AMD Radeon R7 SSD in the test

While AMD can already come up with CPUs, graphics cards and RAM in the field of PC components, a permanent memory component was still missing. And since the Radeon brand focuses on performance, AMD did not get a hard drive manufacturer on board, but an SSD manufacturer. This resulted in the AMD Radeon R7 SSD series, which then has a few surprises in its luggage, as our test shows.

Intro

Like the Radeon main memory products, AMD does not manufacture the Radeon R7 SSD series itself. Entering the crowded consumer SSD market would simply be too expensive and risky. Instead, AMD has entered into a cooperation with OCZ. This is not hidden behind OEM abbreviations, but is printed in large letters on the back of the Radeon R7 drive. AMD didn't just buy an OCZ product and relabel it. Instead, the Radeon R7 is different from the OCZ ARC 100 and Vector 150 on some points.

Only the basic recipe is the same: a controller from OCZ (for which the controller specialist Indilinx was taken over some time ago) and NAND flash from the new OCZ parent company Toshiba. There is also the OCZ guarantee ShieldPlus. The ARC 100 is positioned as a device for beginners, the Radeon R7 is intended to make gaming computers steam, and the Vector 150 is intended for absolute enthusiasts.

While we probably won't be able to figure out why the marketing gave the product the "R7" name that was already used for another Radeon product - a series of graphics cards. But we'll see how this R7 SSD performs and how it stacks up against its competitors.

SSD bookmarks:

• Structure
• Empty or free?
• Wear leveling
• The root of all evil: Read-Modify-Writes garbage disposal
• Spare area and overprovisioning
• TRIM does not delete!
• Speed
• SLC, MLC, eMLC, TLC
• SSD caching - how-to

Recent SSD Reviews:

• OCZ ARC 100 with 256 GB
• Crucial MX100 with 256 GB
• Corsair Force LX with 256 GB
• Crucial M550 with 256 GB and 1 TB
• SanDisk Extreme II and Ultra Plus
• Samsung 840 Pro and EVO
• SSD reboot

The test candidate

Key data and technology

When studying the technical data, one immediately notices the technical relationship between the Radeon R7 SSD and the ARC 100. While the Flash NAND used is the same, a higher clocked version of the Barefoot 7 controller is used in the Radeon R3. In addition, the 480 GB version has more DRAM.

While the ARC 100 was somewhat more restricted when processing sequential data, either structurally or via firmware, OCZ promises significantly higher read and write rates for the performance model.

Equipment

The SSD encrypts data with 256-bit AES, but unfortunately does not support the TCG Opal specifications, which are necessary for the use of Microsoft's eDrive standard, for example. Unfortunately, there is also no support for the additional energy-saving modes such as DevSleep and HIPM + DIPM. When you pick up the drive, you notice that, at 115 grams, it feels heavier compared to competing models (e.g. Samsung's 850 Evo at 66 grams). Both factors don't make it an ideal drive for notebooks, but since it's a performance model, these two points are less important here. In a gamer PC (Radeon target group!) Or a workstation, the weight and 1 watt difference in idle are negligible.

Lifespan

Now let's take a look at durability. OCZ promises the customer an average write volume of 30 GB per day over four years "with typical end user loads". The manufacturer also explicitly mentions workstations as an area of ​​application. The bottom line is that you get a guaranteed write volume of almost 44 terabytes.

When testing the ARC-100 drives, we referred to the endurance test on the Kitguru.net website, which received five ARC-100 drives. When you write these lines, all five drives already have the 300 terabyte mark exceeded. Since the Radeon R7 and the ARC 100 use the same flash and the controller of the R7 is "only" clocked higher, one can assume - at least very cautiously - that the Radeon R7 also packs a multiple of the promised write volume on average. However, it is not uncommon for SSDs to achieve many times their promised write volumes in endurance tests.

Impressions

Apart from the key for the Acronis software (see software features), the SSD also comes with a 3,5 ″ mounting frame.

