Velocity Solo X2 Now Available

[sunnyj]

Originally Posted by letscopro

I installed the Velocity x2 yesterday and I have to say I'm not that impressed with what I'm seeing. There is a definite increase in the read time, but the write time is almost the same as the internal sata as you can see from the image. The ADATA SX900 maximum sequential read and write speeds are 550/530 MB per second.

I was expecting closer to 500 MB per second on both read and write speeds.

I'm waiting to hear back from Apricorn about this.

can you do a speed test using the aja system test?

[letscopro]

Originally Posted by sunnyj

can you do a speed test using the aja system test?

Thanks for the suggestion.

[VirtualRain]

Originally Posted by matthewtoney

For some reason the Black Magic Test is refusing to run on my RAID0 stripe on the X2. (tells me that drive is read-only, although it plainly is not - anyone with any idea on that?)

I've just run a bunch of tests on all my disks though in Xbench and I'm getting read/writes on it with the X2 very similar to what the other person was seeing. In my case - 566MB/s read and 579MB/s write.

Originally Posted by letscopro

Thanks for the suggestion.

It would seem that AJA is a much better benchmark than BlackMagic... (which incidentally I've been having problems with as well)

Although the Adata drive is not achieving it's full potential, I wonder about the M4...

According to StorageReview... a single M4 is capable of
Read/Write: 497/243

And in RAID...
Read/Write: 870/520

They use IOMark on a PC though... http://www.storagereview.com/crucial_m4_ssd_raid_review

[sunnyj]

ive also had problems with the blackmagic system test and thats why i prefer aja's application.

[sunnyj]

Is the Velocity x2 a reliable card in general? I'm thinking of getting one for my Mac Pro.

[letscopro]

Originally Posted by sunnyj

Is the Velocity x2 a reliable card in general? I'm thinking of getting one for my Mac Pro.

It was just released and I've been using it for less than 48 hours. So far, so good, but time will tell.

[VirtualRain]

Originally Posted by matthewtoney

For some reason the Black Magic Test is refusing to run on my RAID0 stripe on the X2. (tells me that drive is read-only, although it plainly is not - anyone with any idea on that?)

I've just run a bunch of tests on all my disks though in Xbench and I'm getting read/writes on it with the X2 very similar to what the other person was seeing. In my case - 566MB/s read and 579MB/s write.

----------




No this is just on 1 X2 card and yes I'm using the power from the optical bay to power the second drive connected to it.

Dude, what happened to you? Lookin forward to an AJA bench on your setup

[matthewtoney]

Originally Posted by VirtualRain

Dude, what happened to you? Lookin forward to an AJA bench on your setup

Oh I'm still here I haven't posted any results with AJA because it doesn't want to run against my RAID0 stripe either. (so far xbench is the only that does for me) I'll try it again tonight when I'm back in front of that machine to see just what it was that AJA was doing - it was different than the complaint I got trying to run the Black Magic test on the stripe.

[albert1028]

For those who had the previous version of the Velocity Solo and for current users of the new version. I was wondering if you have had any stability issues.

On my mac pro, I realized that I had stability issues with the SSD installed on the older card. For example, almost everytime I ran disk utility it would hang. And often times it would randomly hang when engaging in intensive processing tasks, running photoshop, indesign, or any other adobe software for that matter.

I have decided to go back to using one of the 4 HDD slots on my Mid 2010 Mac Pro.

Any stability issues or freezes or hangs?

Does it matter which slot one uses as well?

Thanks.

[matthewtoney]

Originally Posted by albert1028

For those who had the previous version of the Velocity Solo and for current users of the new version. I was wondering if you have had any stability issues.

On my mac pro, I realized that I had stability issues with the SSD installed on the older card. For example, almost everytime I ran disk utility it would hang. And often times it would randomly hang when engaging in intensive processing tasks, running photoshop, indesign, or any other adobe software for that matter.

I have decided to go back to using one of the 4 HDD slots on my Mid 2010 Mac Pro.

Any stability issues or freezes or hangs?

Does it matter which slot one uses as well?

Thanks.

