OCZ Vertex Review

[NoStateofMind]

AnandTech

Very fast! :shocked:

 

[shabby]

They're not using retail firmware and ocz will most likely be changing it to 230mb/sec read - 85mb/sec write with much higher iops.
There's lots of talk about this on the ocz forums.

 

[kensiko]

When the review from Anandtech is coming? This week?

 

[coolVariable]

Originally posted by: kensiko
When the review from Anandtech is coming? This week?

Would like to know too.

 

[Idontcare]

Originally posted by: coolVariable

Originally posted by: kensiko
When the review from Anandtech is coming? This week?

Would like to know too.

They've been very quiet on the SSD reviews since their Intel SSD review which had the OCZ smackdown. No Samsung, Vertex, Apex reviews.

Maybe they got blacklisted and have to wait to get the drives thru retail channels versus getting advanced review samples like all the other shills that refuse to bench small file random writes. :roll:

The OP's link for instance, page after page of sequential read/write benches. Yeah, gee I wonder who politely asked the reviewer to restrict themselves to those benches in the review...:roll:

 

[kensiko]

We have the confirmation from OCZ that Anandtech has a Vertex in their hands.

 

[Blain]

Yes, it's fast.
But for that kind of $$$, I'd rather use two VelociRaptors in RAID 0 (with a solid backing-up, of course).

One other issue...
Why didn't Benchmark Reviews include the Intel X25-E in the testing? :roll:

 

[kensiko]

Maybe because of the huge price difference ??

Blain you never had a SSD I think. Yes a velociraptor is fast, but for reading, a SSD is MUCH faster, believe me!

 

[Blain]

Originally posted by: kensiko
Maybe because of the huge price difference ??

Blain you never had a SSD I think. Yes a velociraptor is fast, but for reading, a SSD is MUCH faster, believe me!

* The Intel X25-E is $404. the OCZSSD2-1VTX120G is $450 AR.
No "huge price difference" there.
* The Intel X25-E for $404 is 32GB, other drives in their benchmarks were 16GB, 32GB, 64GB, etc.
So capacity doesn't seem to have factored in.
> Which brings me right back to my original question...
Why didn't Benchmark Reviews include the Intel X25-E in the testing?

Sure "read" times are a bit faster, but the RAID 0 VelociRaptor's mopped the floor with the OCZ drive in write times.
My tasks required writing data as well as reading it.

 

[Idontcare]

Originally posted by: Blain

Originally posted by: kensiko
Maybe because of the huge price difference ??

Blain you never had a SSD I think. Yes a velociraptor is fast, but for reading, a SSD is MUCH faster, believe me!

* The Intel X25-E is $404. the OCZSSD2-1VTX120G is $450 AR.
No "huge price difference" there.
* The Intel X25-E for $404 is 32GB, other drives in their benchmarks were 16GB, 32GB, 64GB, etc.
So capacity doesn't seem to have factored in.
> Which brings me right back to my original question...
Why didn't Benchmark Reviews include the Intel X25-E in the testing?

Sure "read" times are a bit faster, but the RAID 0 VelociRaptor's mopped the floor with the OCZ drive in write times.
My tasks required writing data as well as reading it.

Ouch! mine eyez! :shocked:

Seriously though maybe the review didn't include the X25-E simply because they didn't have one in-house at the time of the review.

Many of these second-tier review sites are going to be nothing more than a website manned by some dude with a computer in his mom's basement. Sure he's an enthusiast and is getting sponsors, but until he gets his site to big-time status the operating cash is going to be kind of low.

For example just look at techreport, here they are reviewing quad-raid0 X25-E (no way they bought those at newegg) but their testbed is an antiquated Pentium4 CPU with an equally ancient ICH7R raid controller. Sure Intel pimped them four X25-E's for the review, and they managed to convince Adaptec to spot them a raid controller card as well, but the best rig they have in their basement is a P4!?

Guess what the next techreport SSD review is going to look like considering that those X25E's and Adaptec card were no doubt required to be returned to the sponsors once the review was published.

You can't expect premium rib-eye steaks from a butcher who thinks hamburger is fine dining.

 

[fyleow]

..

 

[fyleow]

..

 

[Blain]

Originally posted by: fyleow
Here's a review with random writes:

http://www.pcper.com/article.p...670&type=expert&pid=12

The Vertex does random writes at ~20MB/s while the Intel does it at ~70MB/s

The Intel SSD used in the PC Perspective review is the X25-M, same as in the Benchmark Review article.
The Intel SSD drive that I was asking about is the X25-E.

 

[kensiko]

Originally posted by: fyleow
Here's a review with random writes:

http://www.pcper.com/article.p...670&type=expert&pid=12

The Vertex does random writes at ~20MB/s while the Intel does it at ~70MB/s

This review is not with the latest firmware. In the OCZ Forum, we voted for higher I/O than sequential write for the shipping firmware, and we won! I expect the random write to be better now.

 

[Idontcare]

Originally posted by: kensiko

Originally posted by: fyleow
Here's a review with random writes:

http://www.pcper.com/article.p...670&type=expert&pid=12

The Vertex does random writes at ~20MB/s while the Intel does it at ~70MB/s

This review is not with the latest firmware. In the OCZ Forum, we voted for higher I/O than sequential write for the shipping firmware, and we won! I expect the random write to be better now.

That review also isn't benching small-file random writes, the smallest filesize Yapt v0.3 benches is 64KB.

We are interested in 4KB random writes (i.e. small file), exactly the sort provided by crystaldiskmark.

