Must read article for SSD enthusiasts

Vertex gets about 2.4MB/s not 12-21MB/s in random writes.
Compare that to Intel's 23.1MB/s.

So at 4K Vertex vs X-25M:
Random read: 32 vs 54 (Intel wins)
Random write: 2.4 vs 23 (Intel is in its own league)

Vertex get about 12MB or so in random writes. Sorry to disappoint. Here is a single 120GBers in my laptop. (The SATA is limited to 120MB~ in case anyone was wondering about the larger sequential read/write being low; it is a X61 Tablet)

2jg3tk8.png



Here a few benches from the 30GBers. Keep in mind the crystals are on a running system with even outlook up and running. I don't care that much about benches really, I care about how my applications run - which with these Vertexes for the $ is the best I've ever had, especially with MFT in use on the secondary partition!

On C: without MFT. Note the random writes, about 11MB/drive.
29ncjno.png


With MFT on D:
2py43dw.png


An ATTO of C:, no MFT:

rgvxnl.jpg
 
:LOL:

edit: the only thing a Vertex is going to "decimate" a X25-M in is number of firmware updates

I decided to try an X25-M and will have one tomorrow.
Can you confirm that it is flashable, in case INTEL decides to update firmware in the future?
Or, have they ever been updated since release that anyone knows of?

I would put more faith in INTEL for future support.
 
I decided to try an X25-M and will have one tomorrow.
Can you confirm that it is flashable, in case INTEL decides to update firmware in the future?
Or, have they ever been updated since release that anyone knows of?

I would put more faith in INTEL for future support.

My logic towards OCZ is the SSD market is something that is SO new, and will change a great deal even in a few months time. A company that is small like OCZ can adapt more quickly.

It is obvious OCZ is being quite agressive with their firmware. I know some don't like this - but I prefer it for hardware that is so new on the marketplace.
 
Results 233.20
System Info
Xbench Version 1.3
System Version 10.5.6 (9G55)
Physical RAM 4096 MB
Model MacBookPro5,1
Drive Type OCZ-VERTEX 1199
Disk Test 233.20
Sequential 204.74
Uncached Write 317.36 194.86 MB/sec [4K blocks]
Uncached Write 228.79 129.45 MB/sec [256K blocks]
Uncached Read 105.55 30.89 MB/sec [4K blocks]
Uncached Read 393.47 197.76 MB/sec [256K blocks]
Random 270.87
Uncached Write 95.19 10.08 MB/sec [4K blocks]
Uncached Write 386.70 123.80 MB/sec [256K blocks]
Uncached Read 2495.42 17.68 MB/sec [4K blocks]
Uncached Read 783.97 145.47 MB/sec [256K blocks]

You have to fill up the drive first.
Looks like the drive is fairly fresh. :p
 
You crack me up.

Will this work? Or do you need a photo with today's paper as well? I was one of the very few people that had called into support with stability issues and got it before they had to pull it. No, I will not distribute it to anyone. It should be public in a few days anyway.
2pob30p.png

I'm sorry, but did I ever say that you Didn't have the firmware? I knew some people had already gotten ahold of it. My point was that until the firmware is finalized, it's almost pointless to compare because it could change at any time.

And, as you can see in your own benchmark, the X25-M will still win in the 4k r.write benchmarks against 4x Vertex.
 
I'm sorry, but did I ever say that you Didn't have the firmware? I knew some people had already gotten ahold of it. My point was that until the firmware is finalized, it's almost pointless to compare because it could change at any time.

And, as you can see in your own benchmark, the X25-M will still win in the 4k r.write benchmarks against 4x Vertex.

You said prove it - suggesting I was lying about it. Why are you suggest you weren't implying it?

I agree - until finalized there isn't any point in comparing. For the cost, I'll take the set of Vertex for overall performance compared to one X25M

I'll have several X25-E's here in a week or so, and will compare one of those against a RAID0 set of Vertex's just for fun and post the results here.
 
You said prove it - suggesting I was lying about it. Why are you suggest you weren't implying it?

I agree - until finalized there isn't any point in comparing. For the cost, I'll take the set of Vertex for overall performance compared to one X25M

I'll have several X25-E's here in a week or so, and will compare one of those against a RAID0 set of Vertex's just for fun and post the results here.

I was saying prove your statement about performance, not whether or not you have the firmware.
 
I decided to try an X25-M and will have one tomorrow.
Can you confirm that it is flashable, in case INTEL decides to update firmware in the future?
Or, have they ever been updated since release that anyone knows of?

I would put more faith in INTEL for future support.

It is "flashable" as PCPerspective mentioned that they had updated the firmware before the review.
 
1275 drops performance significantly in the 512k random read and write areas.

Shoulda stayed with 1199 dammit.
 