I've only had mine in for a week (and I did not have the previous version) but I'm not having any stability issues or seeing that behavior in any of those apps at all. My performance level with 2 SSDs on it in a RAID0 stripe is slightly less that I hoped for but still quite good. I have mine in slot 4 - the one closest to the drive sleds.

[letscopro]

Originally Posted by VirtualRain

That's disappointing for sure... let us know what Apricorn has to say for themselves.

It will be interesting to see what matthewtoney has to report with his dual M4s.

Here is the reply I received from Apricorn.


Hello letscopro,

I apologize for the delay, I was awaiting a response from a coworker. The Blackmagic disk speed test is using compressed data, not uncompressed data. Because of this you will see drastically reduced speeds. If you us an older version of the speed test software you will see much higher speeds since it is using uncompressed data.

The Blackmagic disk speed test is using compressed data, not uncompressed data. Because of this you will see drastically reduced speeds. If you us an older version of the speed test software you will see much higher speeds since it is using uncompressed data.

What's New in Disk Speed Test Version 2.1
Some SSD's use hidden compression when writing data to make their benchmarked speeds appear faster. Disk Speed Test will now measure the true speed of these SSD's so you know if they are suitable for high quality uncompressed video capture.

Last paragraph in the help says...

"Important note about Solid State Disk (SSD) speeds

Some models of SSD cannot save video data at the speed indicated by the manufacturer because the disk uses hidden data compression to reach these higher write speeds. This data compression technique can only save data at the manufacturer’s claimed speed when storing simple files or simple data, such as blank data. Video data includes video noise, and more random pixel data which does not compress much, so the true speed of the disk is seen.
Some SSD’s can have up to 50% lower write speed than the manufacturer’s claimed speed, so even though the disk specifications claim an SSD is fast enough to handle video, in reality the disk is not fast enough for real time video data capture. Hidden data compression mostly affects capture and often these disks can still be used for real time playback.
Use Blackmagic Disk Speed Test to measure accurately if your SSD will be able to handle uncompressed video capture and playback. Blackmagic Disk Speed Test uses data to simulate the storage of video so you get results similar to what you will see when capturing video to a disk. This will let you find models of SSD that work well for video capture. In our testing, we have found larger newer models of SSD, and larger capacity SSD’s are generally faster. Blackmagic Disk Speed Test also tests the speed of disks connected to eSATA docks and other interfaces, which can affect disk performance."


Testing my Vertex 3 with pre 2.1 version

(see image)

and with 2.1

(see image)

Same SSD, 2 very different results.

Sincerely,

Brandon
Apricorn Support"

[VirtualRain]

Originally Posted by letscopro

The Blackmagic disk speed test is using compressed data, not uncompressed data.

Yeah, of course... makes total sense. Glad that's sorted out. I had forgot about that and wasn't sure what controller your Adata drive uses... It must be a Sandforce drive.

I also asked Apricorn if there was any bottleneck on the card and what was limiting it to a total of 800MB/s of throughput (with dual SSDs in RAID0) this is what they said...

The bottleneck is the PCI lane. If you look at our x1 products, their advertised speed is 400. The x2 is just twice the PCI capacity, so 400x2=800
Makes sense?

Real world test have shown us that each PCI lane will realistically deliver about 400MB/s.

Please let me know if you have any questions.

I'd still like to see Mr. Matthewtoney get a better benchmark of that dual M4 setup as that's what I'd be running

Any resume from sleep issues?

[matthewtoney]

Originally Posted by VirtualRain

Yeah, of course... makes total sense. Glad that's sorted out. I had forgot about that and wasn't sure what controller your Adata drive uses... It must be a Sandforce drive.

I also asked Apricorn if there was any bottleneck on the card and what was limiting it to a total of 800MB/s of throughput (with dual SSDs in RAID0) this is what they said...



I'd still like to see Mr. Matthewtoney get a better benchmark of that dual M4 setup as that's what I'd be running

Any resume from sleep issues?

I wish I could give (and see) some results - just tried grabbing the latest version from their website for AJA, and still the same. Both Black Magic and the AJA System Test refuse to run against a RAID0 stripe of 2 Crucial M4's on this Velocity X2 card. So far xBench is the only thing that runs but if anyone has any other benchmarking suggestions I can try that.

