First OCZ SSD Core Review

8steve8 said:
I'm astonished by how fluffy hard drive reviews are, including guru3d's,

i guess thats why i started doing my own... I have a raid article in the works...

anyway.. it seems OCZ continues to dodge the issue, claiming its this or that.

My analysis is still 100% valid.
there is a ~250ms latency to do a random write.
for huge writes that's not a big deal (obviously)

the random write performance they claim is using large request size, 100MB per request in one benchmark!
Click to expand...
looking forward to the RAID article, what are you testing?
 
8steve8 said:
i guess thats why i started doing my own... I have a raid article in the works...
Click to expand...

I'm glad to hear that. Though, hopefully, it isn't on fakeraid. I posted on the OCZ forums about RAID controller compatibility and got like nothing.
 
8steve8 said:
I'm astonished by how fluffy hard drive reviews are, including guru3d's,

i guess thats why i started doing my own... I have a raid article in the works...

anyway.. it seems OCZ continues to dodge the issue, claiming its this or that.

My analysis is still 100% valid.
there is a ~250ms latency to do a random write.
for huge writes that's not a big deal (obviously)

the random write performance they claim is using large request size, 100MB per request in one benchmark!
Click to expand...

8steve8, if you notice any data corruption issues, also please let us know. That seems to be a big thing in the OCZ forums, so if you can vouch for data integrity, at least we can say that there isn't something inherently wrong with all of the drives.
 
when i first saw this drive i was pretty intrigued, but after following this thread i just don't think this thing is ready to be used...i'd be interested to know how many others are getting data corruption and other issues..i guess i'll have to wait until the next revision to get one of these
 
in all of my testing on ich9r (raid enabled in bios, which apparently enables ahci)
i saw no data corruption...



when i installed vista sp1 x64 on the disc i saw some signs of corruption, as well as a sort of stuttering , like when opening tabs in firefox and such...

the most sense i could make of it was ocz claiming ahci and its accompanying NCQ causes real problem with these drives, but with all intel ichxR southbridges, u cannot turn off ahci and still use raid, the bios has 3 options: ide, raid, or ahci, and according to intel raid=raid+ahci..

i have not tried installing vista in IDE mode, so i cannot confirm if this fixes the issue...


it seems like u can't just buy one of these things, plug it in, install vista and have everything work...
 
i have a thread on ocz's forums with my issue, their responses are very vague, if not purplexing at times.
 
I got one and can't even get vista to run on it, with AHCI off and all that jazz...

it's speedy as a 2nd drive though for music, videos, etc. :D

Gonna replace my raptor once SLC solid state drives become affordable. MLC the future my ass.
 
Jester FPS said:
I got one and can't even get vista to run on it, with AHCI off and all that jazz...

it's speedy as a 2nd drive though for music, videos, etc. :D

Gonna replace my raptor once SLC solid state drives become affordable. MLC the future my ass.
Click to expand...

What kind of issues were you having?
 
Cyrilix said:
What kind of issues were you having?
Click to expand...

I formatted and installed vista about 4 times on it, and each time it would hang in dos, I could not get it to boot into vista at all and installing it took longer than usual, due to the slow write speeds.

also, just having it as secondary drive and though I formatted it already, I get a Vista error boot screen in DOS. It says the installation is corrupted and I have to manually select which Vista to boot (have to select my raptor install).

I can't test the write benchmark with hdtune.. This is a minor annoyance but I've heard SLC drives act just like hard drives without any of these problems so that's the standard I'll uphold for any future SSD purchases..

on a positive note, it makes for an excellent media storage device. Just not with an htpc since most htpc will have a tv tuner and the tv tuner will require faster write speeds for optimum performance.

hdtuneoczcoressd6480108ee6.png
 
Why are you using expensive SSD's for reasons other than your OS ? Unless you do a lot of media editing or other intensive things that pressure your hard drive then you dont really benefit from the low access times and high throughput rates.
 
I had no idea the drive was going to do so poorly on the small files. This review just confirms all the others. If this is the expected performance from OCZ, I am extremely disappointed. Once I get my 2 64GB Core drives back from RMA, I will probably try to sell them on E-bay. Unless they find someway to improve the small file performance and ofcourse the file corruption issue-

I'm so pissed off-
 
Interesting, results.
Does this seem just as unreliable than the SuperTalent drives?
On Newegg they had quite a few negative reviews b/c the drives were failing after a few weeks.
 
I guess I'm going to have to delay my upgrade to SSD a bit. Oh well. Lots and lots of work I need to get done before I'm building the system in question anyway.
 