 

[kensiko]

http://www.ocztechnologyforum....p=344962&postcount=221

That is already a good benchmark. But Anandtech has a trustful comparison database, so this is why I'm waiting for their review.

 

[Idontcare]

Originally posted by: kensiko
http://www.ocztechnologyforum....p=344962&postcount=221

That is already a good benchmark. But Anandtech has a trustful comparison database, so this is why I'm waiting for their review.

I see even the sequential 4K writes are pretty meager there in the Atto bench screenshot.

I see OCZ practices what they preach to the reviewers of their SSDs - no random small-file write benching.

 

[fyleow]

..

 

[frostedflakes]

Originally posted by: Idontcare

Originally posted by: kensiko

Originally posted by: fyleow
Here's a review with random writes:

http://www.pcper.com/article.p...670&type=expert&pid=12

The Vertex does random writes at ~20MB/s while the Intel does it at ~70MB/s

This review is not with the latest firmware. In the OCZ Forum, we voted for higher I/O than sequential write for the shipping firmware, and we won! I expect the random write to be better now.

That review also isn't benching small-file random writes, the smallest filesize Yapt v0.3 benches is 64KB.

We are interested in 4KB random writes (i.e. small file), exactly the sort provided by crystaldiskmark.

Exactly. Notice that even the JMicron based Apex drive does fair in those 64kB random write benchmarks. 4kB writes are where the crappy drives fall on their face, though. Unless we see some decent 4kB random write benchmarks, I wouldn't get too excited guys. Vertex is looking to be nothing special. I'm guessing random write performance will be on par with the Samsung controller drives -- good enough to prevent stuttering and other random write bottlenecks, but definitely not at a level that compares to the Intel drives.

 

[IntelUser2000]

Originally posted by: frostedflakes

Originally posted by: Idontcare

Originally posted by: kensiko

Originally posted by: fyleow
Here's a review with random writes:

http://www.pcper.com/article.p...670&type=expert&pid=12

The Vertex does random writes at ~20MB/s while the Intel does it at ~70MB/s

This review is not with the latest firmware. In the OCZ Forum, we voted for higher I/O than sequential write for the shipping firmware, and we won! I expect the random write to be better now.

That review also isn't benching small-file random writes, the smallest filesize Yapt v0.3 benches is 64KB.

We are interested in 4KB random writes (i.e. small file), exactly the sort provided by crystaldiskmark.

Exactly. Notice that even the JMicron based Apex drive does fair in those 64kB random write benchmarks. 4kB writes are where the crappy drives fall on their face, though. Unless we see some decent 4kB random write benchmarks, I wouldn't get too excited guys. Vertex is looking to be nothing special. I'm guessing random write performance will be on par with the Samsung controller drives -- good enough to prevent stuttering and other random write bottlenecks, but definitely not at a level that compares to the Intel drives.

The Intel drive can actually stutter because its too fast, but it will be very random and sporadic unlike the OCZ drive because the problem is different.

See the Intel drives to Write Combining. Write Combining is using bursts rather than sending data little by little. Write Combining effectively reduces write amplification(the ratio of number of pages/block) because it can burst close to the size of the block. The drive uses log table inside the controller to manage what to do.

Imagine what happens though if the random write speeds greatly exceed the garbage collection speed. If there are enough IOs being fed to the drive that is random writes the lookup table will be optimized for random write speeds and possibly flood the drive because it will become faster.

Of course it can be changed over long time by changing the usage pattern. That sounds similar to what Intel's talking about when saying it optimizes according to usage doesn't it?

 

[Idontcare]

Originally posted by: IntelUser2000
The Intel drive can actually stutter because its too fast, but it will be very random and sporadic unlike the OCZ drive because the problem is different.

See the Intel drives to Write Combining. Write Combining is using bursts rather than sending data little by little. Write Combining effectively reduces write amplification(the ratio of number of pages/block) because it can burst close to the size of the block. The drive uses log table inside the controller to manage what to do.

Imagine what happens though if the random write speeds greatly exceed the garbage collection speed. If there are enough IOs being fed to the drive that is random writes the lookup table will be optimized for random write speeds and possibly flood the drive because it will become faster.

Of course it can be changed over long time by changing the usage pattern. That sounds similar to what Intel's talking about when saying it optimizes according to usage doesn't it?

Are there any reported observations of this form of performance degradation? (that's not a challenge to your claims, I am curious to see how one characterizes this facet of SSD stutter)

One thing we can count on is being in the murky dark on anything about Intel drives that aren't super until Intel releases their next-gen drives. Intel likes to demonize their past offerings a little when they release a new product, so you can bet when the new SSD's are released then Intel marketing will start the demonization process on the existing SSD line.

 

[kensiko]

Here is a good review:
http://www.cdfreaks.com/review...ries-120GB-SSD-Review/

Look at real world test.

I am definitely buying an OCZ Vertex.

 

[Idontcare]

Originally posted by: kensiko
Here is a good review:
http://www.cdfreaks.com/review...ries-120GB-SSD-Review/

Look at real world test.

I am definitely buying an OCZ Vertex.

Kudos to you for finding this review, and kudos to the author for running and reporting the crystaldiskmark bench results:

http://www.cdfreaks.com/review...Review/Benchmarks.html

So we are looking at ~12MB/s random write for 4K files. About 1/4 the reported random writes reported for the Intel drive that everyone likes.

 

[coolVariable]

The CDfreaks benchmark was done by someone who practically is a mod on the OCZ tech forums.
She would never ever say anything critical about the vertex.
Take the review for what it is worth ... IMO it might as well be an OCZ in-house review.

 

[kensiko]

We will see with Anandtech. But I'm sure it's not an OCZ in-house review.