by Aleksey Meyev , Nikita Nikolaichev
09/09/2009 | 02:47 PM
High-tech products usually pass a few typical stages of development on their way to conquering the market. First, there appear pilot and small-series samples in which the manufacturing technology is being polished off, the characteristic traits are being defined and various ideas, both lucky and not, are being tested. Next goes the period of maturing when product models change one another very quickly, their functionality constantly improving and their childhood diseases getting cured. And then goes the mature period proper. It is when the technology is developing at a slower rate, resulting in a steady improvement of key characteristics. It is during this period that the manufacturers come up with alternative technologies that might lead to new type products. The maturity will eventually transform into a decline, and it is very important not to miss this moment. All such processes go on very quick in the PC hardware industry and everyone can easily name a few examples of such cyclic development. This was the case with floppy diskettes and data CDs. The same goes for CPUs with their ever-changing architectures.
And for the last couple of years we have been witnessing one more example of this cycle as a new product type – a Solid State Drive (SSD) – is coming to market. The newcomer is so aggressive that hard disk drives, although obviously at their prime now, are already not superior to it in some aspects. Early SSDs could only rival HDDs in a few parameters, but the latest SSD models with multi-channel controllers deliver such a high performance that not only the fastest of HDDs but even RAID arrays made out of them fail to equal!
To be specific, early SSDs had the following indisputable advantages over HDDs:
There were obvious downsides, though:
The first problem could only be solved by a reduction of the price of flash memory. Watching the chip prices constantly declining in the last year, many analysts were very optimistic about the possibility of SSDs to quickly win a large market share. However, in the first half of 2009 the price of flash memory chips was growing up and it is only in the summer that there was a certain decline. A tough price competition was the main reason for that. Coupled with the reduced demand, it made the supply exceed the sales while the profitability was at its minimum. The market just could not digest that much of memory. To get any profit, the manufacturers increased their factory prices. As a result, SSDs have not been declining in price as quickly as predicted in the last year. Generally speaking, we don’t think that SSDs will become competitive to HDDs in terms of pricing any time soon. Today, the cost of 1 gigabyte of SSD storage varies from 2 to 5 US dollars whereas HDDs are as cheap as 0.1 US dollar and getting even cheaper. With such a huge gap in terms of pricing, most users are going to prefer the cheaper alternative and will only choose SSDs when they really need high performance.
The introduction of multi-channel controllers has helped SSDs not only overtake but even outperform HDDs in terms of sequential speeds. Even with relatively cheap MLC memory it is possible to create an SSD that has two times the sequential read speed of the best of modern HDDs. SSDs are not so brilliant at writing but at least comparable to HDDs in this aspect, too. And if you really want a high speed of writing, you may consider multi-channel SSDs based on SLC memory which is more expensive than the MLC variety and comes in smaller-capacity chips but ensures a better write speed and higher reliability.
The limited number of write cycles of flash memory cells is aggravated by the nonuniformity of typical loads (some cells work harder than others) and the lack of SSD-oriented optimizations in modern OSes. Many enthusiasts that used early SSDs as system disks were annoyed at their short service life due to the decay of memory cells. The first SSD that tried to solve that problem by utilizing effective controller algorithms was Intel’s X25-M. It was quite a surprise for us when we tested it (you can refer to this review for details about the deterioration of flash memory with use).
It is during that test session that we also encountered a specific feature of the current generation SSDs. Their performance depends on what the previous load has been and may degenerate greatly in some situations. We will talk about this problem later. Right now let’s take a look at the devices we are going to test. These are Corsair’s P128 and three SSDs from Intel.
Corsair, the well-known maker of high-performance RAM, USB flash drives and power supplies, has taken to SSDs very seriously. The company now offers as many as three SSD series: Legacy, Performance and Extreme. Interestingly, these three series are all based on MLC memory chips. The Legacy series includes early SSDs from Corsair which are based on somewhat outdated controllers (this is how quickly things change in the industry: multi-channel controllers came out just a couple of years ago, but now there are already outdated models in the product range). The company doesn’t make a secret of the differences between the other two series. The Performance line includes models with capacities of 64, 128 and 256 gigabytes based on a Samsung controller. The Extreme series is based on an Indilinx Barefoot controller and comes in smaller capacities: 32, 64 and 128 gigabytes. Corsair claims the former series to deliver stable performance and the latter, maximum performance.