(actually I could use HD Tune when booted up in Windows from another drive to test the array - I have MacDrive 9 installed and it can read/write the mac drives, even the RAID0 stripes but I don't know how much overhead that adds)

[TableSyrup]

I had the previous version of the Solo - ran in a 1,1 without issues - was streaming audio sessions from it in Pro Tools - rock solid

I have the new version on it's way, and will be running it in a 5,1 for the same purpose - I am expecting it to be trouble free.... don't see why there would be issues... but you never know

[Yetihunter]

My "2012" mac pro is still sitting in the box, since Monday.....
I'm assuming it's still 5.1, right?...pretend it's 2010, I'll know when I
start her up.

A.) Does it boot?
B.) Is it hardware controlled striped raid 0 ?
C.) If it boots, are apps snappier than ssd in sata II slot?
D.) Are read/write for heavy bandwidth files quciker?

It supports two ssds, why doesn't it allow for two mounted ssds?
That really gets me. These certainly don't look like uber expensive
PCB's. You only see those with DIY audio projects.

[lastmile]

I've been waiting for this to be released but I haven't bought one yet. Debating going to a Hackintosh...

I was thinking about how to add two disks to this. I only need it for a boot drive but being able to add a second drive in the future would be nice.

I suppose you could get a piece of plastic as long as the PCI-E slot and use it to mount a second disk. You'd drill two sets of four holes in it for mounting the disks. It would sit under the Velocity Solo and be attached using the screws that hold the SSD on the card. The other end would be supported by the notch in the grey plastic piece used to support long PCI-E cards like graphics cards. The second disk would be mounted on the plastic just as the first is mounted to the card itself.

Then you'd need to connect the drive to the card. It'd be nice to have the SATA connector on the drive facing the SATA connector on the card and just use a very short SATA cable for a straight connection. But I don't know if you can find a SATA cable that short and making your own by cutting down a cable doesn't seem very reliable. Then you need power. If you're not using both 6-pin graphics card power connectors there are 6-pin to SATA power adapters available. But since the Mac Pro uses a different 6-pin connector you'd have to hack your own Mac Pro to SATA power version by cutting the wires and reattaching the right side from each type of cable.

I'm using both 6-pin connectors for my 5870 but I figure there's enough power left for an SSD. I just need to find something that lets me branch off a SATA power connection. Or you could pull power from the second optical bay. That only makes sense if you have two optical drives and can still get enough power to power an SSD. Otherwise just put the second SSD in the second optical bay and run a SATA cable up to it from the Velocity Solo. Of pull power from one of the internal drive bays but that seems like a waste of a bay.

Given the lack of additional power connectors inside the Mac Pro and the lack of an eSATA port I'd have preferred to see the additional connection on the Velocity Solo be eSATA. That's easy enough to do with one of those SATA to eSATA cables with the eSATA port stuck in the PCI-E slot cover. But it wastes a slot unless you can hack the eSATA port into available space in a currently installed PCI-E card. It doesn't look like there's room to put a port on the slot cover of the Velocity Solo itself - the SSD is too close.

[VirtualRain]

Originally Posted by lastmile

Given the lack of additional power connectors inside the Mac Pro and the lack of an eSATA port I'd have preferred to see the additional connection on the Velocity Solo be eSATA. That's easy enough to do with one of those SATA to eSATA cables with the eSATA port stuck in the PCI-E slot cover. But it wastes a slot unless you can hack the eSATA port into available space in a currently installed PCI-E card. It doesn't look like there's room to put a port on the slot cover of the Velocity Solo itself - the SSD is too close.

I wish they had simply duplicated the SATA/power connector on the other side of the card so you could attach two drives to one card without screwing around trying to power the second drive... or put two connectors on the same side so you could stack two drives... whatever was easiest and allowed it to fit in a standard slot.

Given all this, I wonder if you can run dual x2's in two different slots and RAID0 the drives on both. I think I'll ask them. You might actually get better performance this way since each drive will have PCIex2 at it's disposal so two 500MB/s drives could theoretically result in 1GB/s throughput in RAID0.