 

Image: AMD Radeon R7 SSD in the test

 

In the 256 GB version, the NAND flash memory was divided into 16 packages, 8 each on the front and rear. As a result, there are more dies in the individual packages in the 480 GB variant.

The ShieldPlus guarantee

For the newer models, OCZ offers the ShieldPlus guarantee. This applies to the ARC 100 and Vertex 460A for three years each, and four years for the AMD Radeon R7 SSD. In practice, the procedure works like this: If you have a problem with the SSD, contact support. The serial number of the model is sufficient as legitimation, no proof of purchase is required. If the support determines that there is obviously a defect, a new SSD will be sent directly to the customer, together with a return slip for the free return of the old SSD.

Software equipment

Together with the Radeon R7, the customer receives Acronis True Image HD in a bundle. This image program should support the migration of partitions to the SSD. The OCZ Toolbox is also included again, a tool for updating firmware and checking the SSD properties. A YouTube video illustrates this process. The OCZ Toolbox can be used for Microsoft Windows 7 and 8 (.1), Linux and Mac OS downloaded become. Acronis True Image HD is available for Windows versions from XP to 8.

If you want, you can use other means to ensure that the operating environment is optimally matched to the SSD drives. Important parameters are:

• Does the SATA port run in AHCI mode?
• Does the operating system support TRIM?
• Has any automatic defragmentation of the operating system been deactivated?

Test environment

Hardware

Test station:

• CPU: Intel Core i3 3220 - 2 x 3,3 GHz (Turbo: off) [Amazon offers]
• Motherboard: ASUS P8H77M (H77 chipset) [Amazon offers]
• Memory: 8 GB (4 x 2 GB) Team Xtreem - SPD operation: DDR3-1333 9-9-9-24-1T at 1,5 volts [Amazon offers]
• Power adapter: NZXT 650 Watt HALE82 Series [Amazon offers]
• Boot drive: OCZ Vertex-2-SSD as boot drive [Amazon offers]

The test candidate:

• MD Radeon R7 240GB (Amazon offers)

Comparison models:

• Corsair Force LX 256GB (HT4U-Test / Amazon offers)
• Corsair GTX 480GB (HT4U-Test / Amazon offers)
• Crucial M550 256GB (HT4U-Test / Amazon offers)
• Crucial M550 1TB (HT4U-Test / Amazon offers)
• OCZ ARC 100 2406 GB (HT4U-Test / Amazon offers)
• Samsung 840 120GB (HT4U-Test / Amazon offers)
• Samsung 840 EVO 250GB (HT4U-Test / Amazon offers)
• Samsung 840 Pro 256GB (HT4U-Test / Amazon offers)
• SanDisk Extreme 240GB (HT4U-Test / Amazon offers)
• SanDisk Extreme II 240GB (HT4U-Test / Amazon offers)
• SanDisk UltraPlus 256GB (HT4U-Test / Amazon offers)

Software

Our benchmark course

Our benchmark course aims to answer the following questions:

• How fast is the SSD reading and writing large files sequentially and reading and writing small files at random?
• How do fragmented blocks (not to be confused with file fragmentation!) And the resulting read-modify writes affect performance after a heavy write load?
• How fast is the SSD in a continuous load scenario (steady state)?
• Can TRIM restore full performance?
• How effective is garbage collection?
• How fast is the SSD when certain mixes of large and small blocks occur?

Synthetic benchmarks

The use of synthetic benchmarks cannot be avoided, since only with these the technical limits of the SSDs become visible. They show the maximum achievable.

With these benchmarks we determine the performance in the following states:

In this way it can be seen whether and to what extent the performance of the SSD is falling and whether TRIM can restore the original performance.

It doesn't matter whether you copy a few hundred MP3 or video files or simulate this work with Iometer, the effort is the same for the SSD. Differences resulting from the file system of the operating system then affect all SSDs equally, so that the ratios of the performance differences remain the same.

Trace benchmarks

Real life, on the other hand, can be simulated using trace benchmarks such as PCMark or Iometer profiles, which simulate use cases. With these tests, practical accesses are carried out in a reproducible manner.

