the game over to just 1 core, instead of using both cores... ...the 3 raptors in raid 0 had very little impact (compared to the single raptor), or not that i noticed really anyhow...

as far as the raptors i have, yes, i did purchase them a few years ago to benefit game loading times, among other things, such as just helping my computer to be more responsive, lol... but, i never saw much change for games, even when i had all 4 in raid 0...

now, for older games, i suppose this might not quite be the case, because the game levels might not take that much processing and decompressing to load, and raid 0 may benefit better there... but for current games, a faster processor/memory configuration definetly seems to help speed things up, such as this...

1.RAID 0 and 1 can only be done with "2" disks, not three or more, so you don't see difference because maybe you didn't do a RAID 0.
2. If you have 3 or more disks, you can do a raid 5, it's not that speedy, but you can have security and more speed[/b] at the same time.
3.You should HALF the loading time speed with raid 0. Example: Oblivion in the Data directory have almost 4GB of data in only 6 files!. 8O
Compressed or uncompressed it's still 4 GB the game should load. :roll:

 

[arima]

I've seen people on the forums state that running the page file off a RAID 0 disk is unsatisfactory because the latency of the drives increase (worsens) with RAID, but in my experience, (at least with BF2) this has not been the case.

Those are precisely my concerns. I've been a little hesitant to raid 0 my disks because of the increased latencies that brings. Can someone explain how, despite the latencies, how it can bring better read/write performances? :roll:

 

[choirbass]

...i also have 4*36GB raptors (3 of which are in raid 0,

i dont have a watch to time the loading with, so i just had to go by how long it seemed to take to load each level... where the pauses in loading happened, guaging the difference between just one core, and both cores in use, and the difference between 3 raptors in raid 0, and the single raptor... ...apparently the only real difference in loading times, was when i switched the game over to just 1 core, instead of using both cores... ...the 3 raptors in raid 0 had very little impact (compared to the single raptor), or not that i noticed really anyhow...

as far as the raptors i have, yes, i did purchase them a few years ago to benefit game loading times, among other things, such as just helping my computer to be more responsive, lol... but, i never saw much change for games, even when i had all 4 in raid 0...

now, for older games, i suppose this might not quite be the case, because the game levels might not take that much processing and decompressing to load, and raid 0 may benefit better there... but for current games, a faster processor/memory configuration definetly seems to help speed things up, such as this...

1.RAID 0 and 1 can only be done with "2" disks, not three or more, so you don't see difference because maybe you didn't do a RAID 0.
2. If you have 3 or more disks, you can do a raid 5, it's not that speedy, but you can have security and more speed[/b] at the same time.
3.You should HALF the loading time speed with raid 0. Example: Oblivion in the Data directory have almost 4GB of data in only 6 files!. 8O
Compressed or uncompressed it's still 4 GB the game should load. :roll:

raid 0 can be done with any amount of drives, as long as its 2 or more... you could make a 100+ drive raid 0 array if you had that many available channels and drives available, even completely unmatching drives, if you really wanted to... ...i wasnt using raid 1... the seperate 4th raptor, was just a standalone drive, not associated with the raid array directly

the raid controller i have is just built into the nforce4 chipset, only supporting raid 0, 1, 0+1 and jbod... so raid 5 wasnt an option via hardware

as far as oblivion though... ill have to test it later on, i have it installed, but my brothers on my computer right now

as far as game loading times being cut in half just because of raid 0?, thats all speculation, and why this thread was posted to begin with, to show whether there is any credence to raid 0 actually giving the benefit people are wanting it to give... without bolstering outragous claims that cant be reproduced

when it comes to game loading times, if the data does not need to be uncompressed to play, the level will load much faster with just raid 0 as its just loading from the hdd, but when you start including the cpu and such into the loading too, the hdd then takes more of a back seat...

 

[Mobius]

Let me get this right: not content to simply halve the safety of your data, you are happy to reduce it to 33% of the reliability of a single drive?

And you do all this for zero, or trivial performance increases? Increases which would be spanked to death if you spent half that money on faster CPU and RAM?

Obviously you are not in the trade, or you would NEVER do that. If you knew the number of HDDs I've had to return on behalf of clients over the years, you'd be horrified.

Thanks for giving me a good giggle. I really needed one!

