I feel a little silly saying this, but the Crucial M500 960GB and Samsung 840 EVO 1TB are two of the best values in the SSD market. Both drives sell for north of $500, which puts them firmly in high-end territory. However, they also have loads of storage. If you do the math, these terabyte flagships actually cost less per gigabyte than lower-capacity models with cheaper price tags. And they have higher performance ratings, too.
Considering the manufacturers, we probably shouldn’t be surprised that these drives are such relative bargains. Samsung is the largest producer of flash memory in the world, while Crucial is the consumer brand of memory giant Micron. Both firms have cutting-edge NAND fabrication facilities, and their SSD divisions are first in line for the latest chips. Those divisions no doubt get a nice family discount, too.
The question, of course, is which of these flash giants makes the best SSD? We’ve pitted the M500 960GB against the 840 EVO 1TB to find out. We’ve also included smaller variants from each lineup to provide a little extra flavor for folks with tighter budgets. Join us as we crown the king of the terabyte titans.
TLC vs. MLC flash
Since we’ve already covered lower-capacity versions of the Crucial M500 and Samsung 840 EVO in earlier reviews, we won’t rehash all the details here. I suggest reading those initial articles for a detailed look at each drive’s underlying architecture. Today, we’ll highlight a few key differences and similarities before moving on to the results of our exhaustive performance testing.
By far the biggest difference between the M500 and the 840 EVO is the flash memory under the hood. Like most consumer-grade SSDs, the M500 features MLC NAND with two bits per cell. The 840 EVO is based on TLC flash that squeezes an extra bit into every memory cell.
Flash cells represent data with a negative charge created by trapped electrons. Because there’s some unavoidable variation in the nanoscale characteristics of each cell, all data is verified after it’s written. The verification process takes longer with TLC NAND, which must differentiate between eight possible values between 000 and 111. MLC NAND only needs to worry about four values between 00 to 11.
Due to the more involved verification process, TLC NAND typically has slower write speeds than its MLC counterpart. However, the 840 EVO isn’t a typical TLC implementation. Although much of its flash has three bits per cell, a portion is addressed as server-style SLC NAND with only one bit per cell. This SLC zone totals 12GB and is reserved for TurboWrite, which caches incoming data to mitigate the slower write speeds associated with TLC flash. Cached data is held in the TurboWrite buffer until the EVO slips into an idle state, after which the data is passed along to main storage.
|Max 4KB random (IOps)||Price||$/GB|
840 EVO 250GB
840 EVO 500GB
840 EVO 1TB
Thanks to TurboWrite, the 840 EVO 1TB has a peak sequential write speed of 520MB/s. Writes proceed directly to the TLC flash if the TurboWrite buffer is full, but that only drops the EVO’s sequential speed rating to 420MB/s—20MB/s higher than the M500’s peak write rate. At least according to the manufacturers’ specifications, the 840 EVO is faster than the M500 not only in every category, but also across the multiple capacities we’ve tested. In a moment, we’ll see if that dynamic holds up in the wide variety of benchmarks that make up our storage test suite.
Testing performance is easy enough, but measuring endurance is too time-consuming for a comparison like this one. That’s unfortunate, because write endurance is an important consideration given the contenders. As data is written to NAND, the individual memory cells degrade, shrinking the range of voltages that can be used to represent data. This shrinkage is more problematic for TLC flash, which must differentiate between more values within that narrowing window.
Samsung doesn’t publish an official endurance specification for the 840 EVO, but Crucial says the M500 can withstand 40GB of writes per day for five years, or 72TB total. Based on what we’ve seen in our ongoing SSD Endurance Experiment, Crucial’s estimate is a conservative one. Thus far, we’ve written over 400TB to a collection of MLC-based drives, and they’re all in excellent shape. I can count the number of bad blocks on one hand.