So, we’ve got a Performance series sample. For each model in the series, the manufacturer declares sequential read and write speeds of 220MBps and 180MBps, respectively. The junior models do not have fewer channels. Instead, they just use smaller-capacity memory chips. The SSD does not differ from others of its kind externally. It is a standard box measuring exactly like a 2.5-inch HDD and having the same mounting holes. The four screws in the cap called to be unfastened, so we removed the cap (which is made of some aluminum alloy):
We can see a PCB with controller, RAM chip and 16 flash memory chips located on both sides of the PCB. The controller is Samsung S3C29RBB01-YK40 also known as PB22-J. It is the latest generation controller which is used in modern SSDs from Samsung and some other brands. The other chips are manufactured by Samsung, too. These are the 128MB SDRAM chip Samsung K4X1G323PD-8GC6 clocked at 166MHz and 16 NAND MLC chips K9HCGZ8U5M from this series. Quite expectedly, Corsair did not develop its own SSD design but used a partner offer in the way of a ready-made Samsung SSD.
This is a familiar product for us as we had it in our hands when making our acquaintance with Intel’s SSDs. We must note that we test this SSD with older firmware (version 8160) although there is new firmware (version 8820) available for it. We have a reason for that. We will be able to evaluate the new firmware by the next product whereas the test results of two SSDs with the same controller and very similar chips but with different firmware will help us in our evaluation.
Next goes the 160GB SSD from Intel based on MLC chips. It comes out with the same specs as Intel had promised: 250MBps for sequential reading and 70MBps for sequential writing. The case of the device has become slimmer but there is a protruding plastic frame at the edges which increases the thickness of the case to the standard thickness of 2.5-inch drives, i.e. 9.5 millimeters.
Recently Intel transitioned its MLC-based SSDs from 50nm to 34nm tech process. But in a few days the new SSDs were withdrawn from the market due to a bug in their firmware: data might be lost on the drive in some cases if the user set an access password. This is an example of a childhood disease, just to remind us that the life of SSD technology has begun but recently. It is good that this bug was noticed so quickly although it is unclear how it could have passed quality assurance control.
The last product in this review is the Intel X25-E. This SSD is smaller than the others at only 64 gigabytes but it is based on SLC memory. SLC chips have lower capacity and this model does not have large-capacity counterparts. We have already mentioned the highs of this memory type above: high speed and lower access time at writing and a much longer service life. MLC chips are specified to have 100,000 write cycles whereas SLC chips can last through as many as 1 million write cycles.
This SSD is declared to have a read speed of 250MBps and a write speed of 170MBps. The MLC-based Corsair P128 has a similar specified write speed, so we are going to see if the two SSDs based on different types of flash memory chips can equal each other at writing.
You might think that SSDs should be tested in the same way as HDDs, but that’s not exactly so. With HDDs, there is usually an excellent repeatability of test results. It is easy to minimize potential deviations: use the same or as similar as possible testbeds, warm HDDs up before running your tests, and clear the buffer memory up by means of a cold reboot prior to launching the most sensitive tests. That’s all. It is not that simple with SSDs.
The problem is in the controller’s algorithms that are responsible for leveling out the load among different memory cells. Besides, operations with SSD memory are carried out in blocks or pages even if small chunks of data are processed. If an SSD has been working long under load with a large share of random writes, data get highly “fragmented”. This is not a file fragmentation as on HDDs: a file may reside in sequential LBA addresses but they will belong to different memory chips and different pages in those chips.