 

[Siba]

1.RAID 0 and 1 can only be done with "2" disks, not three or more, so you don't see difference because maybe you didn't do a RAID 0.
2. If you have 3 or more disks, you can do a raid 5, it's not that speedy, but you can have security and more speed[/b] at the same time.
3.You should HALF the loading time speed with raid 0. Example: Oblivion in the Data directory have almost 4GB of data in only 6 files!. 8O
Compressed or uncompressed it's still 4 GB the game should load. :roll:

Wait, so I need 4gb of ram to play oblivion? Maybe that's why it brings everyone's computers to a screeching halt! And I can't use more than 2 drives in raid-0? Well damn! I guess all those people with 4 raptors in raid-0 must be idiots!/sarcasm

Most of the data is compressed to save space, and thus requires time for your cpu to uncompress it and store it in ram. Close that mouth

 

[choirbass]

um... ok... ...i do have all my data backed up elsewhere in several places, so redundancy isnt the issue, or a concern at all to be honest currently, so im risking nothing by doing so... ...however, if someone is running their computer, with no backups whatsoever, for any of their important data, that would be foolish

my reasoning for originally purchasing multiple raptors a few years ago, was because i was unaware of the nominal benefits for the intended purposes... admittedly it was a mistake... but, if i can still put them to use and incur no penalty doing so, why wouldnt i?... windows boots up in close to no time flat, thats worthwhile in itself, lol

...did anyone even read the link at the top though?... thats all i was posting based on, and then drew conclusions from that as well... people dont need to get offended or upset or anything because of it... as if i was just making things up...

 

[SuperFly03]

...did anyone even read the link at the top though?... thats all i was posting based on, and then drew conclusions from that as well... people dont need to get offended or upset or anything because of it... as if i was just making things up...

Yes. I understand that, and that's all I am trying to speak to, but other items need to be clarified as well.

1.RAID 0 and 1 can only be done with "2" disks, not three or more, so you don't see difference because maybe you didn't do a RAID 0.
2. If you have 3 or more disks, you can do a raid 5, it's not that speedy, but you can have security and more speed[/b] at the same time.
3.You should HALF the loading time speed with raid 0. Example: Oblivion in the Data directory have almost 4GB of data in only 6 files!. Shocked
Compressed or uncompressed it's still 4 GB the game should load. Rolling Eyes

No more from you.

Here ya go

Another one for ya

Yet another link

Hardware Requirements for RAID

Simple striping (RAID 0) requires two or more drives; mirroring (RAID 1) requires two drives; and striping with parity requires at least three drives (two or more for data stripes and one for parity, whether it is dedicated to a single drive or distributed). Striping with double parity (RAID 6) requires at least four drives

Next:

Those are precisely my concerns. I've been a little hesitant to raid 0 my disks because of the increased latencies that brings. Can someone explain how, despite the latencies, how it can bring better read/write performances? Rolling Eyes

In RAID 0 the drives are striped based on a predetermined stripe size specified when you originally build the array in the RAID BIOS. They range from 16k-1024k (I think). Tests show that the best stripe size for normal usage is I think 16k or 32k, and if you deal with really large files, 1024k. Once the array is built, the OS sees the two disks as one and addresses them as a single drive. So when you load a game the system sends a request to the RAID controller and says "I want data1.cab" (for example). The RAID controller says "ok, you got it" it then begins reading the drives alternately. Meaning, the first 16k (or whatever stripe size you specify) is read from drive 0 and the next 16k is read from drive 1, then the next 16k comes from drive 0 and the next 16k comes from drive 1, so on and so forth.

So when you move large files there is a decided advantage to the increased throughput that comes from RAID 0, but seek times don't change because the disks themselves are still operating at the same speed, its just that you are alternating the loads to artificially increase throughput. I say artificially because there is no physical change in the throughput, just the way it is allocated changes, instead of 2 point to point connections, they are combined into 1 big one.

Most of the data is compressed to save space, and thus requires time for your cpu to uncompress it and store it in ram. Close that mouth

Bingo. The bottleneck is the CPU decompression speed, not the hard drive in game loading.

To non-believers: Here is a quick test that confirms it: Set your system to stock and load BF2, then OC your proc and load BF2, compare the times. Then RAID 0 (with stock CPU) your drives and compare to your system stock and you will realize the truth

 

[choirbass]

exactly

 

[SuperFly03]

word.

lol, ok it is definitely time to sleep now. Peace!