Well, a guy - Jeff_rys- over on the OCZ Forums has posted his results for his system:

Intel duo-core 3 ghz
2 x 2 GB 1333mhz OCZ memory
Asus P53E board
Radeon 2600 XT 512 mb graphic card
Adaptec 5805 controller (256 mb cache)
6 x OCZ 128 gb SSD in Raid0
Vista Ultimate X64

http://ocztechnologyforum.com/forum/attachment.php?attachmentid=6276&d=1218664695

How are only 6 drives giving that high of read/write speeds???

Is ATTO reading the information incorrectly? Is his controller card causing incorrect readings? 6x 140MB=840MB/S Where is the 1042MB/S coming from?
 


System:
  • E8500 Q740A479T 2L7 / 1313
  • CPU Cooling: Dtek Fuzion Rad Tower Box
  • Gigabyte GA-EP45 Extreme rev 1.0 F6c bios
  • 128MB Gainward FX5200 PCI
  • 2x1GB Crucial Ballistix PC2-8500 dual channel kit
  • Memory cooling: 120x25mm Thermaltake 81cfm fan
  • 4x 32GB OCZ Core SSD 2.5" solid state disks raid 0 - 128K stripe
  • Pioneer 215 DVD-RW
  • 1200W Silverstone OP1200
  • WinXP Pro SP3 Nlite with integrated Intel Storage Manger 8.5.0.1032 raid drivers

bootup1.jpg
bootup2.jpg






more tests at http://i4memory.com/showthread.php?t=8944
 
Count me in for data corruption. Was running OSX on my macbook pro for ~3 weeks. Then all the sudden I started getting beach ball constantly. SSH started saying it couldn't write out to the known_hosts file (which was now gone). Couldn't even touch a file in ~/.ssh/ without it generating an i/o error. Looks to me like their write leveling is junk to be quite honest... and that it's exhausting sectors nearly immediately.
 
Update:

Got the drives back from OCZ. And guess what? Same exact problems. I ended up purchasing an Adaptec 31605 to see if the onboard cache would help, as well as maybe handle the SSD better than onboard controllers. Unfortunately they don't even work at all with the high-end Adaptec controller. Can't write anything to the drives without the raid array tripping an alarm on the card. Can create the array fine and format, but once you try to write anything, it locks up and fails. Tried different controller BIOS's and everything, nothing worked.

OCZ Has done nothing to address the issues that many people are having. Been over a month since i've had my drives and no single fix has been stated- Doesn't seem like they are even close to figuring out what's going on either. All OCZ does and tell you to try this, then try that, then try this- No concrete configuration seems to compatible so far- Every chipset out there has been reported to have problems by forum members.

OCZ is blaming Windows and Motherboard manufacturers. However, other MLC drives, such as the super talent mlc drives, seem to not have these issues at all. It's obvious OCZ did not test these drives before they started selling them, and expect their customers to do the Beta testing for them.

Also Core V1 and Core V2 are having the same exact issues- So please be aware if you're thinking that the V2 is going to be better- Now... what can I do with $600 of file corrupting, windows stuttering SSD drives? Can't sell them on Ebay, as the buyer would have the same problems.

I'm posting this so nobody else has to throw away money like I just did on a product that I thought was from a reputable company- Call me Biased if you want, and Irate, but I've spent over 30-40hours installing, configuring, updating again and again on a product I really wanted to work. So, yeah, I'm a bit Irate-
 
Bummer!

I was hoping you've have better luck with the returned drives.

Looks like these drives aren't quite ready for prime-time.

Unfortunately, this isn't the first time and won't be the last that a company pushes thru an immature technology and the unsuspecting consumer gets the weiner.

OCZ should step up to the plate, wipe the egg off their face and at the very least offer full credit for the drives. Of course, the ultimate apology would be a full refund for the purchase price, but I'm not holding my breath for that one.

I have no stake in it, but if you want to start screaming on all the techie boards, I'll post right behind you as a concerned consumer that will shun all OCZ products 'till the sun goes down.
 
Griff805 said:
Update:


OCZ Has done nothing to address the issues that many people are having. Been over a month since i've had my drives and no single fix has been stated- Doesn't seem like they are even close to figuring out what's going on either. All OCZ does and tell you to try this, then try that, then try this- No concrete configuration seems to compatible so far- Every chipset out there has been reported to have problems by forum members.

OCZ is blaming Windows and Motherboard manufacturers. However, other MLC drives, such as the super talent mlc drives, seem to not have these issues at all. It's obvious OCZ did not test these drives before they started selling them, and expect their customers to do the Beta testing for them.
-
Click to expand...


