Western Digital’s RE2-GP hard drive

When we looked at Western Digital’s “GreenPower” Caviar GP last month, we heralded the drive’s low noise levels and frugal power consumption. Given the GP’s slower spindle speed and energy-efficient billing, those results were largely expected. What surprised us, however, was how well the GreenPower Caviar handled demanding multi-user loads. It’s not often that you find low power consumption with strong performance riding shotgun.

That combination makes the GreenPower drives particularly enticing for enterprise environments looking to maximize storage capacity within a limited power budget. Don’t think that potential went unnoticed at Western Digital. The company has been massaging the GP to better suit enterprise applications, and the result of those efforts is the new RE2-GP.

As one might expect, the RE2-GP has much in common with the Caviar GP. However, the RE2 packs a few additional perks, including a longer warranty, optimization for RAID environments, and firmware tweaks that aim to deliver better performance with enterprise-class applications. Read on to see how those firmware tweaks affect the GreenPower’s performance and why this drive might be a slam-dunk for enterprise storage.

Tweaked for enterprise
“IntelliPower” lies at the core of the GreenPower line, a marketing name that signifies a delicate balance of transfer rate, cache size, and spindle speed that ensures low power consumption along with what Western Digital calls “solid” performance. Two of the RE2-GP’s IntelliPower key specs are easy to pin down: the drive has 16MB of cache, and its spec sheet boasts sustained transfer rates of 672Mbps, or 84MB/s. However, the third and arguably most important element—spindle speed—remains shrouded in mystery.

Western Digital won’t reveal the exact spindle speed of its GreenPower drives, saying only that they run somewhere between 5,400 and 7,200 RPM. Spindle speeds vary depending on capacities, too, although all GP drives at a given capacity run at the same speed. We’ve also confirmed that RE2-GP drives share the same spindle speeds as their like-capacity Caviar GP counterparts.

Spindle speed factors heavily in overall drive performance, so it’s disappointing the GreenPower spec sheets aren’t more forthcoming. That’s not the only factor in the performance equation, though. Areal density also plays a major role, and in that respect the RE2-GP is well-equipped, with the same 250GB platters as the Caviar GP. These relatively high density platters allow the drive head to access data over shorter physical distances, helping to offset the GP’s slower spindle speed.

We’ll see just how well a high platter density can mask the performance impact of a lower spindle speed in a moment, but first, it’s important to highlight why Western Digital opted to turn down the GreenPower’s spindle speed in the first place. As one might expect, slower spindle speeds require less juice, giving GreenPowers lower rated power consumption than any other terabyte drives on the market. This reduction in power consumption is a nice perk for the Caviar GP, but it’s an even bigger boon to the RE2-GP, which is likely to be deployed in multi-drive environments where power savings can really add up. In fact, Western Digital is so keen to push the RE2-GP’s power efficiency in multi-drive environments that it’s cooked up a web-based power consumption calculator to help prospective customers quantify the potential savings.

Despite being targeted at very different environments, the Caviar GP and RE2-GP are actually mechanically identical. The drives share similar feature sets, too, including support for an IntelliPark mechanism that moves the drive heads off the media at idle to reduce aerodynamic drag. WD’s IntelliSeek just-in-time actuator delivery scheme is also under the hood, taking advantage of the rotational latency inherent to spinning media by only moving the drive head as fast as necessary—rather than as fast as possible—to get it into position for the next data point. WD also deploys a StableTrac motor in the 750GB and 1TB flavors of the RE2-GP, which secures its bad self shaft at both ends rather than just one.

The RE2-GP comes with a few features missing from the Caviar GP. These RAID-specific optimizations include Rotary Acceleration Feed Forward (RAFF), which detects and compensates for the ambient vibrations typical of multi-drive environments, and Time-Limited Error Recovery (TLER), which restricts the amount of time a drive will attempt solo error recovery, lest it be dropped from a RAID array for being unresponsive for too long. Heroic as they may be, prolonged error recovery attempts are unnecessary in RAID environments, since RAID controllers prefer to handle error recovery on their own.

Western Digital also provides additional validation testing for its enterprise-class drives, and although the RE2-GP isn’t necessarily more reliable than its desktop counterpart, the drive benefits from two more years of warranty coverage. That won’t guarantee the safety of your data for an additional two years, but with the RE2-GP, you should have a RAID array ensuring redundancy anyway.