 

[choirbass]

ok lol

 

[Chainzsaw]

Let me give you guys an example...


My old computer was a P4- 3.06Ghz HT (very first hyper threaded one), and 80GB x2 in RAID 0, 512MB RAM. It seemed pretty snappy, loading games, etc...

I was working on a system one time that had an AXP 3000+, 512mb RAM, and an 80GB HD. So it should be fairly similar (the system set up with the exception of the RAID 0, they both had XP PRO).

As I was using this AXP3000+ I noticed it took visibly 2-3 seconds longer to open up the "Start->All programs" than my own system with RAID 0, AND it had LESS programs installed on it.

Anyways, i'm never going back to non-raid. I have been happily using RAID 0 since 2003, and never will turn back.

Currently i'm Running 2X 150GB Raptors in RAID 0, and it screams!
While playing games, i'm usually the first one to load, as an example in Warhammer 40k, I load in under 20 seconds, usually waiting for 1-2 minutes for other people to load

During the whole time i've been using my RAID 0 system not once has it ever failed on me. FYI I have a 500GB secondary drive that stores all of my critical DATA, so i'm not too concerned about RAID corruption (it has never happened so far to me).

FYI, it's good to have a battery back up system, currently I have a UPS ($250 investment to protect my $3600CDN computer) for my computer, because computer's are highly susceptible to power-fluctuations. This could be why my systems have always run solid.

Just thought I would post my little FYI.

 

[Siba]

Let me give you guys an example...


My old computer was a P4- 3.06Ghz HT (very first hyper threaded one), and 80GB x2 in RAID 0, 512MB RAM. It seemed pretty snappy, loading games, etc...

I was working on a system one time that had an AXP 3000+, 512mb RAM, and an 80GB HD. So it should be fairly similar (the system set up with the exception of the RAID 0, they both had XP PRO).

As I was using this AXP3000+ I noticed it took visibly 2-3 seconds longer to open up the "Start->All programs" than my own system with RAID 0, AND it had LESS programs installed on it.

Anyways, i'm never going back to non-raid. I have been happily using RAID 0 since 2003, and never will turn back.

Currently i'm Running 2X 150GB Raptors in RAID 0, and it screams!
While playing games, i'm usually the first one to load, as an example in Warhammer 40k, I load in under 20 seconds, usually waiting for 1-2 minutes for other people to load

During the whole time i've been using my RAID 0 system not once has it ever failed on me. FYI I have a 500GB secondary drive that stores all of my critical DATA, so i'm not too concerned about RAID corruption (it has never happened so far to me).

FYI, it's good to have a battery back up system, currently I have a UPS ($250 investment to protect my $3600CDN computer) for my computer, because computer's are highly susceptible to power-fluctuations. This could be why my systems have always run solid.

Just thought I would post my little FYI.

What you are seeing here is most likely the effect of a very much reduced seek time. Try a single raptor 150 and see if there's any difference between that and your raid 0. Tell me exactly what you think 150+ mB/s transfer rate means when the files for icons etc on your startmenu probably don't even exceed 1mB in size? If you're the first one to load in WH40k, then it's probably a combination of a faster cpu(which usually goes along with an upgrade to raptors, since people will purchase raptors at the same time they purchase a whole new computer) and faster seek times on the hard drives (which would make a big difference when loading tons of small files that aren't sequentially arranged on the hard drive in the order they are loaded).

 

[Chainzsaw]

What it means that...instead of spending 2-3 seconds out of every minute in a 3-6 hour day using computers waiting, i'm able to do thing's faster.

All this time adds up, even if it seems really small in the short run.

Like I said I'm not going to break my RAID 0 set up because I don't want to "just" use a single raptor.

Of course I have a fast CPU, it all helps in the long run. But consider this, a HD is 1000 times slower than the next fastest component in a computer. If I can give the HD an advantage in speed, it will definitely help out with the overall system performance by making the slowest component faster.

Of course the seek time is reduced, it's 10,000RPM vs 7,200. after all. The reduced seek time + read + write speed of the raptors of course affect the loading times. Reduced seek time is directly related to read and write speed, tell me i'm wrong. Not to mention each raptor drive has 16MB of cache, further increasing speed by allowing the information to be read as quickly as possible and stored in the RAM if the computer cannot read the information fast enough.