Our experiment doesn’t include the M500 or the 840 EVO. However, it does feature Samsung’s first-generation TLC drive, the 840 Series. That SSD showed its first bad blocks after 100TB of writes, and it’s accumulated over 1300 of ’em through 400TB of writes. We also encountered unrecoverable errors while performing a data retention test at the 300TB mark.
TLC NAND clearly isn’t as robust as the MLC alternative. That said, hundreds of terabytes is an awful lot of writes for a consumer-grade SSD. The drive in my primary desktop writes only about 2TB per year, so it would take me half a century just to get to 100TB.
SSDs employ advanced signal processing and error correction algorithms to make the most of their NAND’s limited lifespan. Crucial also endows the M500 with an extra layer of protection inherited from Micron’s enterprise-class products. Dubbed RAIN, this RAID-like redundancy scheme uses a portion of the flash to house parity information. If the user’s data is compromised by a physical flash failure, it can be reconstructed with the parity information stored on the drive.
The flash reserved for parity is inaccessible to the user, which explains the M500’s lower capacity. The drive actually has a terabyte of flash onboard. RAIN and overprovisioned spare area each take a slice, leaving only 894GB for the user.
Even though the 840 EVO doesn’t have to worry about parity data, its TurboWrite cache monopolizes some of the flash. With the same 1TB total NAND capacity as the M500, the EVO delivers 932GB to the user. The 38GB difference works out to only 4%, but that’s still enough to store a lot of applications and games—or one blockbuster title with high-resolution textures.
Despite their differences, the Crucial M500 960GB and Samsung 840 EVO 1TB have quite a lot in common. Their flash is built on similar fabrication technologies, for example. The EVO’s flash is fabbed on a 19-nm process, while the M500’s is made with 20-nm technology.
Also, both drives have 128Gb (16GB) NAND dies that double the capacity of the 64Gb (8GB) dies commonly found in other SSDs. These 128Gb dies are key to reaching terabyte territory. They’re also a liability for the lower-capacity models. The M500 240GB and 840 EVO 250GB use only 16 dies each, but 32 dies are required to take full advantage of the internal parallelism in their respective controller chips. That’s why those models have lower write performance ratings.
With 64 dies apiece, the 960GB and 1TB variants have more than enough flash to saturate their controllers. The M500 uses a Marvell 88SS9187 chip, while the EVO sports a proprietary Samsung MEX design. Both controllers have eight parallel NAND channels.
Crucial deploys the M500’s NAND across 16 physical packages, while Samsung squeezes the same capacity into just eight. The 840 EVO’s circuit board is smaller as a result, though the cases are the same size. Each one measures 7 mm thick, allowing the drives to slip into slimmer notebook bays.
Speaking of notebooks, the 840 EVO 1TB is also available in mSATA form. Crucial doesn’t have a bite-sized version of the M500 960GB, though. The M500’s mSATA and M.2 flavors top out at 480GB.
When Samsung introduced the 840 EVO mSATA, it released new firmware for the 2.5″ version. This revision enables support for the TCG/Opal and IEEE 1667 encryption standards required by eDrive, Microsoft’s hardware-accelerated BitLocker implementation for Windows 8. The Crucial M500 has supported eDrive since its debut, so the EVO is a little late to the party. In fact, its eDrive firmware was originally scheduled for September. Better late than never, though.
Software may be a secondary concern for solid-state drives, but there’s a big difference between what comes with the M500 and 840 EVO. The M500 drive is a barebones affair. There’s no accompanying Windows utility, and cloning software is only available with an installation kit that’s sold separately.
Samsung’s Magician utility and data migration software can be downloaded by any 840 EVO owner. The cloning tool is designed specifically for migrating from larger mechanical drives; it allows some data to be transferred to secondary storage rather than the target SSD, which is almost always smaller than the source hard drive. Then there’s the Magician utility, which can download and apply firmware updates, optimize system settings, monitor drive health, and read SMART attributes, among other things.