One more performance-affecting factor is that today’s OSes when removing data only change the allocation table but do not send requests to clean up the specific LBA addresses. Thus, from the controller’s point of view, the SSD soon becomes fully filled up, all of its cells being occupied. The controller does not know that from the OS’s point of view there are no valuable data at most of addresses. The consequence is that the SSD’s performance degenerates greatly. And then the controller’s algorithms come into play once again. Having identified the type of load, a modern SSD controller tries to adjust data access and data placement in such a way as to achieve maximum performance. Easy to guess, it depends on the controller’s algorithms how long and how successful this process is going to be. For a hardware tester, it means that an SSD’s performance depends heavily on the type of previous load and on the time period between two subsequent tests.
How big can this discrepancy be in practice? Let’s see. First, we will show you the variance of test results in PCMark05 at five successive runs of the benchmark.
The Corsair SSD passed these tests after a long idle period (during which it could have carried out all planned optimizations) without a previous random write load and without storing any data. Easy to see, even under such comfortable conditions the benchmark provokes some changes in the data placement structure and calls for the controller’s optimization algorithms because there are as many as five different loads. As the consequence, the results of some iterations of the benchmark differ up to 10% from the average of the five iterations.
And what if we make things more complicated by benchmarking an SSD that already has some data which occupy some of its storage space and having previously tormented it with a generous portion of write requests in IOMeter?
Yes, the SSD feels worse now. The average result across the five tests is obviously lower now. You can see this even from the graphs, without looking at the numbers. And what is the most disturbing thing for a hardware tester, the variance of results between the different runs of the benchmark has changed. Some tests now deliver repeatable results, but the results of the file write test vary by 15% from the average. The results of the Windows XP startup test vary by over 30% even! The SSD’s speed is fluctuating, without showing any pattern.
Perhaps this is only a problem of the Corsair P128’s controller? Alas, all modern SSDs behave in the same way. But again, how big can this variance in performance be? Let’s check out the performance hit provoked by the storing of data and the preliminary load on two SSDs: a Corsair P128 and an 80GB Intel X25-M.
So, the performance decreases depending on the type of load and on the controller installed in the specific SSD. The value of the performance hit varies from negligible to serious (up to 46%)!
This diagram suggests that SSDs’ controllers might be compared according to the performance hit, but we won’t do such a comparison because the performance hit depends on the type of load and the history of previous requests. Our results won’t be repeatable and thus won’t be verifiable.
Well, Intel’s early SSDs indeed had problems with performance and the firmware update was meant to target this issue.
To sum up this section, we will show you one more interesting thing we have found. It is the performance hit of the SSDs in FC-Test:
In fact, there is no performance hit altogether! The SSDs even speed up somewhat. We suspect that the SSD’s controller takes the time between the tests (the reboot of the testbed plus two minutes of waiting) to perform an optimization. Anyway, this is one more example of the unpredictable behavior of modern SSDs that lowers our confidence in the repeatability of test results.
After all those problems we have discussed in the previous section, it is not so easy to find the best way to benchmark SSDs. Of course, we might go the easiest way and “defragment” (i.e. level out) the SSDs by means of sequential reading and writing to all of the SSD’s addresses (as the manufacturers suggest in the reviewer’s guide). But then we would have a refined and unrealistic result. Instead, we use the following method:
We guess this method will help compare the SSDS under identical conditions and yet produce real-life results.
The following testing utilities were used:
We installed the generic OS drivers for the drives and formatted them in FAT32 and NTFS as one partition with the default cluster size. For some tests 32GB partitions were created on the drives and formatted in FAT32 and NTFS with the default cluster size, too. In every test the drives were connected to the mainboard’s ICH7 controller.
We will compare only the four SSDs in IOMeter’s tests but will throw in a 1TB Western Digital Caviar Black for the sake of comparison in tests like FC-Test, PCMark and defragmentation which are more important for ordinary users. This HDD seems to us the best of desktop 7200rpm HDDs as it came out the winner of our comparative test session. We do not show you its IOMeter performance because it wouldn’t look good at all in comparison with the SSDs.
We will use WinBench 99 for low-level tests. That’s not much of testing, though. We will only take a look at pretty flat graphs and will learn the exact value of speed. This benchmark goes first and we “organize” the flash memory cells before it (by a stream of read requests as the SSD manufacturers suggest), so it is indicative of the maximum speed you can have with your SSD.
