Intel X25-M, round 2 : 10 SSDs compared - BeHardware
>> SSD

Written by Marc Prieur

Published on May 5, 2009

URL: http://www.behardware.com/art/lire/761/


Page 1

Introduction



Launched in September 2008, Intel’s X25-M was a real bomb on paper. With the help of MLC flash, Intel was in a position to offer an 80 GB model at €500, while also giving amazing performance for that time with read speeds of 250 MB/s and, more modestly, writes of 70 MB/s.

To achieve this, Intel put the accent on its 10 channel architecture, very effective wear levelling as well as more efficient usage of blocks of memory by including a cache for writing. This allowed Intel to announce that its X25-M supported writing over 100 GB/day for 5 years. At a time when everyone doubted how dependable SSDs were in the long term, an announcement of this type from Intel amounted to manna from heaven.


However, although the Intel X25-M blew us away in the first tests, we did note a very significant fall in performances when we subjected the SSD to very different loads (see this article). Intel’s only response was to qualify these results as “expected”. Expected for Intel perhaps but this didn’t correspond to our expectations with regard to SSDs: to see performance fall is one thing, especially as the X25-M was way out in front of the competition anyway, but it is quite another for performances, under some circumstances, to fall lower than those of 5400 rpm drives!

Unfortunately, at the time, not many in the tech media world had noted the problem and it was only after an article by an English language site came out in February and created a bit of a stir on the other side of the Atlantic, that Intel decided to correct what they’d told us was “expected behaviour”. The corrrection came in April in the form of the new 8820 firmware, available on this page.
Inside the Intel
Here is a reminder of what an X25-M looks like under the bonnet.


Inside the drive, you find a 20 chip PCB (10 on each side) marked with the Intel and Micron logos, both companies being partners in the production of flash memory, an Intel controller and a 16 MB Samsung SDRAM chip.


Page 2
Usure and the X25-M, the test

Usure of the X25-M in practice
Here are the results of usure of the X25-M with the new firmware. For a full recap of this phenomenon as well as the results obtained on the Vertex with the same protocol, read this article (pages 2 and 3).

Case 1: Sequential performance – Sequential access only

First of all, here are the sequential read and write speeds (in KB/s) when we carry out only this type of access (one read, one write, one read, one write, one read and a final write, each time on the whole of the SSD, ie 80 GB):


There isn’t a great variation, with read speeds around 211 MB/s and writes at 75 MB/s to 72 MB/s.

Case 2: Sequential performance – Random then sequential access

This time we started with a new drive (after HDD Erase) that is then “used” by writing by block of 4 KB randomly for 30 minutes. We carried out the same test as in case 1.


This time there was a significant fall in speeds with reads dropping from 211 MB/s to 147 MB/s on the first run. As you can see however, after carrying out a sequential write the SSD regains its original level of performance on the second run.


In writes, note the big variations on the first run, which correspond to the time when the SSD is reorganising itself. The fall isn’t as significant however; 66 MB/s on the first run and then back to 72 MB/s. With the previous firmware the average write speed sometimes went under 20 MB/s in this type of test!
Case 3: Random performance
Now here is an observation of the change in performance (in I/Os per second) during random writes of 4 KB files. We launch the test on a new SSD, 6 times 5 minutes, then a second time on an SSD whose cells have been previously filled seqentially.


On a new SSD, the X25-M starts with a very good level of performance, more than 6500 I/Os per second and then falls rapidly to 1262 on the 6th run. On a used SSD, it starts at 3022 and falls quickly to 370, but bizarely climbs back to 500 on the last run. In any case this is more than ok for desktop use but you’ll note that with the old firmware random write performace was higher: stability has then been gained at a cost.
The test
For this test, we compared the Intel X25-M with different models:

- OCZ Core V2
- OCZ Apex
- Samsung PM410
- Samsung PS410
- Samsung PB22-J 64 GB
- Samsung PB22-J 256 Go
- Mtron MOBI 3500

The OCZ Core V2 represents what is “best” with the JMicron JMF602/MLC combination, with the Apex a RAID version (two SSDs in one). The Vertex is based on the most recent Indilinx controller. The Samsung SLC and MLC SSDs are in a way the references for the previous generation, and the Mtron MOBI 3500 is an interesting alternative. The Samsung PB22-Js are the new generation of Samsung SSD, here in two capacities as their specs are different. For information we also include the performances of a VelociRaptor, a 3"1/2 Samsung SpinPoint F1 640 GB drive and a 2"1/2 Samsung SpinPoint M5 160 GB.