I’m not sure I trust the Magician utility’s general health indicator—it’s been far too optimistic about the state of the 840 Series drive in our endurance experiment. However, I do like how the main interface displays the total number of bytes written. Users still have to dig into the SMART data to track the number of bad blocks, but at least that’s easily done with through the Magician utility.
Third-party software is required to get similar information out of the M500. Even then, it’s not that simple. Check out the SMART attributes exposed by Hard Disk Sentinel, the utility we’ve been using to monitor wear in our endurance experiment:
The reallocated sector count adds up the number of bad blocks. There are no attributes for tracking writes, though. One of those vendor-specific attributes can probably be translated to total bytes written, but I wouldn’t trust an unlabeled attribute, and I’m surprised that such important information is obfuscated in the first place.
TR reader Karol tipped me off about the fact that the M500 does tally the number of logical sectors written in its extended SMART device statistics. Those stats aren’t accessible with common utilities like HD Sentinel. However, they are available via Smartctl, a command-line tool included in the free Smartmontools package. Getting a bead on total writes shouldn’t be this difficult, but at least it’s possible.
Crucial should really develop a Windows utility with basic monitoring capabilities. An integrated firmware updater would be nice, too. Samsung’s Magician software displays a notification when new firmware is available, and it handles both downloading the update and installing it.
The Magician utility also has a feature called RAPID mode, which uses a slice of system memory as a high-speed drive cache. We took a closer look at RAPID mode in August, and the caching solution did improve performance in some benchmarks. However, it slowed the EVO in other tests, and we’re not crazy about writes being cached in volatile DRAM. Unless you’re looking to set benchmark records, we recommend keeping RAPID mode disabled. The EVO is fast enough without it.
Our testing methods
If you’re familiar with our testing methods and hardware, the rest of this page is filled with nerdy details you already know; feel free to skip ahead to the benchmark results. For the rest of you, we’ve summarized the essential characteristics of all the drives we’ve tested in the table below. In addition to the M500 960GB, we’ve tested 240 and 480GB models from the same family. We’ve also benched the 250 and 500GB 840 EVO alongside the 1TB variant.
|Corsair Force Series GT
|NA||SandForce SF-2281||25nm Intel sync MLC|
|Corsair Neutron 240GB||256MB||LAMD LM87800||25nm Micron sync MLC|
|Corsair Neutron GTX
|256MB||LAMD LM87800||26nm Toshiba Toggle MLC|
|Crucial M500 240GB||256MB||Marvell 88SS9187||20nm Micron sync MLC|
|Crucial M500 480GB||512MB||Marvell 88SS9187||20nm Micron sync MLC|
|Crucial M500 960GB||1GB||Marvell 88SS9187||20nm Micron sync MLC|
|Intel 335 Series 240GB||NA||SandForce SF-2281||20nm Intel sync MLC|
|Intel 520 Series 240GB||NA||SandForce SF-2281||25nm Intel sync MLC|
|OCZ Vector 256GB||512MB||Indilinx Barefoot 3||25nm Intel sync MLC|
|OCZ Vertex 4 256GB||512MB||Indilinx Everest 2||25nm Micron sync MLC|
|OCZ Vertex 450 256GB||512MB||Indilinx Barefoot 3 M10||20nm Intel sync MLC|
|SanDisk Extreme II 240GB||256MB||Marvell 88SS9187||19nm SanDisk Toggle SLC/MLC|
|Samsung 830 Series 256GB||256MB||Samsung MCX||27nm Samsung Toggle MLC|
|Samsung 840 Series 250GB||512MB||Samsung MDX||21nm Samsung Toggle TLC|
|Samsung 840 EVO 250GB||256MB||Samsung MEX||19nm Samsung Toggle
|Samsung 840 EVO 500GB||512MB||Samsung MEX||19nm Samsung Toggle
|Samsung 840 EVO 1TB||1GB||Samsung MEX||19nm Samsung Toggle
|Samsung 840 Pro 256GB||512MB||Samsung MDX||21nm Samsung Toggle MLC|
|Seagate 600 SSD 240GB||256MB||LAMD LM87800||19nm Toshiba Toggle MLC|
|WD Caviar Black 1TB||64MB||NA||NA|
Apart from the drives sharing the spotlight, our collection of results includes some of the most popular SSDs around. You can find full reviews of most of the drives in our storage section.