I guess we could go on forever debating but if you haven't/ don't use RAID 0, i suggest you give it a try before making those kinds of comments.
You might like RAID 0.

 

[SuperFly03]

What it means that...instead of spending 2-3 seconds out of every minute in a 3-6 hour day using computers waiting, i'm able to do thing's faster.

All this time adds up, even if it seems really small in the short run.

Ok.... but what did you pay in order to save 1 minute per hour?

Like I said I'm not going to break my RAID 0 set up because I don't want to "just" use a single raptor.

That is your choice.

Of course I have a fast CPU, it all helps in the long run. But consider this, a HD is 1000 times slower than the next fastest component in a computer. If I can give the HD an advantage in speed, it will definitely help out with the overall system performance by making the slowest component faster.

Overall the system may seem more zippy, but there is more of an effect on load times moving from a 7200RPM drive to a Raptor than there is going from single disks to RAID 0.

Of course the seek time is reduced, it's 10,000RPM vs 7,200. after all. The reduced seek time + read + write speed of the raptors of course affect the loading times. Reduced seek time is directly related to read and write speed, tell me i'm wrong.

Ok.

Seek time is one of the several delays associated with reading or writing data on a computer's disk drive, and somewhat similar for CD or DVD drives. The others are rotational delay and transfer time. In order to read or write data in a particular place on the disk, the read/write head of the disk needs to be physically moved to the correct place.This process is known as seeking, and the time it takes for the head to move to the right place is the seek time.

Seek time isn't related to read/write speed. The only thing is measures is the latency from moving the read/write head across the disk. Read/Write speed is how quickly can you extract data once the read/write head has gotten to the position it needs to on the disk. So a lower seek time helps in non sequential file loading where files are spread out all over the disk, but it doesn't inherently change the read/write speed of the disk.

Not to mention each raptor drive has 16MB of cache, further increasing speed by allowing the information to be read as quickly as possible and stored in the RAM if the computer cannot read the information fast enough.

Check the HDD Charts Out

Cache is important, but more important than that is overall construction. Check out the hard drive charts under read performance, why is it that 7 of the top 10 have 8MB of cache not 16MB? Hard drives are a combination of a lot of factors, but the overall impact of cache isn't as great as you might think. I am going to head you off, before you say but look at the top 2, the old Raptor is 11MB/s slower than the new Raptor. The new raptor was a complete redesign including a native SATA connection instead of a SATA bridge used on the previous two generations of Raptors. Also the areal density has doubled, again providing a solid increase in performance because the smaller distance between the magnetic bits allows more data to be read per unit time as compared to a lower areal density.

I guess we could go on forever debating but if you haven't/ don't use RAID 0, i suggest you give it a try before making those kinds of comments. You might like RAID 0.

I have used RAID 0, for almost 2 years on first gen raptors. I wasn't all that impressed. Other than a few synthetic scores I didn't really notice that much difference. Windows did load faster, but first you have to load the RAID BIOS and detect the array so your total load time actually increases. All the data points to CPU's bottlenecking load times in games (which I agree with based on my experience), and I don't do huge file transfers, nor do I page during games, I don't do video editing, so I really have no use for RAID 0.

 

[choirbass]

i dont deny that using raid 0 does smooth out system responsiveness... thats not what was in question... ...when youre using two different cpu architectures to do the same thing however, neither of which are even clocked remotely similar (regardless of efficiency), there is automatically going to be a difference, im not emphasizing that clock speed is the be all and end all, we know that... but, taking a 2GHz cpu rated at 3000+, and comparing it to an actual 3GHz cpu w/HT no less, is going to impact performance to a different degree, and motherboard achitectures matter too... im assuming they both had the same memory speed, since you didnt say they were any different, which means your p4 3.06 w/HT was also using ddr400 at best (not ddr2 533 or so), didnt specify whether they had 2*256MB modules to possibly be in dual channel... ...the point being that those 2 compared systems simply arent only a cpu change in difference, or just using raid 0, and a single drive... now, if you were able to install the 3000+ inplace of your 3GHz/HT... that would have been a more worthwhile comparison (but still off)... and even more worthwhile, again, would have simply been to overclock your pentium 4, for the tightest spec'd comparison... instead of a possibly extreme difference between 2 different cpu architectures and systems in general, both of which have their own advantages, doing some things faster than one another... and HT is going to simulate a second cpu core anyhow, which is going to releave the burdon on the cpu as a whole, inevitably resulting in increased productivity... 'eg. multiple threads being able to be processed simultaneously', instead of just one after the other after the other, in sequence... again, the systems being compared were simply not the same, they could have had similarities, as you pointed out, but by no means were they comparable for identical performance, to see how much raid 0 by itself really helped for what you were testing.