Shame, i thought OCZ could bring affordable SSD drives to mass market.. obviously they are failing and dodging the issues to try and sell as many as possible before word really gets out!
 
MrGuvernment said:
http://www.directron.com/ssdhd.html
Click to expand...

Ya know, I'm not opposed to throwing a little money to where it needs to go. This should be obvious when I've spent 300.00 for a boot drive and only used 80 GB of it to get fast times.

Just for shits and giggles I followed your Directron link to gander at the SSd drives avaliable. For the mere pittance of 627.00 I can get a Super Talent DuraDrive AT 3.5-inch 32GB SATA Solid State Drive that has the drive stats of

supertalentpk6.jpg


Read at a max of 65 MB/sec and write of 50MB/sec! :eek: Pretty frigging ridiculous.

I'll grant you, access time of 0.1 ms is great, but gimme a break.

Unless you put at least two of these things in RAID 0, it's gonna be a downgrade to a 5400 RPM drive.

Unless I'm missing something here, for now, WD VR FTW.
 
58,675 IOPS is extremely high though-

This is what 2x OCZ core drives do in RAID 0 on an Intel ICH10R controller:

 
Griff805 said:
58,675 IOPS is extremely high though-

This is what 2x OCZ core drives do in RAID 0 on an Intel ICH10R controller:

Click to expand...

OK, I'll go with ya. I had to investigate this IOPS stuff because I've never heard of it.

I'm not proud, and have always readily admitted that the entire scope of my knowledge would be the size a grape pit in the world's largest stadium.

But, how does that spec translate into real world performance?

From my limited searching it doesn't apply to any transfer speeds. It seems to be, pardon my crude analogy, akin to "thinking speed".

Or, as usual, am I waaayyyy off base?
 
IOPS is what servers love!!! the higher the better, the raw speed isnt great but does seem to those supertalent ones are more for servers, and i am sure they actually work like they say, dont corrupt data and work fine in raid and any controllers :)
 
MrGuvernment said:
IOPS is what servers love!!! the higher the better, the raw speed isnt great but does seem to those supertalent ones are more for servers, and i am sure they actually work like they say, dont corrupt data and work fine in raid and any controllers :)
Click to expand...

So you don't know either. :D
 
http://en.wikipedia.org/wiki/IOPS

From what I can tell it is akin to the metric of contrast ratio for monitors. A completely useless statistic that makes no sense when you translate it from one technology to another.

Correct me if I am wrong, but I think typical disk drives take a huge hit in IOPS because of their terrible performance on lots of small random accesses where the seek times mean a lot.
 
IOPS is, from what I remember, the number of independent seek (most often followed by read) requests carried out per second. This is great for servers, because by definition they are handling many unrelated requests concurrently (say, the [H]|F server feeding Thread requests to 500 people at a time). It is related to the access time, but not in a directly linear fashion since a seek is just the amount of time that it takes the head to jump to the right place on the platter (obviously not the same for SSD), before it can start reading in the information.

This is less of an issue for home users, since the vast majority of their requests are linear, or at least closely related. The nearest thing I could really think of would be either a swap file (where seek time is still key), or perhaps a couple of large, concurrent BitTorrent downloads.
 
@Griff805: You said that the V2 drives have simmilar issues.
Can you maybe point a thread that had this mentioned.

Over at XS there is a guy who says everything seems to be ok...
 
Thanks for the links!

How's the simultaneous read and write performance?
Can somebody that has V2 maybe measure that and file copy to and from SSD to another drive?
 
Old Hippie said:
So you don't know either. :D
Click to expand...


I know, i just didnt explain it as good as below:

(i have been doing IOPS testing all week on some new servers to get the best configuration out of it, IOMeter tests it all!


mavalpha said:
IOPS is, from what I remember, the number of independent seek (most often followed by read) requests carried out per second. This is great for servers, because by definition they are handling many unrelated requests concurrently (say, the [H]|F server feeding Thread requests to 500 people at a time). It is related to the access time, but not in a directly linear fashion since a seek is just the amount of time that it takes the head to jump to the right place on the platter (obviously not the same for SSD), before it can start reading in the information.

This is less of an issue for home users, since the vast majority of their requests are linear, or at least closely related. The nearest thing I could really think of would be either a swap file (where seek time is still key), or perhaps a couple of large, concurrent BitTorrent downloads.
Click to expand...
 
You must log in or register to reply here.
Back
Top