The solid-state crowd is augmented by a couple of mechanical contenders. From the old-school hard drive camp, we have WD’s Caviar Black 1TB. We’ve included Seagate’s Desktop SSHD 2TB, as well. The SSHD combines mechanical platters with 8GB of flash cache, but like the Caviar Black, it’s really not a direct competitor to the SSDs. Those drives are meant to provide additional context for our SSD results.
If you’ve made it this far, you’re probably the sort of detail-oriented person who appreciates naked circuit board shots. So, here are a couple of the M500 960GB and 840 EVO 1TB. You can see even larger versions of these and other images from the article in the image gallery at the bottom of the page.
We used the following system configuration for testing:
Core i5-2500K 3.3GHz
Asus P8P67 Deluxe
|Platform hub||Intel P67
|Platform drivers||INF update
|Memory size||8GB (2
Corsair Vengeance DDR3 SDRAM at 1333MHz
ALC892 with 2.62 drivers
Asus EAH6670/DIS/1GD5 1GB with Catalyst 11.7 drivers
Desktop SSHD 2TB with CC43 firmware
WD Caviar Black 1TB with 05.01D05 firmware
Corsair Force Series GT 240GB with 1.3.2 firmware
Corsair Neutron 240GB with M206 firmware
Corsair Neutron GTX 240GB with M206 firmware
Crucial M500 240GB with MU03 firmware
Crucial M500 480GB with MU03 firmware
Crucial M500 960GB with MU03 firmware
Intel 335 Series 240GB with 335s firmware
Intel 520 Series 240GB with 400i firmware
OCZ Vector 256GB with 10200000 firmware
OCZ Vector 150 256GB with 1.1 firmware
OCZ Vertex 450 256GB with 1.0 firmware
SanDisk Extreme II 240GB with R1131
Samsung 830 Series 256GB with CXM03B1Q firmware
Samsung 840 Series 250GB with DXT07B0Q firmware
Samsung 840 EVO 250GB with EXT0AB0Q firmware
Samsung 840 EVO 500GB with EXT0AB0Q firmware
Samsung 840 EVO 1TB with EXT0AB0Q firmware
Samsung 840 Pro Series 256GB with DXM04B0Q firmware
Seagate 600 SSD 240GB with B660 firmware
|Power supply||Corsair Professional Series Gold AX650W|
|OS||Windows 7 Ultimate x64|
Thanks to Asus for providing the systems’ motherboards and graphics cards, Intel for the CPUs, Corsair for the memory and PSUs, Thermaltake for the CPU coolers, and Western Digital for the Caviar Black 1TB system drives.
We used the following versions of our test applications:
- Intel IOMeter 1.1.0 RC1
- HD Tune 4.61
- TR DriveBench 1.0
- TR DriveBench 2.0
- TR FileBench 0.2
- Qt SDK 2010.05
- MinGW GCC 4.4.0
- Duke Nukem Forever
- Portal 2
Some further notes on our test methods:
- To ensure consistent and repeatable results, the SSDs were secure-erased before almost every component of our test suite. Some of our tests then put the SSDs into a used state before the workload begins, which better exposes each drive’s long-term performance characteristics. In other tests, like DriveBench and FileBench, we induce a used state before testing. In all cases, the SSDs were in the same state before each test, ensuring an even playing field. The performance of mechanical hard drives is much more consistent between factory fresh and used states, so we skipped wiping the HDDs before each test—mechanical drives take forever to secure erase.
- We run all our tests at least three times and report the median of the results. We’ve found IOMeter performance can fall off with SSDs after the first couple of runs, so we use five runs for solid-state drives and throw out the first two.