Different measurements were taken in the course of this comparison. First of all we were interested in the “synthetic” performances of these drives: average access time, sequential speeds and I/O in random and sequential access. Next, were more practical tests, first of all involving an applicative performance index based on PC Mark Vantage and then an evaluation of writing, reading with various groups of files These groups were composed of:

- A collection of large files: 6 files (on average 2.2 GB) totalling 13.2 GB
-Medim sized: 7.96 GB of 10,480 files (each averaging 796 KB)
-Small sized: 2.86 GB of 68,184 files (each averaging 44 KB)

The source or target of reading or writing on the drive was a RAID of three VelociRaptor 150 GB drives that replace the raid of two 150 GB Raptors used in previous tests. This type of measurement is worthwhile because, while the sequential speed gives us an idea of the performance in copying large files, things can be different with smaller ones.

The test machine was based on an X38 chipset mounted on an ASUSTeK P5E motherboard while Serial ATA ports were configured in the bios in AHCI (Advanced Host Controller Interface) so that NCQ could be used, all operating with Vista SP1.

We also added new practical tests so as to provide more useful data for those who want to work out if going for an SSD over there standard hard drive is worthwhile. For this we timed various operations on an other machine based on a P5QC, QX9770, GTX 280 and 2x2 GB of DDR2-1066.


Page 3
Synthetic performance

Synthetic performance
Mesured with h2benchw, SSD cache speed is really astonishing. At 0.13 ms, the X25-M gets a good scrore, very close to that of the Vertexes.


We continue with the speed, still with the help of h2benchw. In contrast to other benchmarking software such as HDTune or HDTach, h2benchw carries out a true sequential test because it reads or writes the entire drive, whereas the others jump between zones to reduce test time.


The X25-M has read speeds of 200 MB/s, in line with the competition, but write speeds are not as good at 74.9 MB/s.

Now on to measurements of the number of 100% random accesses that can be supported by these storage systems, using the IOMeter. We carry these out in small blocks of 4 KB and see how many the SSD can support each second, first using a type of sequential access and then random access. For this test we use a configuration file that carries out 10 tests for 10 minutes per test: 0, 25, 50, 75 then 100% random reads then 0, 25, 50, 75 and 100% random writes.


Sequential reads: the X25-M leads with 13366 I/Os per second, ie over 52 MB /s. Random performance is lower as with all the SSDs, but much less than with hard drives. The X25-M therefore has a good level of performance but the Vertex range and, above all, the MOBI 3500 do much better.


The hard drives do much better in writes than reads, their advantage being essentially linked to the cache. For SSDs however, things are harder, especially as the test was carried out on a device on which the pages had already been filled previously with sequential writes. In spite of everything the X25-M does quite well as it compares to a VelociRapter with only the Vertexes in front. Note that in the tests on the previous page that were made up of random writes only, the X25-M attained 500 I/Os per second.


Page 4
PC Mark Vantage

PC Mark Vantage
We now move on to less synthetic tests, starting with an index of hard drive performances in PC Mark Vantage. FutureMark reproduces a set of reading/writing tasks on the drive, namely a Vista start-up, loading of applications (Word, Photoshop, IE, Outlook), the manipulation of multimedia files (photo, video, music), games (loading of Alan Wake) and disk scan with Windows Defender.


With 26,594 points the X25-M is in the lead, in front of the OCZ Vertexes.


Page 5
Management of files

File copying
This brings us to file copying. We measure reading and writing of diverse groups of files. These groups were composed of:

- A collection of large files: 6 files (on average 2.2 GB) totalling 13.2 GB
-Medim sized: 7.96 GB of 10,480 files (each averaging 796 KB)
-Small sized: 2.86 GB of 68,184 files (each averaging 44 KB)

The source or target of reading or writing on the drive was a RAID of three VelociRaptor 150 GB drives.