Test notes
We’ll be comparing the performance of the RE2-GP with that of a slew of competitors, including some of the latest and greatest Serial ATA drives from Hitachi, Maxtor, Samsung, Seagate, and Western Digital. These drives differ when it comes to external transfer rates, spindle speeds, cache sizes, platter densities, NCQ support, and capacity, all of which can have an impact on performance. Keep in mind the following differences as we move through our benchmarks:

Note that the 250GB Caviar SE16 and the Raptor WD740GD lack support for Native Command Queuing. The WD740GD does support a form of command queuing known as Tagged Command Queuing (TCQ), but host controller and chipset support for TCQ is pretty thin. Our Intel 955X-based test platform doesn’t support TCQ.

We have test results from several versions of Western Digital’s Caviar SE16 and RE2. To avoid confusion, we’ll be listing their capacities in parentheses in each of our graphs.

Since Seagate makes versions of the 7200.7 both with and without NCQ support, the 7200.7 in our tests appears as the “Barracuda 7200.7 NCQ” to clarify that it’s the NCQ version of the drive. The other drives aren’t explicitly labeled as NCQ drives because they’re not available without NCQ support.

Finally, we should note that our WD1500ADFD has a slightly newer firmware revision than the Raptor X sample we’ve had since February, 2006. The drives still share identical internals, but firmware optimizations could give our newer Raptor an edge over the X in some tests.

Performance data from such a daunting collection of drives can make our graphs a little hard to read, so I’ve highlighted the RE2-GP in bright yellow and its high-capacity competitors—the Barracuda 7200.10, 7200.11, and ES, the Deskstar 7K1000, the Caviar GP, and the Caviar SE16 and RE2 750GB—in pale yellow to set them apart from the others. Since the RE2-GP is targeted at enterprise applications, I’ve highlighted the Raptor WD1500ADFD in pale yellow to show how its performance compares, as well.

We also have two sets of IOMeter graphs: one with all the drives, and another with just the RE2-GP and its direct rivals. Most of our analysis will be limited to how the RE2-GP compares with its direct rivals, so it should be easy to follow along.

Our testing methods
All tests were run three times, and their results were averaged, using the following test system.

Thanks to the folks at Newegg for hooking us up with the DiamondMax 11 we used for testing. Also, thanks to NCIX for getting us the Deskstar 7K1000.

Our test system was powered by OCZ PowerStream power supply units. The PowerStream was one of our Editor’s Choice winners in our last PSU round-up.

We used the following versions of our test applications:

• WorldBench 5.0
• Intel IOMeter v2004.07.30
• Xbit Labs File Copy Test v1.0 beta 13
• TCD Labs HD Tach v3.01
• Far Cry v1.3
• DOOM 3
• Intel iPEAK Storage Performance Toolkit 3.0

The test systems’ Windows desktop was set at 1280×1024 in 32-bit color at an 85Hz screen refresh rate. Vertical refresh sync (vsync) was disabled for all tests.

All 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.

WorldBench overall performance
WorldBench uses scripting to step through a series of tasks in common Windows applications. It then produces an overall score. WorldBench also spits out individual results for its component application tests, allowing us to compare performance in each. We’ll look at the overall score, and then we’ll show individual application results.

WorldBench better simulates typical desktop tasks than enterprise workloads, but the RE2-GP scores higher than the Caviar GP, if only by one point. Still, that puts the enterprise GreenPower at the head of the class, ahead of every other terabyte drive we’ve tested.

Multimedia editing and encoding

MusicMatch Jukebox

Windows Media Encoder

Adobe Premiere

VideoWave Movie Creator

With the exception of Premiere, scores are close throughout WorldBench’s multimedia editing and encoding tests. In Adobe’s video editing app, the RE2-GP fares reasonably well, besting terabyte drives from Seagate and Hitachi in addition to the Caviar GP. Western Digital’s standard RE2, with its full 7,200-RPM spindle speed, is still faster, though.

Image processing

Adobe Photoshop

ACDSee PowerPack

Photoshop doesn’t seem to care for faster drives, but ACDSee takes advantage better of the RE2-GP. Here, the drive only just falls behind the 7,200-RPM Caviar SE16 and RE2, and it’s a good ten seconds faster than the Caviar GP.

Multitasking and office applications

Microsoft Office

Mozilla

Mozilla and Windows Media Encoder

There isn’t much to see through WorldBench’s office and multitasking tests.