- Steps have been taken to ensure that Sandy Bridge’s power-saving features don’t taint any of our results. All of the CPU’s low-power states have been disabled, effectively pegging the 2500K at 3.3GHz. Transitioning in and out of different power states can affect the performance of storage benchmarks, especially when dealing with short burst transfers.
The test systems’ Windows desktop was set at 1280×1024 in 32-bit color at a 75Hz screen refresh rate. Most of the tests and methods we employed are publicly available and reproducible. If you have questions about our methods, hit our forums to talk with us about them.
HD Tune — Transfer rates
HD Tune lets us present transfer rates in a couple of different ways. Using the benchmark’s “full test” setting gives us a good look at performance across the entire drive rather than extrapolating based on a handful of sample points. The full test gives us fodder for line graphs, which we’ve split up by drive maker. You can click the buttons below each line graph to see how the M500 960GB and 840 EVO 1TB compare to their rivals.
To make the bar graphs easier to read, we’ve highlighted the M500 and 840 EVO families in different colors. We’ve also marked the terabyte-class models with brighter shades.
The 840 EVO 1TB gets off to a good start, just edging out the M500 960GB in HD Tune’s sequential read speed test. The difference between the two drives is slim, and the rest of the M500 and 840 EVO models deliver comparable performance.
HD Tune’s sequential write speed test produces some funky results with the M500. Flipping through the line graphs reveals similar behavior from several of the other SSDs, and it seems to be unique to this particular test. The M500’s average write speed suffers as a result, causing the drive to fall well behind the 840 EVO.
Check out the line graph for the Samsung SSDs, and pay particular attention to the far left side of the plot. All the EVOs exhibit an early surge in write speeds thanks to their TurboWrite caches. The 1TB model has the largest buffer of the bunch, so its surge lasts the longest.
Also, note how the 840 EVO 250GB’s write speed falls dramatically after the initial TurboWrite boost. Remember what I said about 128Gb NAND dies being a liability for lower-capacity SSDs? That’s the first evidence.
HD Tune runs on unpartitioned drives, with no file system in place, which probably explains the write-rate spikes exhibited by some of the SSDs. For another take on sequential speed, we’ll turn to CrystalDiskMark, which runs on partitioned drives. We used the benchmark’s sequential test with the default 1GB transfer size and randomized data.
Here, the 840 EVO family has a bigger edge in sequential read performance. It’s faster than the M500 with writes, too, but by a much smaller margin than in HD Tune.
The M500 960GB and 840 EVO 1TB largely match the performance of the other models in their respective families. However, the M500 240GB has a much slower write speed than not only its siblings, but also the 840 EVO 250GB. The EVO’s TurboWrite cache probably deserves some of the credit for keeping the 250GB model competitive.
HD Tune — Random access times
In addition to letting us test transfer rates, HD Tune can measure random access times. We’ve tested with four transfer sizes and presented all the results in a series of line graphs. We’ve also busted out the 4KB and 1MB transfers sizes into bar graphs that should be easier to read without the presence of the mechanical drives throwing off the scale.
As before, click the buttons below the line graphs to compare the M500 960GB and 840 EVO 1TB to different groups of drives.
The SSDs all have similar random read access times. The M500 is a smidgen quicker in the 4KB test, but the difference is only four microseconds. I wouldn’t worry about it.
There is, of course, a much bigger difference between the access times of the SSDs and our lone hard drive. The Caviar Black is at least an order of magnitude slower, if not more. Seagate’s Desktop SSHD hybrid is much more competitive, but it can’t keep up with the SSDs as the transfer size increases.
Although the SSDs are closely matched in HD Tune’s 4KB random write test, the field spreads out in the 1MB test. There, the 840 EVO 1TB has a modest lead over the M500 960GB. All the EVOs perform almost identically, but the M500 240GB lags well behind its higher-capacity siblings.