Meanwhile, has anyone got a decent benchmark from two drives in RAID0 on this thing to work?

[letscopro]

Originally Posted by letscopro

Thanks for the suggestion.

I bought a second Adata SX900 today. Here are the RAID 0 results.

[matthewtoney]

Originally Posted by letscopro

I bought a second Adata SX900 today. Here are the RAID 0 results.

Those numbers are actually a little better than I expected, but my only results are out of xbench and my 2 Crucial M4's. I *still* do not understand why I can't run the AJA test though on my RAID0 set. I've just tried it again with the latest version from their website and although it works against the other drives in my system, choosing my RAID0 SSDs on the X2 and clicking "Start" just does nothing - the write/read scores just sit at 0 Are you running 10.8 or something earlier?

[maximage]

Are you choosing an area on your hard drive that can be read and written to? Don't just choose the drive you want to test, but a folder inside it like your desktop folder, that has access for writing to.

[matthewtoney]

Originally Posted by maximage

Are you choosing an area on your hard drive that can be read and written to? Don't just choose the drive you want to test, but a folder inside it like your desktop folder, that has access for writing to.

You can do that with the Black Magic test, but with AJA you can only choose a volume with the read/write test and not a folder or location. You *can* choose a file itself for the "Disk Read Existing File" test with AJA and that *does* run for me when choosing a file on the RAID0 SSD array but its the read/write test on the array volume that does not.

(btw, using that disk file *read* test, AJA is telling me 612.2MB/s on the RAID0 of the M4s)

[letscopro]

Originally Posted by matthewtoney

Those numbers are actually a little better than I expected, but my only results are out of xbench and my 2 Crucial M4's. I *still* do not understand why I can't run the AJA test though on my RAID0 set. I've just tried it again with the latest version from their website and although it works against the other drives in my system, choosing my RAID0 SSDs on the X2 and clicking "Start" just does nothing - the write/read scores just sit at 0 Are you running 10.8 or something earlier?

I'm running 10.8.

[VirtualRain]

Originally Posted by letscopro

I bought a second Adata SX900 today. Here are the RAID 0 results.

That's disappointing. You would expect 800MB/s. I wonder what's going on.

[matthewtoney]

Originally Posted by VirtualRain

That's disappointing. You would expect 800MB/s. I wonder what's going on.

This is only a PCIe X2 card from what the list for specs and I assume these two SATA ports are not each getting their own X2 bandwidth - the speeds seem pretty much what I'd expect for that personally, but I do wonder where Apricorn gets the 800MB/s that they claim on their site.

[VirtualRain]

Originally Posted by matthewtoney

This is only a PCIe X2 card from what the list for specs and I assume these two SATA ports are not each getting their own X2 bandwidth - the speeds seem pretty much what I'd expect for that personally, but I do wonder where Apricorn gets the 800MB/s that they claim on their site.

I asked my contact at Apricorn what the bottleneck on the card was for RAID0 arrays...

Me:

what's the bottleneck on the X2?… I see the max speed expected advertised on your website is 800MB/s fro two drives connected in RAID0, but that's a bit low for two lanes of PCIe and a pair of SSD's capable of 500+MB/s each. Is there some other logic at play?

Them:

The bottleneck is the PCI lane. If you look at our x1 products, their advertised speed is 400. The x2 is just twice the PCI capacity, so 400x2=800
Makes sense?

Real world test have shown us that each PCI lane will realistically deliver about 400MB/s.

Please let me know if you have any questions.

This makes sense... there should be nothing stopping this card from achieving sequential reads of 800MB/s assuming the drives are up to it... which the two Adata drives in RAID0 certainly are.

I've emailed them again... hoping they can chime in on these results above.

----------

Originally Posted by letscopro

I bought a second Adata SX900 today. Here are the RAID 0 results.

What slot are you running the X2 in and what year Mac Pro? While it shouldn't matter, it might be worth trying this card in one of the slots designed for a graphics card. I'm not familiar with the slot configuration in 2008 and earlier but in 2009 onwards, that would be one of the two bottom PCIe slots. I'm just wondering if there's something limiting the throughput on the Intel chipset or Apples implementation of the PCIe bus on the x4 slots.