The Corsair and Intel X25-E are true to their specs, delivering almost exactly the same speed as specified for them. But neither of the X25-M models shows the promised speed in this test. The graphs suggest that the 160GB model was just not optimized at the time of the test. As opposed to the other models, its graph is not a straight line. The performance of the 80GB model is inexplicable, though.
From the low-level test we will now proceed to the synthetic IOMeter. IOMeter is sending a stream of read and write requests with a request queue depth of 4. The size of the requested data block is changed each minute, so that we could see the dependence of the drive’s sequential read/write speed on the size of the data block. This test is indicative of the maximum speed the drive can achieve.
The numeric data can be viewed in tables. We will discuss graphs and diagrams.
It is only on large data blocks that the SSDs reach their maximum performance. The top speeds are somewhat lower than what we have seen in WinBench 99. Frankly, we suspect that at such high speeds the disk controller’s bandwidth may be a bottleneck. Anyway, each SSD conquers the 150MBps peak while the leading 80GB Intel X25-M is even as fast as 200MBps. We can remind you that even 15,000rpm hard disk drives with SAS interface cannot deliver such a high speed. The Corsair is somewhat worse than the other SSDs, but not by much.
The three drives from Intel deliver their promised speeds at sequential writing, the SLC-based model even exceeding its specification of 170MBps. The Corsair does not make it to the specified 180MBps, yet its 165MBps is an outstanding result for inexpensive MLC memory, anyway. The Intel X25-E is faster irrespective of the data block size, but the difference is not worth the considerable difference in price between these two SSDs.
In this test IOMeter is sending a stream of requests to read and write 512-byte data blocks with a request queue of 1 for 10 minutes. The total number of requests processed by the drive is over 60 thousand, so we get a sustained response time that doesn’t depend on the drive’s buffer size.
It is all very simple here. The read access time is negligibly low with every SSD: the results of the best of HDDs are worse by an order or two! Intel’s SSD controllers deserve our praise for excellent writing: it is difficult to achieve as low an access time with flash memory at writing as at reading, but Intel manages that. The new chip from Samsung employed in the Corsair SSD is somewhat slower at writing, yet is better than any HDD. It is only the controllers of early SSDs that had an access time of tens of milliseconds.
Now we’ll see the dependence between the drives’ performance in random read and write modes on the size of the data block.
We will discuss the results in two ways. For small-size data chunks we will draw graphs showing the dependence of the amount of operations per second on the data chunk size. For large chunks we will compare performance depending on data-transfer rate in megabytes per second. This approach helps us evaluate the disk subsystem’s performance in two typical scenarios: working with small data chunks is typical for databases. The amount of operations per second is more important than sheer speed then. Working with large data blocks is nearly the same as working with small files, and the traditional measurement of speed in megabytes per second becomes more relevant.
Let’s start with reading.
IOMeter: Random Read, operations per second
The 160GB Intel X25-M is in the lead, followed by its 80GB series mate. The Corsair and the Intel X25-E are slower than the leaders, the SLC-based model proving to be the slowest.
Besides comparing the SSDs between each other, take note of the numbers. Thousands of operations per second! Ordinary desktop HDDs cannot deliver even a hundred operations per second in this test whereas the best HDDs, 15,000rpm models with SAS interface and fast heads, can barely notch 200. It is the performance at random-address operations, which is determined by the drive’s access time, is the trump of flash memory based storage. Being limited by the mechanism of the heads and the rotation speed of the platters, hard disk drives cannot compete on this point.
IOMeter: Random Read, megabytes per second
SSDs reach their maximum speed very soon on large data blocks. Again, the two Intel X25-M series models are in the lead whereas the SLC-based model is the slowest.
IOMeter: Random Write, operations per second
The results of this test are far from similar. The two models of the X25-M series perform brilliantly, boasting an excellent performance growth as the data block gets smaller. This is due to the controller and its algorithms, of course. Flash memory proper does not have any exceptional capabilities in terms of random-address writing.