Again, the largest deltas are between the solid-state and mechanical drives. The Caviar Black fares a little better than it did with reads, but the Desktop SSHD falls behind the SSDs even with the smaller transfer sizes.
TR FileBench — Real-world copy speeds
Concocted by resident developer Bruno “morphine” Ferreira, FileBench runs through a series of file copy operations using Windows 7’s xcopy command. Using xcopy produces nearly identical copy speeds to dragging and dropping files using the Windows GUI, so our results should be representative of typical real-world performance. We tested using the following five file sets—note the differences in average file sizes and their compressibility. We evaluated the compressibility of each file set by comparing its size before and after being run through 7-Zip’s “ultra” compression scheme.
|Number of files||Average file size||Total size||Compressibility|
The names of most of the file sets are self-explanatory. The Mozilla set is made up of all the files necessary to compile the browser, while the TR set includes years worth of the images, HTML files, and spreadsheets behind my reviews. Those two sets contain much larger numbers of smaller files than the other three. They’re also the most amenable to compression.
To get a sense of how aggressively each SSD reclaims flash pages tagged by the TRIM command, we run FileBench with the solid-state drives in two states. We first test the SSDs in a fresh state after a secure erase. They’re then subjected to a 30-minute IOMeter workload, generating a tortured used state ahead of another batch of copy tests. Let’s start with the fresh-state results.
In each and every test, the 840 EVO 1TB comes out ahead of the M500 960GB. The Samsung drive only has a narrow lead in the Mozilla and TR tests, but the gap widens when the drives are copying larger files.
Interestingly, the 840 EVO 1TB is slower than the 500GB model in several tests. The 250GB model is slower still, but it doesn’t lag behind its brethren as much as the M500 240GB. Of course, even the low-capacity Crucial offering is much faster than the mechanical drives. It also has an edge over the Samsung 840 Series, which was replaced by the 840 EVO.
The M500 960GB and 840 EVO 1TB perform similarly in our simulated used state. It’s been a while since we’ve seen an SSD turn in substantially slower used-state copy speeds, but the last time was with the 840 Pro, and other Samsung SSDs have exhibited problems in the past. Fortunately, there are no issues with the EVO.
TR DriveBench 1.0 — Disk-intensive multitasking
TR DriveBench allows us to record the individual IO requests associated with a Windows session and then play those results back as fast as possible on different drives. We’ve used this app to create a set of multitasking workloads that combine common desktop tasks with disk-intensive background operations like compiling code, copying files, downloading via BitTorrent, transcoding video, and scanning for viruses. The individual workloads are explained in more detail here.
Below, you’ll find an overall average followed by scores for each of our individual workloads. The overall score is an average of the mean performance score for each multitasking workload.
Though its margin of victory is narrow, the 840 EVO 1TB comes out just ahead of the M500 960GB in DriveBench 1.0 overall. Let’s see what we can learn from the individual test results.
The lower-capacity M500 and 840 EVO models seem to struggle a little in the file copy test, but they’re not that much slower than their terabyte-class kin.
Once again, the hybrid and hard drive are both decimated by the solid-state field.
TR DriveBench 2.0 — More disk-intensive multitasking
As much as we like DriveBench 1.0’s individual workloads, the traces cover only slices of disk activity. Because we fire the recorded I/Os at the disks as fast as possible, solid-state drives also have no downtime during which to engage background garbage collection or other optimization algorithms. DriveBench 2.0 addresses both of those issues with a much larger trace that spans two weeks of typical desktop activity peppered with multitasking loads similar to those in DriveBench 1.0. We’ve also adjusted our testing methods to give solid-state drives enough idle time to tidy up after themselves. More details on DriveBench 2.0 are available on this page of our last major SSD round-up.
Instead of looking at a raw IOps rate, we’re going to switch gears and explore service times—the amount of time it takes drives to complete an I/O request. We’ll start with an overall mean service time before slicing and dicing the results.