Reading large files the X25-M scores over a hugely impressive 240 MB /s plus! Note however that once the size of files falls, performance drops back to a level comparable to other drives and not very far off a standard hard drive.


Write speeds are the weak point on this Intel SSD. Happily however although speeds for large files are slow, performance with small files is ok.


Page 6
Practical tests

Practical tests
We have at last decided to add other practical tests to our SSD tests, in particular so as to provide more useful data for those who want to work out if going for an SSD over their standard hard drive is worthwhile. For this we timed various operations on an other machine based on a P5QC, QX9770, GTX 280 and 2x2 GB of DDR2-1066.

- Windows Vista start-up
We measure the time needed to start up Windows Vista freshly installed with the drivers. The measurement is from running the bios (disappearance of P5QC logo) to full Windows desktop display with the cursor not showing the egg times.

- Installation of Service Pack 1
Here, the time needed to install Service Pack 1, the installation file itself being situated on the SSD.

-Start up of Windows Vista SP1+Kaspersky+Word+Excel+Outlook+Photoshop
After having installed Kapersky Antivirus 2009, the Office suite and Photoshop CS4, we put shortcuts to Word, Excel, Outlook and Photoshop in the Startup directory of the start menu. The time measured is from running the bios to the end of Photoshop CS4 launch.

-Loading of “Train” level in Crysis Warhead
After having installed and patched Crysis Warhead we launch it with the –DEVMODE option and load the train level from the consol with the “map train” command.

We weren’t able to carry out the tests on the 256 GB version of the Samsung PB22-J, as we no longer had this SSD in the lab.


The X25-M is the fastest on starting up a "new" Vista, though the differences are very small between the different SSDs.


The installation of SP1 is to the advantage of the X25-M, giving it a better score than the Vertexes.


While there were no major differences with just the start-up of Vista, it was different when we launched very heavy multi applications. The X25-M is in the leading group along with the OCZ Vertexes, the Mtron MOBI 3500, and the Samsung PS410 is out in front.


As you can see, moving from an HDD to an SSD doesn’t necessarily change much in terms of loading of games. Of course, during loading there are read operations that are carried out on the storage device but also a lot of the time is linked to the creation of the environment by the processor from this data – this time can’t be cut down. Therefore, even if there is a gain it is slight and the X25-M is well placed.


Page 7
Power consumption

Power consumption
All the SSDs tested here are completely silent. We concentrated therefore on measuring consumption at rest and reading/writing with IOMeter. Here are the results for the X25-M 80 GB:


The scores are fine.


Page 8
Conclusion

Conclusion
The new X25-M firmware allows the Intel SSD to put the problem we picked up on last year behind it, namely the lack of stability of performances with varying loads. What Intel originally described us as “expected behaviour” has in fact been reviewed and in the end recognised as a problem that needed solving.

Better late than never and at a time when the Indilinx controller, used by the OCZ Vertexes, looked set to dominate the top end MLC SSD market for a few months, Intel has thrown its hat back into the ring. Indeed, excepting its sequential writes on large files, an area that is pretty negligeable for SSDs given current capacities, the SSDs are very close in performance.


Of course the Vertexes, and we hope eventually all Indilinx SSDs, will include the TRIM command. This isn’t yet sure to be integrated on the X25-M and gives the Vertexes an advantage. However it is less important on the X25-M, as the differences in loads have a much less significant impact on performances.

With the X25-M, then, Intel is proposing a viable alternative to the OCZ Vertex. This is particularly so in view of the fact that the OCZ Vertex has increased in price recently because, among other things, of the increase in the cost of flash memory. Intel has reduced the price of its X25-M. This means that the 80 GB X25-M is now priced between the 60 and 120 GB OCZ Vertexes, which makes much more sense. It’s a shame that a more affordable 40 GB version is not available.



Copyright © 1997-2014 BeHardware. All rights reserved.