Interestingly, the SLC-based Intel X25-E is overall somewhat slower than its opponents. It only has an advantage on rather large data blocks. The Corsair is poor in this test: Intel’s controller is superior as yet. On the other hand, an HDD would be even slower here. We can recall early SSDs again: they used to be inferior to HDDs in this test.
IOMeter: Random Write, megabytes per second
The SLC-based SSD is beyond competition on large data blocks. The Corsair is much slower than the MLC-based models from Intel until very large data blocks.
In the Database pattern the drive is processing a stream of requests to read and write 8KB random-address data blocks. The ratio of read to write requests is changing from 0% to 100% with a step of 10% throughout the test while the request queue depth varies from 1 to 256.
You can click this link to view the tabled results for IOMeter: Database pattern.
We will build diagrams for request queue depths of 1, 16 and 256.
We’ve got an interesting picture at the minimum queue depth. The Corsair is better than its opponents at high percentages of reads, but rolls back to last place at high percentages of writes. The Intel X25-E is just the opposite. Take note of the jagged graphs of all the SSDs from Intel: they are obviously trying to adapt for the load during this long test.
When the queue depth increases (although it is hard to imagine a situation when such a fast drive would need a requests queue), it is the Intel X25-E that goes ahead. This model is obviously up to its enterprise targeting and worth its high price. The Corsair takes last place.
There are no big changes when the queue gets longer except that the two X25-M models get closer to the leader.
Winding up this group of tests, we will show you diagrams for each SSD with graphs for five different request queue depths.
Frankly, we did not expect an SSD to deliver such a repeatable result. The behavior of the Corsair is easy to explain. It begins to reorder requests when there appears a request queue. This improves its read performance but worsens its write performance.
Although drawing different graphs, the two Intel X25-M models behave in a similar way. They are not very fast at short queue depths, but have very effective request reordering algorithms. As opposed to the Corsair, their speed of writing does not deteriorate at that.
The Intel X25-E behaves in a similar way. It wins more than the others from having a request queue and has an excellent performance boost at reading. This HDD obviously prefers high loads. It is an enterprise model indeed.
The drives are tested under loads typical of servers and workstations.
The names of the patterns are self-explanatory. The Workstation pattern is used with the full capacity of the drive as well as with a 32GB partition. The request queue is limited to 32 requests in the Workstation pattern.
The results are presented as performance ratings. For the File-Server and Web-Server patterns the performance rating is the average speed of the drive under every load. For the Workstation pattern we use the following formula:
Rating (Workstation) = Total I/O (queue=1)/1 + Total I/O (queue=2)/2 + Total I/O (queue=4)/4 + Total I/O (queue=8)/8 + Total I/O (queue=16)/16.
The picture is most unusual because HDDs behave completely different in this test. The Corsair delivers stable results at any queue depth. As a result, it is slower than the X25-M models at short queue depths but is ahead of them at long queue depths.
Intel’s SSDs all behave in the same way. They are brilliant at short queue depths but slow down at a queue depth of 128 requests. The 160GB model (with newer firmware) wins at short queue depths while the 80GB model, at long ones. The X25-E is surprisingly the worst of the three. These SSDs have some problems in this test. It must be some flaw in the controller’s operation.
The X25-M and Corsair have similar performance ratings whereas the X25-E has the slowest rating.
The picture is different when there are write requests in the load. The Corsair delivers stable performance but is slower than the opponents among which the SLC-based model is superior. Intel’s SSDs are far from stable. Their performance is constantly fluctuating.
The Workstation pattern is a more variegated load and it provokes dramatic changes in the behavior of the SSDs although there are no serious changes in terms of our performance ratings except that the 80GB X-25M gives way to the 160GB model. Particularly, Intel’s SSDs now deliver stable performance, although the 160GB model suffers a performance hit as the request queue gets longer. Corsair is still on the losing side. It is now clear that it prefers to work under low loads.
When the test zone is reduced to 32 gigabytes, every SSD, save for the Corsair, has performance fluctuations. The Corsair loses its speed at short queue depths while the X25-E is slower at long queue depths. As a result, the X25-M models take the two top places according to our performance rating.