The M500 960GB and 840 EVO 1TB may not have the lowest mean service times in DriveBench 2.0, but they’re pretty close to each other, and they’re not too far off the fastest SSDs we’ve tested. The 480 and 500GB variants are a little slower, while the 240 and 250GB models lag further behind. We can look at the read and write service times separately to get a better sense of what’s going on.
Interesting. The M500 960GB has a slight advantage over the 840 EVO 1TB with writes, but it’s a little slower with reads.
The read results are pretty close both within the M500 and 840 EVO families and for the SSDs at large. Writes are a different story, however. The field is more spread out, the lower-capacity M500 and 840 EVO models have higher mean service times, and the M500 240GB even trails the mechanical drives.
Note that the M500 240GB and 840 EVO 250GB have much slower write service times than the other SSDs, most of which have similar capacities. Those other drives sport smaller 64Gb NAND dies, so they can take advantage of more controller-level parallelism at lower capacities. With 128Gb dies, the M500 and 840 EVO use half as many memory chips as their peers.
There are millions of I/O requests in this trace, so we can’t easily graph service times to look at the variance. However, our analysis tools do report the standard deviation, which gives us a sense of how much service times vary from the mean.
Greater variance isn’t necessarily a problem if service times are low overall, as they are with DriveBench read requests. However, the M500 240GB has longer write service times and more variance among them. That isn’t an attractive combo, especially since the 840 EVO 250GB fares better according to both metrics.
The M500 960GB has more variance in its write performance than the 840 EVO 1TB, but since the mean service times are low for both drives, I wouldn’t worry too much about the discrepancy.
We can’t easily graph all the service times recorded by DriveBench 2.0, but we can sort them. The graphs below plot the percentage of service times that fall below various thresholds. Once again, the buttons below each graph switch between different sets of drives.
The mechanical drives are the only ones that deviate substantially from the curves followed by the SSDs. They have fewer service times under each threshold, at least until we hit 100 milliseconds.
We’re particularly interested in service times beyond 100 ms, since those could be long enough for users to notice. The following graphs tally the number of extremely long service times exhibited by each drive in DriveBench 2.0.
The 840 EVO 1TB has fewer 100+ ms service times than the M500 960GB with both reads and writes. In fact, all of the EVOs come out ahead of their M500 competition. The most worrying results are produced by the M500 240GB, which has nearly 15 times more extremely long service times than its 960GB counterpart. Ouch.
Our IOMeter workloads feature a ramping number of concurrent I/O requests. Most desktop systems will only have a few requests in flight at any given time (87% of DriveBench 2.0 requests have a queue depth of four or less). We’ve extended our scaling up to 32 concurrent requests to reach the depth of the Native Command Queuing pipeline associated with the Serial ATA specification. Ramping up the number of requests also gives us a sense of how the drives might perform in more demanding enterprise environments.
We run our IOMeter tests using the fully randomized data pattern, which presents a particular challenge for SandForce’s write compression scheme. We’d rather measure SSD performance in this worst-case scenario than using easily compressible data.
There’s too much data to show clearly on a single graph for each access pattern, so we’ve once again split the results by drive maker. You can compare the performance of the M500 960GB and 840 EVO 1TB to that of the competition by clicking the buttons below each graph.
The web server access pattern is comprised entirely of read requests, and the 840 EVO 1TB has a clear advantage over the M500 960GB. The various M500 and 840 EVO capacities perform similarly here. While the EVOs have some of the highest I/O rates of the bunch, the M500s are less competitive, especially under heavier loads.
Our remaining IOMeter tests mix reads and writes, and this time, the M500 960GB trumps the 840 EVO 1TB—at least until the load really ramps up. The M500 hits its stride earlier and starts to plateau after eight concurrent I/O requests. The EVO starts slower, but its performance continues to scale through 32 simultaneous requests.