Vertex get about 12MB or so in random writes. Sorry to disappoint. Here is a single 120GBers in my laptop. (The SATA is limited to 120MB~ in case anyone was wondering about the larger sequential read/write being low; it is a X61 Tablet)

2jg3tk8.png



Here a few benches from the 30GBers. Keep in mind the crystals are on a running system with even outlook up and running. I don't care that much about benches really, I care about how my applications run - which with these Vertexes for the $ is the best I've ever had, especially with MFT in use on the secondary partition!

On C: without MFT. Note the random writes, about 11MB/drive.
29ncjno.png


With MFT on D:
2py43dw.png


An ATTO of C:, no MFT:

rgvxnl.jpg
what's MFT?
 
what's MFT?

It allows extremely high random writes even with MLC SSD's.

http://www.easyco.com/mft/index.htm

I'll be comparing X25-E's (SLC) in RAID0 to a set of Vertex's (MLC) with MFT technology as soon as I have them here in my possession.

My testing with what I have in my possession is MFT turns the "D" drive into something like one huge RAMdisk, at least for writes. Writes just go into memory prior to being reordered and written to the SSD.

There is certainly a concern however for what that data is doing while hanging out in RAM should there be an issue in the system via powerfailure or other software issue.

For desktop usage though.. who cares. It is smacking fast.
 
awesome article, my thanks to Anand.

when i can get an anand approved SSD with 160GB capacity for 200 GBP along with TRIM support in both the drive and the OS i will probably get one.

160GB is enough for 32GB system and 128GB for games, big enough for me.
 
awesome article, my thanks to Anand.

when i can get an anand approved SSD with 160GB capacity for 200 GBP along with TRIM support in both the drive and the OS i will probably get one.

160GB is enough for 32GB system and 128GB for games, big enough for me.

I would only use SSD's for OS drive, for pure data reads(games) SATA HD's is sufficient...
 
I would only use SSD's for OS drive, for pure data reads(games) SATA HD's is sufficient...

I disagree. I've got my OS, all apps and all games running off a pair of 120GB Vertexes. Everything benefits tremendously. Transformed my computer from my Velociraptors. Speed is speed, not just for the OS.
 
I would only use SSD's for OS drive, for pure data reads(games) SATA HD's is sufficient...

For pure data reads, SSDs are at their best. It is actually on an OS drive where their write issues tend to manifest themselves. That is why you need to make some tweaks to reduce the number of writes to the SSDs.

Don
 
For pure data reads, SSDs are at their best. It is actually on an OS drive where their write issues tend to manifest themselves. That is why you need to make some tweaks to reduce the number of writes to the SSDs.

Don

This is actually not really the case any longer with SSD's that have some cache. Intel's M and the OCZ Vertex are two such examples of MLC disks with some cache on the controller. Benches do not show how they will work in the real world very well.

With the Vertexes you can just align them, partition, install, and turn off defrag. Other stuff is optional. Those basics done it will rock.
 
This is actually not really the case any longer with SSD's that have some cache. Intel's M and the OCZ Vertex are two such examples of MLC disks with some cache on the controller. Benches do not show how they will work in the real world very well.

With the Vertexes you can just align them, partition, install, and turn off defrag. Other stuff is optional. Those basics done it will rock.

Rememeber that every write to the disk is reducing it's lifetime so it might be better to remove as many of them to ramdrives as possible.
 
How many PC/servers do you have with +5 year old drives in?

Around here? Still a couple, plus we still use 5+ year old HDDs for backup purposes.

It's nice to throw a 5 GB HDD into an old clunker used for testing purposes.
 
You backup to old media? Seriously? Wow....

Personally none of my "spinning" spindles are IDE. That includes my box, the GF's, the linux server, and the WHS. I do still have a few small IDE drives laying around (couldn't sell them for anything worth while). I have flash drives that are bigger.
 
Around here? Still a couple, plus we still use 5+ year old HDDs for backup purposes.

It's nice to throw a 5 GB HDD into an old clunker used for testing purposes.

You backup to old media? Seriously? Wow....

Personally none of my "spinning" spindles are IDE. That includes my box, the GF's, the linux server, and the WHS. I do still have a few small IDE drives laying around (couldn't sell them for anything worth while). I have flash drives that are bigger.

I have to second that "wow"....backing up to really OLD media kinda negates backing up in the first place.
 
I have to second that "wow"....backing up to really OLD media kinda negates backing up in the first place.

Not really. It's best to have multiple backups; utilizing old media for additional backups makes sense if the drive would just be sitting around unused anyway (or sent to a landfill).

I'd agree with you if one only used old media for backups.
 
Not really. It's best to have multiple backups; utilizing old media for additional backups makes sense if the drive would just be sitting around unused anyway (or sent to a landfill).

I'd agree with you if one only used old media for backups.

We use daily, incremental backups on rotating media, in our experience even old HDDs are still reliable enough to be used for such purposes. We use SD cards for backups as well, though. With multiple GBs per card they're large and cheap enough that we can use two of them for each backup (mirror).
 
Looks like Intel has something for the X25 series. It will be interesting to see what they release in the next few weeks.
 
Back
Top