The multithreaded tests simulate a situation when there are one to four clients accessing the virtual disk at the same time – the clients’ address zones do not overlap. We will discuss diagrams for a request queue of 1 as the most illustrative ones. When the queue is 2 or more requests long, the speed doesn’t depend much on the number of applications. You can also click the following links for the full results:
Multithreaded reading is one more test where SSDs are superior to HDDs because they do no have to move a read/write head between two or more access zones. As a result, each SSD slows down but slightly with the addition of a second thread, the Corsair being somewhat worse than the others here. And when there are even more threads to be processed, the SSDs all speed up.
The SSD controllers behave differently at multithreaded writing. Intel’s SSDs are indifferent to the number of write threads, delivering almost the same speed. The Corsair is excellent at one thread where it is but slightly slower than the X25-E, but suffers a threefold performance hit when processing multiple threads.
For this test two 32GB partitions are created on the SSD and formatted in NTFS and then in FAT32. A file-set is then created, read from the SSD, copied within the same partition and copied into another partition. The time taken to perform these operations is measured and the speed of the SSD is calculated. The Windows and Programs file-sets consist of a large number of small files whereas the other three patterns (ISO, MP3, and Install) include a few large files each.
We’d like to note that the copying test is indicative of the drive’s behavior under complex load. In fact, the SSD is working with two threads (one for reading and one for writing) when copying files.
This test produces too much data, so we will only discuss the results achieved with the Install, ISO and Programs file-sets in NTFS. You can use the following link to view the results in FAT32.
We have the same standings when writing any file-set. The X25-E is in the lead, followed by the Corsair, which in its turn is ahead of the MLC-based models from Intel. Comparing this to the HDD’s performance, we can see that modern SSDs are equal to or better than the best of modern HDDs in this test. The tested SSDs are all faster than the WD Caviar Black, which delivers average results in this test for an HDD.
The tested SSDs are two times as fast as the HDD irrespective of the file-set. The three models from Intel go neck and neck, the Corsair lagging somewhat behind. The 25MBps gap doesn’t look large considering that the SSDs are as fast as 150MBps and more.
There are no surprises at copying files. The X25-M is in the lead due to its higher write speed. Next goes the Corsair. The SSDs are all much better than the HDD. The Corsair rather fails on large files, though. Is it some flaw in its caching mechanisms?
PCMark 2005 has the same tests as the 2004 version (not only in names, but also in results as we have seen a lot of times), so we only use one test from PCMark 2004 which is not available in the 2005 version. It is called File Copying and measures the speed of copying some set of files. The other tests are:
The final result is the average of ten runs of each test.
We’ve got the same standings are in the copying subtest of FC-Test.
We see the same standings in the Windows XP startup test. Take note that the 160GB X25-M is somewhat better than its 80GB mate in quite a lot of tests. This must be due to the updated firmware. The SSDs are also much better than the HDD. They are five times as fast as the latter! In our HDD tests different models do not differ more than twofold.
Oddly enough, the X25-M with older firmware goes ahead and pushes the X25-E back to second place.
The standings change again in the General Usage test. The X25-E is on top, followed by the 80GB X25-M whereas the 160GB X25-M sinks to last place, giving way to the Corsair.
This scanning for viruses test is highly sensitive to caching mechanisms and SSDs are still better than the WD Caviar Black. And if you look up the results of the best HDDs in this test, you will see that the worst SSD is no worse than them! Interestingly, it is the 160GB Intel X25-M that is the worst SSD in this test: its series mate with older firmware passes this test faster.
Here is one more file writing test. And there are no surprises: the SLC-based Intel X25-E and the Corsair are ahead, delivering two times more speed than the others. The slower SSDs are as good as the HDD, though.
The Intel X25-E is expectedly first in terms of overall score. The X25-M (the 80GB model with older firmware) and the Corsair share second place. The SSDs are superior to the HDD: the worst SSD scores three times as many points as the Western Digital Caviar Black.