As one might expect given the results we’ve seen thus far, the lower-capacity M500 and 840 EVO drives are slower than the flagship models. The differences are relatively small compared to the gaps between those families and the fastest SSDs, though. The LAMD-based Corsair Neutron and Seagate 600 SSDs offer much better performance in these tests, and OCZ’s new Vector 150 delivers even higher I/O rates. You can blame the Vector for throwing off the scale for all the other drives.
Before timing a couple of real-world applications, we first have to load the OS. We can measure how long that takes by checking the Windows 7 boot duration using the operating system’s performance-monitoring tools. This is actually the first test in which we’re booting Windows off each drive; up until this point, our testing has been hosted by an OS housed on a separate system drive.
Level load times
Modern games lack built-in timing tests to measure level loads, so we busted out a stopwatch with a couple of reasonably recent titles.
No more than a second separates the fastest SSDs from the slowest ones in our load time tests. The 840 EVO 1TB technically comes out ahead of the M500 960GB, but the difference between those contenders is less than half a second. Yeah, you’re not going to notice that.
You will, however, notice the difference in load times between a solid-state and mechanical drive. The Caviar Black is much slower than all of the SSDs. The Desktop SSHD almost eliminates the gap, but its hybrid design can only accelerate access to data that users load frequently. SSDs provide wicked-fast load times regardless.
We tested power consumption under load with IOMeter’s workstation access pattern chewing through 32 concurrent I/O requests. Idle power consumption was probed one minute after processing Windows 7’s idle tasks on an empty desktop.
The M500 960GB sips a little less power than the 840 EVO 1TB at idle but slightly more under load. The lower-capacity drives in each family are more power-efficient, but the deltas amount to only one watt or less. Considering their capacities, the M500 960GB and 840 EVO 1TB draw relatively little power.
We’ve waded through a lot of performance data, and we’ll indulge a couple more graphs before weighing in with our final verdict. The following scatter plots use an overall performance score derived by comparing how each drive stacks up against a common baseline. This score is based on a subset of our performance data described here, and we’ve mashed it up with per-gigabyte prices from Neweegg. The best solutions are found in the upper left corner of the plot, which signifies high performance and low pricing.
Solid-state and mechanical storage have vastly different performance and pricing, and those disparities make the main plot a little busy. Click the buttons below the plot to switch between all the drives and a cropped look at just the SSDs—and keep in mind that we’ve trimmed the axes for the SSD-only plot.
The Crucial M500 960GB sells for $530, while the Samsung 840 EVO 1TB rings in at $570. Those prices put the drives on nearly equal footing on the price-per-gigabyte axis, but the EVO is faster overall. The EVO also maintains its performance advantage at lower capacities, and its 250GB incarnation is way ahead of the M500 240GB.
Zooming in on the SSDs reveals a lot of drives with comparable performance to the terabyte titans. Notice that none of those alternatives has a lower cost per gigabyte, making the flagship drives relative bargains.
There are considerations beyond where these drives fall on our value plots, of course. The M500 has built-in redundancy, and its two-bit MLC NAND should be more durable than the three-bit TLC flash in the 840 EVO. As we’ve learned in our endurance experiment, though, Samsung’s first-gen TLC drive can withstand hundreds of terabytes of writes without failure. The 840 EVO should last for years under typical client workloads.
Samsung bolsters the EVO with a strong software suite that includes a handy cloning tool and Windows utility. Those might seem like small additions, but they go a long way toward making the EVO feel like a complete package. Crucial doesn’t bundle any software with the M500, and the firm makes monitoring more difficult by masking the drive’s SMART attributes.
If you’re looking for a king-sized SSD to put inside a typical desktop or notebook, the Samsung 840 EVO 1TB is the better all-around option. It performs better overall, offers a slightly higher capacity, and comes with software that improves the user experience. The M500 960GB is an excellent drive, too, and it has my enthusiastic recommendation. But the M500’s strengths seem better suited to write-heavy workloads and entry-level enterprise applications. The 840 EVO 1TB is the more cohesive consumer product, and that’s why it’s our Editor’s Choice.