but, again, simply overclocking your p4 /w HT wouldve been a more worthwhile comparison to do (as opposed to using the 2GHz cpu as a comparison)... taking it from 3.06GHz, maybe up to 3.8 or around there (or higher), and, then compare the difference.

and, i see you do have an E6600 anyways now... so, you can test exactly what i was saying in my first post, you have 2 cpu cores, and theres a decent amount of multithreaded games out now... when running a multithreaded game, just switch the cpu affinity from both cores, to one core, see if it takes any longer for a game level to load, only having one core available to process the level, and then switch back to have 2 cores enabled (though the efficiency of how well the game was coded for multithreading has an impact too)... or, compare your cpu at stock speed of 2.4, and then run it at 2.8 like youre doing (or higher), see if it affects game loading time at all, that way you dont 'have' to have a multithreaded game on hand to test. i dont think youd want to break up your raid 0 raptor array, so, doing that may be a bit much, just for the sake of testing... so, cpu speed differences and testing with one core, and both core would be easier to do then... now, even if you dont want to do this, or mention if theres any difference, or even be honest about it if there is any difference... its okay either way, but, doing this will allow you to see firsthand, for your own benefit... ...it could very well be (TBH), that the reason youre entering the games as fast as you are, is because of undoubtably how fast your cpu is, compared to theirs, more than likely.

im assuming the game isnt old... ...if it was old however, more than likely there wouldnt need to be much decompressing by the cpu, the cpu would no longer be the bottleneck, and you can just go simply reading data from the hdd, with no processing necessary really, and, raid 0 will benefit there a great deal... thats more than likely not the case however, but just to give the benefit of the doubt, incase the game is rather small, or outdated, and then raid 0 would be in fact giving the benefits we all look to experience when using it, for every game no less.

 

[Synergy6]

Surely the point that 95%+ of respondents here aren't in the "trade", as you put it, invalidates your point. Most of the people here are home users with little/no mission critical data, and no customers to please. If the drives go tits up, the worst thing they're likely to lose are game save files, password cookies or holiday snaps.

I do quite a bit of video editing, and also regularly move 5GB+ files. I'm not using RAID0 atm, but I will in the next build. If halving the time it takes me to move my files is "trivial", I'd like to see something else which could make such a difference to large file transfers.

I don't have any "clients", so their experiences, and your knowledge of them, doesn't really bother me. I back up any really critical stuff to DVDR, so even if my array went down, I wouldn't lose anything of importance. If that sentiment makes you giggle, it's nice to create a little happiness in the world.

Synergy6

 

[SuperFly03]

I do quite a bit of video editing, and also regularly move 5GB+ files. I'm not using RAID0 atm, but I will in the next build. If halving the time it takes me to move my files is "trivial", I'd like to see something else which could make such a difference to large file transfers.

That is one circumstance which I would say RAID 0 really is beneficial. I agree with you. Moving 5GB+ files consistently would really benefit from RAID 0. I would like to make one suggestion, use as large a stripe size as possible when you build your array, probably 1024k.

Side note: RAID is one of the most oversold gimics that mobo manufacturers have sold. I don't deny that RAID has its place, its just a very specific place.

Also home users may not have mission critical data, i agree, no one wants to lose family portraits (esp if they are loved ones who have passed on) or their music collection (talk about hours upon hours of encoding). So it may not be critical in the sense that others depend on the data, but there are some things that you really don't want to lose.

 

[Chainzsaw]

"Ok.... but what did you pay in order to save 1 minute per hour? "

So what, sure I paid 3600CDN for my whole setup but that was my choice. It's like someone buying a 17" monitor, vs a 20" monitor, do you really need it? No, but does it help in the long run by being that much better? Yes.