To make this part of our test session complete, we are going to run the latest version of PCMark called Vantage. Compared with the previous versions, the benchmark has become more up-to-date and advanced in its selection of subtests as well as Windows Vista orientation. Each subtest is run ten times and the results of the ten runs are averaged.
Here is a brief description of each subtest:
Basing on these subtests, the drive’s overall performance rating is calculated.
This test is a multithreaded load and SSDs cope with it nicely. This time we see them enjoy a fivefold advantage over the HDD. The Corsair is the leader now. It is somewhat faster than the Intel SSDs, including the X25-E which proves to be the slowest model in this test.
The SSDs all deliver similar and very high results. These drives are going to be appreciated by gamers who don’t like waiting for the next game level to load up.
The SSDs are also good at adding photographs into a media gallery. The Intel X25-M models are somewhat better than the others here. HDDs are generally fast in this test, yet the SSDs are almost three times as fast as one of the best HDDs available.
The 80GB Intel X25-M falls behind the other SSDs when booting Windows Vista up, yet it is still six times as fast as the best 7200rpm HDD!
The drives are all different at movie making. The X25-E wins this test, the 160GB X25-M is second and the Corsair is third. The X25-M is the slowest of the SSDs and comparable to the HDD.
This test is highly sensitive to the caching algorithms, too. And it is one of the few tests where the HDD can compete with the SSDs. It is actually second, being only slower than the X25-E. The 160GB X25-M is the best of the MLC-based SSDs.
The standings are familiar here: the X25-E is first, followed by the Corsair. Next go the two X25-M series models (the model with newer firmware is somewhat faster). The HDD is much slower.
It is different in the Application Loading test: the 160GB Intel X25-M goes ahead of its 80GB cousin while the Corsair slows down. The HDD is only one tenth as fast as the SSDs here.
The Intel X25-E has the highest overall score, but the other standings differ from those of the previous version of the benchmark. The 160GB X25-M is second while the Corsair is third. The X25-M with older firmware is but slightly better than the HDD.
Next goes our homemade test of defragmentation speed. We created a very defragmented file system on a 32GB partition of a disk by loading it with music, video, games and applications. Then we saved a per-sector copy of the disk and now copy it to the disk we want to test. The tested disk is connected to the mainboard’s SATA controller whose operation mode (AHCI/Standard SATA) is controlled from the mainboard’s BIOS. Next we run a script that evokes the console version of the Perfect Disk 8.0 defragmenter and marks the time of the beginning and end of the defragmentation process. Thus, each drive is tested twice – with AHCI support turned on and off on the controller. You can refer to this article for details about this test.
Strictly speaking, this test makes no practical sense for solid state drives because there is nothing to defragment on them. Every memory cell is equivalent to any other, so defragmentation won’t have any effect. Moreover, the location of a file in LBA sectors with sequential numbers won’t affect the physical location of data because the translator’s table can assign these addresses to completely different cells. So, the only effect you can achieve by defragmenting your SSD is to reduce its service life.
However, this test will allow us to compare how much time ordinary HDDs and SSDs spend to move the same amount of data.
This test is indicative of how firmware and memory type affect the speed of moving small data portions around the drive. The X25-E is excellent here. It passes the test within 10 minutes. The X25-M with new firmware is 2 minutes faster than its 80GB cousin. The Corsair has a less efficient controller and is but slightly better than the HDD in this test.
Now we are going to show you one more interesting test in which we use WinRAR version 3.8 to compress and then uncompress a 1.13GB folder with 8118 files in 671 subfolders. The files are documents and images in various formats. These operations are done on the tested drive. WinRAR is usually used to benchmark CPUs, but it can make a good test for HDDs/SSDs if you select the lowest compression level and use a huge amount of files. The drive’s performance should affect the speed of compressing/uncompressing then.
We have already tried this test with 2.5-inch HDDs and 3.5-inch HDDs and know what to expect.
The 160GB model from Intel is again better than its 80GB cousin with older firmware. The Corsair is surprisingly slow. The Intel SSDs are half a minute faster than desktop HDDs in this test whereas the Corsair took as long as 5400rpm HDDs to archive data. Perhaps it hadn’t had enough “rest” after the previous test.