Other applications

WinZip

Nero

WinZip and Nero are different stories altogether, though. In the former, the RE2-GP ties its desktop counterpart and manages to squeak ahead of the terabyte Barracuda and Deskstar. Nero provides more fertile ground for the RE2-GP, which returns to third place behind the Caviar SE16 and RE2. That puts the drive a good 23 seconds ahead of the Caviar GP and faster than all the other terabyte offerings in the pack.

Boot and load times
To test system boot and game level load times, we busted out our trusty stopwatch. Each of these tests was run five times to ensure consistent results.

Our test system takes its sweet time booting with terabyte drives under the hood, and the RE2-GP can’t escape this trend. The drive is only marginally slower than the Caviar GP here, but all the terabyte drives are well behind their closest competitors.

The results are more mixed for game level load times. In Far Cry, the RE2-GP falls to the back of the field and is more than three seconds slower than the Caviar GP. Things don’t get much better in Doom 3, where the RE2-GP lands in the middle of the field. There, it’s still slower than the terabyte Barracuda and Deskstar, but at least a little quicker than the Caviar GP.

File Copy Test
File Copy Test is a pseudo-real-world benchmark that times how long it takes to create, read, and copy files in various test patterns. File copying is tested twice: once with the source and target on the same partition, and once with the target on a separate partition. Scores are presented in MB/s.

To make things easier to read, we’ve busted out our FC-Test results into individual graphs for each test pattern. We’ll tackle file creation performance first.

Areal density can only do so much to offset the RE2-GP’s spindle speed handicap, so it’s no surprise to see the drive struggling against competitors spinning at a full 7,200 RPM. What’s more interesting here is how the RE2-GP’s performance compares with that of its Caviar cousin. With the Install, ISO, and MP3 test patterns, the RE2-GP is actually slower than the Caviar GP. However, with the Programs and Windows test patterns, which feature a greater number of smaller files, the RE2 pulls out ahead.

The RE2-GP actually does pretty well, all things considered. Despite its spindle speed disadvantage, it’s more often than not faster than the Deskstar 7K1000 and Barracuda 7200.11.

FC-Test’s, er, read tests paint a different picture for the RE2-GP, which easily beats the Caviar GP across all five test patterns. That’s an impressive result considering the drives share identical hardware. However, it doesn’t improve the RE2-GP’s performance relative to its terabyte competition. When confined to read operations, the Deskstar 7K1000 proves faster with each test pattern.

FC-Test – continued
Next, File Copy Test combines read and write tasks with some, er, copy tests.

Copy tests combine read and write operations, and that seems to suit the RE2-GP just fine. The drive is faster than its Caviar counterpart through teach test pattern, and it’s usually quicker than the Deskstar 7K1000. However, Seagate’s terabyte Barracuda does make up some ground here, ensuring that GreenPower doesn’t dominate the terabyte lineup.

FC-Test’s second wave of copy tests involves copying files from one partition to another on the same drive.

The RE2-GP continues to fare better than the Caviar through our partition copy tests. It only comes out at the head of the terabyte field with one test pattern, but considering its slower spindle speed and the fact that the 7K1000 and 7200.11 have twice the cache, that’s still pretty impressive.

iPEAK multitasking
We’ve developed a series of disk-intensive multitasking tests to highlight the impact of command queuing on hard drive performance. You can get the low-down on these iPEAK-based tests here. The mean service time of each drive is reported in milliseconds, with lower values representing better performance.

Our first wave of iPEAK workloads run reasonably well on the RE2-GP, whose mean service times are consistently quicker than those of the Caviar GP. The margins aren’t huge, but Western Digital’s firmware tweaks clearly make a difference here.

The terabyte competitors from Seagate and Hitachi are tough rivals here, no doubt thanks to their potent combinations of larger caches and faster spindle speeds. Still, the RE2-GP manages to hold its own in most workloads.

iPEAK multitasking – continued

Our second batch of iPEAK tests again shows the RE2-GP edging out its Caviar cousin. The drive doesn’t lead the terabyte field in any test pattern, though; it either falls behind the Deskstar 7K1000, the Barracuda 7200.11, or both.

IOMeter – Transaction rate
IOMeter presents a good test case for command queuing. To keep things easy to read, we’ve busted out two sets of graphs here. The first includes the Caviar GP and its closest competitors, while the second has results for all the drives we’ve tested. With over 20 drives, those latter graphs are a little difficult to read, so we’ll focus our attention on the first set and the Deskstar’s direct rivals.

