• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

So... X25-M or Vertex.

vhx

Weaksauce
Joined
Nov 23, 2007
Messages
66
I have been debating on which of these drives to go with and with every new review I see, I change my mind about which drive to choose. So what it comes down to at the moment seems to be:

Vertex: 120GB, less expensive per GB, writes advertised as up to 180 MB/s (however this seems to only be seen in synthetic tests only, maybe improved with future firmware?), seems to have random real world and synthetic performance (it looks inconsistent with each review), performance degrades heavily after use more so than the X25-M, has a wiper tool you can run to restore the performance, but not perfected and can corrupt data, constant firmware updates.

Intel X25-M: 80GB, more expensive per GB, writes only up to 70 MB/s but seems to still outperform the Vertex in real world tests, synthetic and real world tests seem more consistent with each other, performance degrades but a lot less of a percentage than the Vertex does, no tool to recover performance (have to wait for TRIM for 7 to be finalized).

Really this is confusing the hell out of me. Every benchmark either just shows synthetic tests, old firmware, or used performance with no reference to new/used. A lot of people are hyping the Vertex, but it seems like it doesn't compare to the X25-M all that much except in a few instances. Any help please?
 
Last edited:
Are you going to be writing massive chunks of data (5 second write x 180MB/s is almost 1GB) to the drive, and is the drive speed going to be the bottleneck for that?

Is cost more important than performance?

If the answer to both is no, get the Intel. If yes to one, get the Vertex.
The Intel smokes the Vertex in random I/O, which is felt most day to day.
 
Agreed. Intel kills the Vertex in random writes. I love my X25. So amazingly responsive no matter what I'm doing.
 
The vertex is not in the same league as the X-25, simple as that. If you have the cash for one, the intel drive is better, period.
 
Is fragmentation the reason performance degrades over time on all SSD drives? Or is this something separate?
 
SSDs have a thing called wear-leveling, which means all available cells get written to so no cells are repeatedly written to. When all the available cells on an SSD have been written to, you reach what's called "steady state performance" which means from now all, in order to write new data, old cells have to be erased. Once this level is reached, performance more or less has leveled off.

Data fragmentation has almost no effects on SSD performance given its low seek times. This is probably a separate topic altogether, but fragmentation regular desktop use rarely effects hard drive performance since most writes are sequential and not random like in a server environment.

The fragmentation hysteria online was specific to Intel drives. The Intel drives are affected by data fragmentation but it was fragmented at a partition table level so it wasn't something de-fragmentation could fix. The new firmware "fixed" this issue on the X25-M. Intel released this new firmware for mainstream users. This fragmentation is actually part of their aggressive wear-leveling algorithm. Enterprise applications are about random read/write performance so the fragmentation does not affect their use--in fact, it prolongs the lifespan of the drive.
 
Intel X-25M is the better of the two as long as you don't mind paying a few bucks more.
 
The only thing that Vertex got any real advantage over is in capacity, aside from price. Though Intel also got their 160GB model, there are also 250GB Vertexes. :p
Both are of course ridiculously expensive, and as stated already before the Vertex is less expensive per GB. I would still get the Intel's drive if I were to buy.
 
I was looking for the OCZ Vertex, the g.Skill Falcon and the Intel X25-M.
OCZ's forum appears to be excellent while the G.Skill looks somehow poor.
And everybody seem to praise the Intel SSD.

Here just a comparison between my old 40GB Seagate HDD (Top) and some results (I found somewhere) about the Intel X25-M SSD (Bottom).

seagate40gbhoir.png


intelx25-mvsqo.png




Found the Intel X25-M (80GB) for $378.- and the X25-E (32GB) for $495.-

Has anyone heard anything about that SSD prices going to fall by 50% still this year ?
 
I'm still testing my Intel X25-M 80GB drives in RAID0. So far I'm VERY pleased with the results. So far they are roughly in line with everything that's been posted here.
 
I'm still testing my Intel X25-M 80GB drives in RAID0. So far I'm VERY pleased with the results. So far they are roughly in line with everything that's been posted here.

I just got two of the X25-M 80GB drives.
Planned also to run them in RAID 0.
Did you try a single drive before you set them up as RAID 0?
If so, did you feel a difference?
(I am not concerned much with benchmarks)
 
I just got two of the X25-M 80GB drives.
Planned also to run them in RAID 0.
Did you try a single drive before you set them up as RAID 0?
If so, did you feel a difference?
(I am not concerned much with benchmarks)

I didn't run an OS off of one but benchmarks show about half the speed of the RAID0 array. I know you said you weren't concerned about that, but everything I've read shows almost precise linear speed improvements from going from a single drive to two in RAID0. My tests echo that information. Right now I'm experiementing with stripe sizes and I've tried using the ICH10R on my motherboard as well as the LSI MegaRAID 8308ELP controller. Clearly the LSI isn't really optimized for SSD's because the performance difference between the ICH10R and it are relatively small. Normally it kills the onboard controllers at anything and everything by large amounts. Not so with the SSD's.

So far the 64k and 128k stripe sizes yield similar and nearly identical performance and a 16k stripe size yields less performance.
 
