Final queries before ordering

Associate
Joined
30 Jun 2009
Posts
1,027
Location
Norwich
I've gone backwards and forwards over ordering what and how to set it up. I've gone with the F3's mainly over price than anything although people have still said these run a bit faster than the caviar blacks.

I wanted to RAID but all I've heard is bad things for a gaming rig, so I may just have to connect the 2 drives and have one as a C: and the other as a D: drive.

So does everyone agree that this is the best way;

2x 500gb F3's in parallel - one for os and other for gaming?
 
If you tend to have a lot of background activity and use a large page file, then split the drives to prevent stuttering in game. Otherwise RAID0 is fine - it'll give a noticable boost to game loading times and you get the speed benefits in your OS too.

Ideally, of course, you get an SSD for your OS and games and put the F3's in RAID1 for your docs and media.
 
Last edited:
F3's will be fine in raid, my 500gb versions reach, up to 240Mb/s.
Ofc ideally a SSD would be nice for an OS.

Although for the price you can't go wrong.
 
Otherwise RAID0 is fine - it'll give a noticable boost to game loading times and you get the speed benefits in your OS too.

If you call 10 - 20% "noticeable" then yes.

I don't.

e.g. 20 secs down to 16 - 18 secs.

To the OP - the biggest benefit for games will be going to 10k (Raptor) or SSD. Yes, have your PF on the first partition of the second drive.
 
If you call 10 - 20% "noticeable" then yes.

I don't.

e.g. 20 secs down to 16 - 18 secs.
Not willing to pay for an SSD as it's just not plausible to do that with what's available.

so what is the issue because now I'm hearing different things. I thought that if you had two drives in RAID then performance should be at maximum twice or fast, access times increased and load time for games take longer.

All I'm after is improved performance over a single drive so I don't mind what the performance boost is, as long as it's an improvement :)

What if I added a third drive into the equation?
 
Never had any issues with onboard RAID0 and I've always seen an improvement over a single drive.

I've had drives fail taking the whole array out, but that's the risk of running a striped array with no redundancy and something I've always been happy to trade off against the better performance.
 
so what is the issue because now I'm hearing different things. I thought that if you had two drives in RAID then performance should be at maximum twice or fast, access times increased and load time for games take longer.

All I'm after is improved performance over a single drive so I don't mind what the performance boost is, as long as it's an improvement :)

What if I added a third drive into the equation?
two drives in RAID0 will definitely feel faster than a single drive. A third drive will be faster again, but less noticable ... bear in mind that the more disks you have in RAID0, the higher the chance that you'll lose the array.

Maximum read/write speeds almost double, access times remain pretty much the same. Game load times will go down becasue it can stream the level data faster.
 
Game load times will go down becasue it can stream the level data faster.

Sorry if I'm getting you to repeat yourself but could you expand on this?

My system like many others on ocuk is just for gaming, I have a year old computer running off an E8500 and has sufficient space to backup/store work on, so I'm not worried about backing up/losing info on this system.

If I used 3 drives in RAID I know it'll be an improvement than 2 but obviously not to the same degree in performance gain, but how can load times be improved for the OS but not for gaming?
 
Why don't you look at my quantitative evidence and decide for yourself if it's worth it? (I posted the link in your other thread).

Game load times will decrease (i.e. be faster) by around 10 - 20%.

If this is worth if for you then go for it.

For me, it wasn't worth the hassle of losing the array, which did happen once or twice.

I have yet to see anyone on these boards show benchmarks of game loading times with/without RAID, which is why I posted my findings. Everyone seems to throw benchmarks around and "yeh m8, your system will feel a lot faster!" without any numbers to back up their subjective statements.

EDIT - Just ordered myself a Crucial 64GB for OS + common games. ;)
 
Last edited:
I have yet to see anyone on these boards show benchmarks of game loading times with/without RAID, which is why I posted my findings. Everyone seems to throw benchmarks around and "yeh m8, your system will feel a lot faster!" without any numbers to back up their objective statements.
I guess it's mainly down to the time it takes to install an os on a single drive, update it, hd tune it and use a game to also benchmark or something.

... then set up a raid and do it all again

... then use a different dive or more and do it again

I haven't RAID'ed myself and want to give this ago with these inexpensive discs before jumping ahead to a SSD, hopefully in the meantime they'll also lower in price.

With what I quoted last time I just read it wrong, getting confused between increased data speed and decreased load time which is essentially the same thing.

I appreciate the effort you went through Jim to do what you did, I'll definety go over that thread again but with different drives and boards available the whole speed thing is relative so I think most people would still give it a go and experience for themselves

Cheers guys

p.s. might try 3 drives lol
 
Yeh, I totally appreciate it's a pain to benchmark RAID vs. non-RAID, but I object to people banding about comments without evidence.

All the evidence I can find (including an Anandtech article and my own) points to RAID 0 being, at best, marginally faster for game loading times.

To be fair, from what I can see in game loading benchmarks with SSDs, they are only marginally faster again, suggesting the bottleneck is elsewhere.

In addition, your point about cost is also valid. If one already has a motherboard with RAID and the time to mess around, why not have a play? Mechanical drives are practically pocket money (relatively speaking).

Will be interesting to compare when I get my Crucial, hopefully tomorrow. I'll try to run the same benchmarks, although I don't have all of the games anymore.

PS - I would expect 3 drives to be 15 - 25% faster than one. i.e. diminishing returns. e.g. 20 second load on one drive = 16-18 second load on two drives = 15-17 second load on three drives.
 
Considering they're now only focusing on Hard Drives being the main bottleneck and from what you're saying it's hard to depict where this speed problem is located.

the only place I'd expect it to be is between the s/nb and the internal bus but this would practically be the same for the runnings of the os.

the only thing I can think of is that it's the os slowing things down and securom based devices and possibly even an optical drive (in the sense the os checks for a disc).
 
In some games, ARMA2 comes to mind here, Additional sequential transfer performance translates almost linearly into faster loading times. Going from a single Sammy F1 to my 80GB Intel halved load times from ~90 seconds to ~45 seconds. From watching Win7's resource monitor i could see most of the load time was dealing with large game files, so you should see similar gains from a RAID0 setup (three modern 7200rpm drives should come close to the 250MB/s sequential reads of the X25-M).

It's not always a cut and dried 10-20%, depends entirely on how the game is designed - in the case of ARMA2 RAID0 should give you a 90% improvement to load times.

disclaimer: I don't have the drives to test it out for myself any more, this theory is based from my observation that the ARMA2 load time is dependent on loading several GB worth of large game files, which would play to the strengths of RAID0.
 
Well, my Crucial M225 arrived, so I'll be benchmarking with the games I have available and comparing to 10k Raptor + 10k Raptor RAID 0 tonight.

Think I've already posted those combinations on Crystaldisk comparisons but with a M225 256gb in the Crucial thread. Benchmarks tell one story, real world use another - but people like to see the (bigger) numbers ;)

To be honest, game loading times is something I've not been overly impressed with on SSD. It seems faster, but not significantly faster...i.e. there's no wow factor - but that could just be the games I play.
 
Back
Top Bottom