Now this is interesting. The RE2-GP should be optimized for exactly the kind of multi-user environments that IOMeter simulates, but the drive actually delivers slightly lower transaction rates than the Caviar GP. That puts the RE2-GP behind Seagate’s Barracuda 7200.11, as well, although it’s at least well ahead of the terabyte Deskstar.

Good luck with these, folks.

IOMeter – Response time

Turning our attention to IOMeter response times, we again find the RE2-GP slightly behind the Caviar GP. Both drives trail the Barracuda 7200.11, with the Deskstar 7K1000 bringing up the rear of the terabyte train.

IOMeter – CPU utilization

CPU utilization is low across the board.

HD Tach
We tested HD Tach with the benchmark’s full variable zone size setting.

Even with the same spindle speed, the RE2-GP manages to sustain HD Tach transfer rates nearly 10MB/s higher than the Caviar GP. That’s some fancy firmware mojo right there. Of course, no one can touch the Barracuda 7200.11 here. The ‘cuda’s unique combination of 250GB platters spinning at a full 7,200RPM is a tough act follow.

Burst performance tends to line up along manufacturer lines, and Western Digital has always been a little conservative here. The RE2-GP is actually a little slower in this test than the Caviar GP.

The RE2-GP’s random access times are also a little higher, although only by a fraction of a millisecond. That still puts the RE2-GP at the back of the field, though.

The results from HD Tach’s CPU utilization test are within the +/- 2% margin of error for this test.

Noise levels
Noise levels were measured with an Extech 407727 Digital Sound Level meter 1″ from the side of the drives at idle and under an HD Tach seek load. Drives were run with the PCB facing up.

For all intents and purposes, the RE2-GP is every bit as quiet as the Caviar GP. Noise levels won’t matter much in a server room loaded with full racks, of course, but they could come into play if you’re building a home storage server that will live in your closet.

Power consumption
For our power consumption tests, we measured the voltage drop across a 0.1-ohm resistor placed in line with the 5V and 12V lines connected to each drive. Through the magic of Ohm’s Law, we were able to calculate the power draw from each voltage rail and add them together for the total power draw of the drive.

The RE2-GP draws only marginally more power than the Caviar GP, which is by far the most power-efficient drive we’ve ever tested. These results are even more striking when you look at the power consumption of our other terabyte drives. The Barracuda 7200.11 consumes quite a bit more power at idle and under load, and the Deskstar 7K1000 nearly doubles the RE2-GP’s seek power draw.

Gigabytes per watt
Inspired by a handy little graph whipped up by TR regular Usacomp2k3 in the discussion thread following our initial Caviar GP review, we’ve done some number crunching to calculate capacity per watt. This is an important metric for enterprise environments looking to maximize capacity within a fixed power budget, and we’ve chosen to express the results in gigabytes per watt.

The GreenPower drives deliver substantially more capacity per watt than the rest of the field. Even the Barracuda 7200.11, which is as good as the competition gets, manages 27 to 33% fewer gigabytes per watt than the RE2-GP.

Conclusions
Western Digital says it will start selling the RE2-GP starting today on its web site with a suggested retail price of $349.99 for the terabyte model—a $40 premium over the Caviar GP. You’re likely to pay much less when the RE2 hits larger online retailers, though. The Caviar GP 1TB can already be found for around $270 from several reputable dealers, and we’d expect the RE2-GP’s price to drop when the drive hits the street, as well.

Given its RAID-specific features and additional warranty coverage, the RE2-GP’s premium over the Caviar GP seems reasonable. The fact that the RE2-GP is a little bit faster in a number of tests helps, too. Although, curiously, the RE2-GP isn’t faster in IOMeter, where one would expect its enterprise credentials to shine. Then again, the GreenPower line isn’t about pack-leading performance. If it were, the drives would spin at a full 7,200RPM and probaby really lay a whuppin’ on their terabyte competition. As it stands, they spin a little slower and still manage to perform better in a number of tests.

The key to the GreenPower line, and especially to this enterprise variant, is energy efficiency. The RE2-GP delivers on that front in spades. With significantly lower idle and seek power consumption than its terabyte rivals, the RE2-GP enables much higher storage densities within a fixed power budget or much lower power consumption than solutions with equivalent capacity. The RE2-GP’s power savings will only add up in multi-drive RAID arrays, multi-server farms, or multi-rack data warehouses, lowering the drive’s cost of ownership relative to its competitors. That makes the RE2-GP an ideal candidate for enterprise applications that rank power efficiency and total capacity above absolute performance.