Romex Software Forum

Thx InquiringMind. I thought I would be safe enough by using the default Windows backup. Am I wrong? It currently copies an image of the disk to my 256GB Backup HDD. BTW, I live in the Dutch region of Belgium, called Flanders, so Dutch isn't a problem for me
I also tore the RAID0 apart, I started feeling afraid of data-loss ( ) and switched data around (moving 4TB of data from A to B, and then from B to A takes A LOT of time). I currently use one of my 2TB Seagates that was in the RAID as the "incomplete" torrent drive, after which I move them towards either the other Seagate drive or the 2TB Western Digital One (unused data drive in the benchmarks), depending on the kind of torrent. The sequential speeds dropped to arounds 180MBps read/write, but the random access speeds doubled! So RAID0 isn't always a winner (unless you buy a 1000$ raid controller ).

@Quigg: Did you make any benchmarks? What about the hit rate in Primocache? I can't use primocache for my Data drives since it's cache gets filled with a lot of useless data. But it DOES help on my OS AND Truecrypt drive!

Neglacio wrote:I thought I would be safe enough by using the default Windows backup.

Have you tested it by doing a restore? (to an unused drive). If so and it works, then great (but see caveats below). I've not used Windows Backup (I'm running WinXP) and have relied on Drive Snapshot, since it was the first (as far as I can tell) utility that could backup a system partition while that partition was in use - it's tiny (350KB, can fit on a floppy if ever needed!), very reliable (even worked on systems with memory issues) and self-sufficient (most other imaging software relies on the Volume Shadow Service).

The most important feature is the ability to verify backups once they have been taken, as the last thing anyone wants is to find their backups to be faulty only when their disk fails. This should not be confused with Acronis TrueImage "validation" (which only checks the backup file is valid, not whether it has the right data) as proper verification (comparing backup with current disk) can only be done when the backup is taken and Windows Backup can't do this. Being able to view and extract individual files from an image is also extremely handy (feature built into Drive Snapshot, possible but slightly less easy with Windows Backup).

Another thing to consider is the worst case scenario - if your system drive fails, you won't have a Windows installation and will need to use alternative means of running your imaging software. Acronis and Drive Snapshot supply you with bootable CD images you can use for such situations, but for Windows Backup it looks like you would have to create a new Windows install first. However you can work around this by creating a backup Windows install (on another disk, either a minimalist one or a clone of your main system) and that would also make restores with Drive Snapshot, Acronis, etc more pleasant (usually faster, and you have the option to play Solitaire while the restore is running!).

Neglacio wrote:@Quigg: Did you make any benchmarks? What about the hit rate in Primocache? I can't use primocache for my Data drives since it's cache gets filled with a lot of useless data. But it DOES help on my OS AND Truecrypt drive!

Sorry, I don't know how to import screenshots, but maybe this will do:

-----------------------------------------------------------------------
CrystalDiskMark 3.0.2 x64 (C) 2007-2013 hiyohiyo
Crystal Dew World : http://crystalmark.info/
-----------------------------------------------------------------------
* MB/s = 1,000,000 byte/s [SATA/300 = 300,000,000 byte/s]

Sequential Read : 6804.461 MB/s
Sequential Write : 9250.088 MB/s
Random Read 512KB : 6706.557 MB/s
Random Write 512KB : 10999.220 MB/s
Random Read 4KB (QD=1) : 1546.126 MB/s [377472.1 IOPS]
Random Write 4KB (QD=1) : 1257.981 MB/s [307124.3 IOPS]
Random Read 4KB (QD=32) : 1565.838 MB/s [382284.8 IOPS]
Random Write 4KB (QD=32) : 1316.082 MB/s [321309.2 IOPS]

Test : 500 MB [C: 37.8% (75.5/200.0 GB)] (x2)
Date : 2014/01/10 21:01:45
OS : Windows 7 Ultimate Edition SP1 [6.1 Build 7601] (x64)

I have tried different setups and this was the best result, the pc also 'feels' very fast. Hit ratio is about 39%, but to be honest I don't know whether that is good or bad (or can be improved).

@InquiringMind: I have a bunch of Windows discs laying around here. And even if that doesn't work, I have 2 other computers in the vicinity, and 6 others scattered trough the house. So if my PC fails horribly, I'll still be able to make some recovery discs.

I changed my setup completely again. I'm just using 6-8GB of RAM for my system drive. When gaming, I notice a serious impact on loading speeds, and with a hit rate of 50-90% in-game, that's just awesome.

(Cache-hit rate topped at 92%, and the cache wasn't even filled completely. It's worth noting that I rebooted the pc and browsed around for some hours, before I started playing. I resetted the statistics just before I started playing.)

I started some real-world tests, and I was unable to find any benefits on the Truecrypt drive so I disabled this cache, allocating it for the system drive.

Now I'm in the final stages of my storage set-up. I'm about to order a small SSD (probably 60GB Kingston) and a 1TB HDD (WD Black) to run my games from. The SSD would be used as an L2 cache with Primocache. What do you guys think about this? Or should I just stick with another 256GB SSD for the same price. I have around 100GB of games installed ATM. What do you guys think?
1TB SSHD with (DIY) 60GB L2 or 256GB SSD?
The self-made SSHD would give me a pretty fast HDD with enough space to keep on installing games for the next 5 years or so. Performance is dramatically increased with Primocache.
The SSD gives speed out of the box, but is "only" 2.5times faster for 4-5 times the price... Guess I'll stick with the SSHD

Now, since I already used all my onboard SATA controllers, I'll need to buy another controller +50$