Last edited:
That's great!
I am planning to used the ICH10R on a P5Q Deluxe with the 2 X 80GB X25-M's.
Since I will run VISTA (and probably Windows 7 later), I should not have to fool with stripe sizes.

Thanks.
 
That's great!
I am planning to used the ICH10R on a P5Q Deluxe with the 2 X 80GB X25-M's.
Since I will run VISTA (and probably Windows 7 later), I should not have to fool with stripe sizes.

Thanks.

If you are going to run them in a RAID-0 configuration, you do. The onboard RAID controller will probably default to a 64k or 128k stripe size if you just accept the default configuration. From what I can see, that is probably the optimal configuration or close to it. The OS itself has nothing to do with the stripe size configuration. One thing you have to pay attention to is 4k write performance as that's the default allocation size for NTFS partitions. Stripe size does indeed effect this. Granted the Intel X25-M kills any mechanical hard drive on the market performance wise but for what these drives cost, not getting the absolute most out of the configuration is a waste. This is why I'm testing various stripe configurations.

I'm pleased to see that the ICH10R performance is close to my expensive RAID controller in regard to SSD performance. This may be due to the fact that the controller is optimized for SAS and not SATA. Normally this wouldn't be the case, but that controller is slated to be placed in my server system and as a result it won't be available for use in my gaming computer. (Additionally, I don't have room for it normally since I typically run 3 Geforce GTX 280 OC's in 3-Way SLI.)
 
We aim to please. :) Just because I'm curious, I plan on testing 512k and 1MB stripe sizes because my MegaRAID supports those. Onboard RAID controllers aren't nearly so versatile, so for those people using those controllers, those stripe sizes won't be an option.
 
Look at the price development over the last 4 months for the X25-M SSD !!

Currency in Euros:

pricedevelopmentm1rm.jpg


Source
 
i personally recommend a X25-M 80g, which I am using.
1st, the writing speed does not really matter cos what you want is that use it as your system hard drive, not a storage. even at 70m/s, its not slow at all.
2nd, it indeed reads faster than 200m/s as advertised, most others do not.
3rd, intel released the new firmware regarding the slow down problem
It is an amazing fast SSD on my rig. you wont disappoint.
a $280 will be a good price
 
Sorry, but I don't understand why the scale suppose to start with zero Euro.

The scale does not start at 260 Euro but at 358 Euro and ends at 263.88 Euro.

What advantage would it have for any scale, to include figures that don't exist ?

By the way, I did not make this scale. So why should I modify something I have no information about in the first place ?
 
Last edited:
Well that can't be right. 4 Raptors in RAID 0 would kill all those SSDs in sequential performance. And RAID doesn't exactly improve access times, in fact it gets worse, so by that metric, it would be:

Intel > Samsung > Indilinix > Raptor > RAID Raptor
 
Yep, my 4x160GB raptor raid 0 pulled over 400MB a sec in sequential reads.

but who cares, that doesnt matter

My machine is SIGNIFICANTLY faster now with a single falcon 128GB than the 4 raptors. I can open ALL of Office 2007 in like 1 second. I can open visual studio 2008 as fast as notepad. I can REboot a winxp sp3 VM in like 12 seconds. (That's including shutting down...)

I couldn't do any of that with the raptors.

The 4x stripe was very noticeably faster than a single raptor also, as I have ran that config in the same box. (Access times don't really get worse, or not much worse)

Intel > Samsung > Indilinix > RAID0 Raptor x4 > RAID0 Raptor x2 > Raptor
 
Last edited:
Sorry, but I don't understand why the scale suppose to start with zero Euro.

The scale does not start at 260 Euro but at 358 Euro and ends at 263.88 Euro.

What advantage would it have for any scale, to include figures that don't exist ?

By the way, I did not make this scale. So why should I modify something I have no information about in the first place ?

Well by "starting the scale at 0 EUR" he means that the graph's sloping and the overall negative regression won't look as significant.

The X-25 still dropped in price like a rock, but it's still much too expensive for my tastes. :(
 
As an owner of an X25-M, I would definitely say buy it over the Vertex. OCZ are constantly trying to improve their drives, which is great; however, to get such improvement requires constant firmware updates on your end, which means reinstalling windows. The x25-m just lets you not deal with that, and gives you better performance as well.
 
Really you can update the x25m without reinstalling? That takes the cake right there. Still vertex raid 0 recloning the drive of the os/apps doesn't take that long and you're back and runnin again :) To me it's worth it. But for a new buyer that's something to consider.
 
Yep, the Intel update for my X25-M was just something I burned on a disc and did outside of the OS. Never needed to reinstall my OS and all my files were OK after the fact. Also, not that it matters much, but my WEI score jumped from 7.4 to 7.7 after I did the firmware update...

As for the price, all I know is that the 80GB X25-M was about $600 back in November/December last year. You could get one from newegg for like $280 with a combo deal when I got one in March this year (haven't checked prices since). So they've already dropped by over 50% of their price as far as I was concerned. I'd guess they'll be approaching the $200 by year's end.
 
Back
Top