The standings are absolutely different in the unzipping test. The X25-E is the best. It is just excellent at writing. HDDs generally take twice the time to complete the task. The 160GB X25-M is good, too. It coped with the task in 45 seconds. The Corsair is third, delivering the same performance as modern HDDs. The 80GB Intel X25-M is 50% slower than the Corsair – its firmware needs updating.
You can refer to our article called Hard Disk Drive Power Consumption Measurements: X-bit’s Methodology In depth for details on this test. We will just list the specific modes we measure the power consumption in:
Let’s check out each mode one by one.
A solid state drive has no platters to spin up, so their power consumption is very low from the very start, even if you compare them with 2.5-inch HDDs. The Corsair requires only two thirds of the current that the other SSDs need.
The Corsair is twice more economical when idle. On the other hand, all SSDs consume very little – less than 1 watt – in comparison with 2.5-inch 5400rpm HDDs.
The power consumption grows up when there are random requests to be performed. However, the SSDs all remain within 1 watt at random reading, the Corsair again being twice as economical as the Intel SSDs.
When we switch from reading to writing, the SSDs increase their power consumption because the controller is working more actively and the memory cells need more power. The Corsair is still the most economical drive, consuming a little more than 1.5 watts. The 80GB Intel X25-M needs about the same amount of juice. The other two models from Intel consume more than 2 watts but less than 2.5 watts. Of course, we don’t think that someone will use such high-performance SSDs over USB, but anyway.
If linear operations are substituted with sequential ones, the SSDs consume less at writing and more at reading.
It is good that solid state drives have made their way from early samples to mature market products. They have got rid from most of their childhood diseases that used to plague early models of flash-based drives such as low speed of sequential operations and low writing performance.
Of course, SSDs are not ideal. Their performance is rather unpredictable and their service life is not very long (it depends on how much writing they have to do). This may repel some customers. However, every SSD tested in this review can be characterized as a compact small-capacity storage device that is capable of delivering excellent speed (far better than that of hard disk drives) at low power consumption.
As soon as SSDs get rid of the two mentioned drawbacks and as soon as they become larger in terms of storage capacity and more affordable, they will have a good chance of ousting HDDs out of the market altogether. We guess users are going to like that. Computers will only be better if cooling fans are the only rotating thing in them. Unfortunately, we cannot expect a quick victory of SSDs over HDDs. The difference in the cost of storage is huge and is diminishing too slowly. So far, we can see the following uses for SSDs:
Of course, you can use your SSD for any other purpose, but you should be aware that this is not only a fast but also a very expensive storage which can also fail quickly from too much writing. If you replace the hard disk with an SSD in your home all-purpose computer, applications will be loaded and data will be copied faster, yet not instantaneously.
Finally, we’d like to say a few words about each product tested in this review.
The Intel X25-E is the fastest model because it is based on SLC rather than on MLC memory as the other tested SSDs. Its capacity is small at 64 gigabytes. That’s about the contents of two Blu-ray discs with movies. However, it is far more reliable and its writing speed is much higher than that of the other SSDs. Alas, its price is high, too. With all these characteristics and with the high speed of writing and excellent random reading it has, this SSD is going to be a perfect choice for enterprise disk subsystems.
The less expensive Intel X25-M with capacities of 80 and 160 gigabytes seem to be universal products. Like every other SSD, they deliver excellent speed of reading, both sequential and random. You should not expect them to be fast at sequential writing, though. They are merely comparable to HDDs in this respect. However, they do cope with random writing nicely, their controller boasting high efficiency. The only downside of these SSDs is that their performance is rather unpredictable, often depending on the type of previous load.
The Corsair P128, representing the latest generation of SSDs based on Samsung’s platform, differs from Intel’s X25-M series. It has a higher speed of sequential writing, which leads to faster processing of large files, but its random writing is slower. This SSD should be given credit for delivering very stable, if not always very high, performance. Looks like Samsung has developed a worthy alternative to Intel’s products.
Hopefully, we will see more products with other controllers very soon.