I'm not trying to say your right, or that i'm right, i'm not trying to argue with you. Just saying I made a choice to not use slower components, because "I" can notice a difference in computer speeds. I'm very sensitive to those things because i've used computers since I was 7-8.

"Windows did load faster, but first you have to load the RAID BIOS and detect the array so your total load time actually increases. "

My system boots up normally like any other one with or without RAID. In fact by the time Windows desktop screen appears I can start using everything on my desktop, can you say the same thing? I like that and i'm used to it. If you don't notice a difference, well thats just you.

"i dont deny that using raid 0 does smooth out system responsiveness... thats not what was in question... ...when youre using two different cpu architectures to do the same thing however, neither of which are even clocked remotely similar (regardless of efficiency), there is automatically going to be a difference, im not emphasizing that clock speed is the be all and end all, we know that... but, taking a 2GHz cpu rated at 3000+, and comparing it to an actual 3GHz cpu w/HT no less, is going to impact performance to a different degree, and motherboard achitectures matter too... im assuming they both had the same memory speed, since you didnt say they were any different, which means your p4 3.06 w/HT was also using ddr400 at best (not ddr2 533 or so), didnt specify whether they had 2*256MB modules to possibly be in dual channel... ...the point being that those 2 compared systems simply arent only a cpu change in difference, or just using raid 0, and a single drive... now, if you were able to install the 3000+ inplace of your 3GHz/HT... that would have been a more worthwhile comparison (but still off)... and even more worthwhile, again, would have simply been to overclock your pentium 4, for the tightest spec'd comparison... instead of a possibly extreme difference between 2 different cpu architectures and systems in general, both of which have their own advantages, doing some things faster than one another... and HT is going to simulate a second cpu core anyhow, which is going to releave the burdon on the cpu as a whole, inevitably resulting in increased productivity... 'eg. multiple threads being able to be processed simultaneously', instead of just one after the other after the other, in sequence... again, the systems being compared were simply not the same, they could have had similarities, as you pointed out, but by no means were they comparable for identical performance, to see how much raid 0 by itself really helped for what you were testing. "

My system I had RDRAM PC 1066- 32bit (single 512MB stick, don't tell me I had to have two sticks, because my system used A SINGLE stick, it was 32bit, thats why I could use 1 stick, with 1 continuity module on the other slot) where as the AXP 3000+ had DDR 333Mhz. I know of course the RDRAM had much more bandwidth BUT, the Athlon system had a dedicated controller onboard, which made communicating with the DDR ram MUCH FASTER. My RDRAM, although it had a lot of B/W, it's latency WAS HUGE. I just wanted to point out that even though my system was Hyper Threaded...it was a proven fact that it was only GOOD with Encoding and media files....VS the Athlon's awesome gaming performance due to it's short latency with the RAM, and somewhat efficient architecture.

I wanted to point out that my P4 3.06 was the FIRST Hyper threaded one, and it worked...but it was more of a performance disadvantage then a good one (except for media files of course). The next update of the P4 (3.0GHZ/800mhz FSB) was superior to my original P4 3.06/533mhz FSB system. I also wanted to mention I tried overclocking it the first time I got my system 3 years ago (the P4) and it didn't overclock so well, because it was a NorthWood, and it was one of the newer HT ones (not as good of a process as the Prescott).

The game I play is 2 years old, but still needs a fairly decent system. Warhammer 40k is the game.

Yea...anyways

I RIP quite a few DVD's movies myself. It really does help me to write fast.
Not to mention I work with quite a few large movie files (600MB-3-4GB)

Actually I use a crappy little program called Super DVD Ripper (anyone know of a better DVD ripper?) and I actually have a problem of my DVD drive unable to keep up with the Encoding of the DVD. It will read the DVD for 10-20 seconds, then it will encode and it will pause again to read from the DVD again.


This is a FYI, I don't want to start a war with you guys, I'm just letting you know about MY experiences.

 

[misry]

These supposed "Uber Geeks" that love to pontificate about the dangers/uselessness of RAID0 all seem to base their points on the premise that we all only have one system.

 

[randomcow]

Actually I use a crappy little program called Super DVD Ripper (anyone know of a better DVD ripper?)

Maybe DVD Decrypter?