Originally Posted by
Lance B
..... What does Crystal Disk Mark do?
It just checks storage system speeds.
It's a small(2.4Mb) exe, rather than a large bloated system performance software suite .. so you don't clutter your PC with needless stuff.
When you run it, it pops open a window, and you click the [ALL] button or whatever the box says, and it gives you a brief rundown of what sort of speeds you can get through to your drives.
It creates a small series of values for each test, and you can save the results to a txt file, which is really only important if you're trying stuff like different configs(eg, raid or not) .. or drivers or firmware .. etc.
Code:
-----------------------------------------------------------------------
CrystalDiskMark 4.1.0 x64 (C) 2007-2015 hiyohiyo
Crystal Dew World : http://crystalmark.info/
-----------------------------------------------------------------------
* MB/s = 1,000,000 bytes/s [SATA/600 = 600,000,000 bytes/s]
* KB = 1000 bytes, KiB = 1024 bytes
Sequential Read (Q= 32,T= 1) : 179.773 MB/s
Sequential Write (Q= 32,T= 1) : 177.051 MB/s
Random Read 4KiB (Q= 32,T= 1) : 0.704 MB/s [ 171.9 IOPS]
Random Write 4KiB (Q= 32,T= 1) : 1.365 MB/s [ 333.3 IOPS]
Sequential Read (T= 1) : 179.102 MB/s
Sequential Write (T= 1) : 175.760 MB/s
Random Read 4KiB (Q= 1,T= 1) : 0.672 MB/s [ 164.1 IOPS]
Random Write 4KiB (Q= 1,T= 1) : 1.389 MB/s [ 339.1 IOPS]
Test : 1024 MiB [X: 50.2% (1403.7/2794.5 GiB)] (x3) [Interval=5 sec]
Date : 2015/07/09 13:17:15
OS : Windows 7 Professional SP1 [6.1 Build 7601] (x64)
So the end result is a txt file that looks something like that above.
That was from my fastest HDD(all mechanical) drive connected in any way to my PC.
If you trust the Windows transfer window for speed of transfer, it's usually wrong(in fact almost always wrong), as there will be some buffering involved somewhere during that transfer.
eg. I see about 250MB/s for a short while if transferring a large amount of data(eg. 200Gig) .. and then it slows down as the buffering in Windows is incapable of handling all that data.
And the Windows transfer window then slowly reverts to the speed that the actual drive is capable of achieving.
So if you ever try to tweak, or tune or change something to achieve a more efficient system, Windows reporting isn't an effective way to work out if you are getting closer to achievign your goal!
Of course there are other HDD performance software .. I just use that because I got onto it and .. well, I now use it.
A scenario as to how you'd use Crystal Disk Mark(CDM): lets say you want to transfer 1Tb of data from the Drobo(to anywhere else).
You say that it probably transfers at 200Mb/s(which would be very nice!), but this doesn't take into account any buffering that the system is doing.
I don't know anything about Drobos, but they'd have to have an amount of internal memory for buffer. Add that to any buffering Windows is doing in the background too, and it's not hard to get that sort of speed. So what you may be seeing in real terms is the speed of the buffering(which is always fast!).
CDM, simply shows you haw fast your disks are in real life .. eg. over a long transfer, once the internal buffers have been exhausted
I use Capture NX2 for my intricate NEF editing, and while I always edit the raw file, the reality is that it's a TIFF file by design. So when you edit a raw file in CNX2, in effect you're opening a 200+Mb tiff file. I also get less than 1 sec opening times on my PC, even tho I know my HDD's can't actually transfer at those rates.
Strangely with CNX2, even tho I showed you the fastest drive in my system(could do 170MB/s, but now has slowed to 140MB/s), my main photo storage drive is a wee bit slower than this faster drive. It's rated at about 125-130MB/s .. but CNX2 renders the files off that internal(to the PC) drive faster compared to the same file off the faster but external(USB3) drive.
FWIW: the reason I want one of those 5 bay HDD's you linked to earlier is that I have 4 devices holding 5 drives sitting on my desktop(next to my desktop).. it's a cluttered mess.
I have a few that I rely on heavily, but one of the devices I have is a HDD docking station. A device that allows you to simply drop a HDD into it with ease(it looks like a toaster).
I have this because in my circle of people I know, are always bringing me something to try to fix .. or un-clutter, or de-clog or wipe viruses .. or whatever, so I need an easy way to add remove drives. So the ability to add/remove drives easily/toolessly is appealing too.
That USB3 docking station drove me crazy. It was my first USB3 device, and I purchased a 3Tb Seagate barracuda drive to go with it. I'm thinking(feeling smug).. now I have those annoying all day 1.5Tb transfers all sorted. On numerous occasions I've had to leave the PC on over two days just to backup my images(and other important data).
The time itself isn't a problem, but the chances of a power failure during that time was.
Anyhow, the USB dock turns out to be a dud, due to a badly engineered internal (USB-SATA) chip.
It was slow and it would drop out at every opportunity. Connected to a USB2 port tho, it was fine. But I didn't get a more expensive USB3 dock to use it in USB2 mode.
I nearly threw it out, but instead persevered with updating firmware/drives, and everything else in the middle. Just wouldn't work in USB3 mode.
Anyhow, fixed that .. but due to the bad design, even after the manufacturer finally released a new firmware that fixed the dodgy chip, it was still much slower than the other USB enclosure I got for the fast barracuda drive. So I know the drive is fast, and I know the dock is slow(half the speed of the enclosure). So when I first transferred that 1.5Tb of images it took nearly 20 hours. As a test, I re transferred the same data(maybe a touch more) to the same drive but now in the fast and stable enclosure, and it only took about 3 hours or so.
I think I lost about a dozen raw files, due to that stupid docking station when it was dropping out on USB3.