For practical results, we carry out these tests after the SSD has already been written with load profiles several times and is occupied with active data except for a remaining 10 GB. This gives you the performance values ​​of an SSD that has already been used and is currently mostly full.

Treatments

We test less per application itself. There are two main reasons for this: First, the CPU limit falsifies the performance gap between the SSDs. For example, when the SSD has to wait for the CPU to process certain data before the SSD can continue working when the application starts. Due to the CPU limit, the SSDs move closer together than would be the case with faster CPUs later. Second, many applications can only be measured with a stopwatch, which is too imprecise for us, especially since the results are sometimes only tenths of a second apart. However, we carry out our long-serving OpenOffice copy test because it is easy to reproduce. We have only increased the amount of data there by a factor of 12. It is now 3,06 GB of data in over 48.000 files of various sizes that will be duplicated on the test drive.

Continuous load measurements

As described in the section “Load behavior”, SSDs collapse under a continuous random write load if the garbage collection cannot provide free blocks quickly enough. Such a load behavior occurs only rarely in normal home use. For one or the other reader, however, it might be interesting whether an SSD is also suitable for somewhat tougher use. For example, as a data carrier for a virtualizer, where a lot of small accesses can occur in parallel, or as a disk for a database test environment.

For this test, we unleash as many 4k writes as possible on the SSD via Iometer and create a graph that shows the performance over time. We repeat this test after a 30-minute or 12-hour break to see whether the garbage collection was able to provide enough free blocks for high performance during this time. Since Iometer works with a large test file, which is never deleted but only overwritten, influences by TRIM in these two repeat runs are excluded. The increase in performance through TRIM itself is then measured in a fourth run. This takes place after a quick format, which "trims" the drive. The test file is then created again.

We would like to point out that this goes well beyond the normal requirements for SSDs for home use. If an SSD does not do so well here, it is therefore not counted negatively. But we want to find out which SSDs stand out positively from the crowd. In addition, this test makes it easier to see to what extent the garbage collection is working.

MByte / s or IOPS?

Usually we give the measurement results in megabytes per second. In the profile tests, however, we give the results in IOPS (Input / Output Operations per Second = input and output commands per second). An input or output command can mean reading or writing a block. This does not affect the comparability. If a data carrier manages 128 IO per second in a write test with 1.000 KB blocks, then mathematically this results in 1.000 * 128 KB = 128 MB per second. When an operating system writes MP3 files or videos, it does so in blocks, too, and the block sizes ultimately depend on the size of the files and the formatting of the file system. With many small files, this may limit the number of IOPS and with large files the maximum write rate of the SSD. Therefore, it makes sense to use the specification of IOPS wherever a high number of read and write operations take place and / or different block sizes are involved.

In the case of continuous load measurements, the indication in IOPS has the additional advantage that the maximum IOPS information usually advertised by the manufacturers can be compared directly with the real results.

Measurement results

Sequential reading

The Radeon R7 is not one of the strongest SSDs in reading either. The 550MB / s announced in the data sheets were based on the Atto benchmark. Iometer and AS SSD deliver noticeably lower values. However, one should not lose sight of the fact that the measurable difference of 3% to the top of the field at Iometer is so small that it cannot be felt in practice. The difference with AS SSD is 10%. This is a bit clearer, but ultimately only means a difference between 1 and 1,9 seconds for the 2 GB test files from AS SSD.

Sequential writing

The sequential write performance clearly shows that the R7 combines high write performance with a low performance loss under load. The two OCZ SSDs have by far the lowest performance drop in the area of ​​the 240/256 GB SSDs. In the AS SSD benchmark with its very short sequential write spurts, the EVO competitor with its TurboWrite mechanism can still outperform, but otherwise the distribution remains more or less the same.

Random reading

The measured values ​​for the random 4k reading clearly show that the Barefoot 3 controller cannot really distinguish itself here. Both Iometer and AS SSD are behind the competition. In practice, that doesn't look that bad. As we will see in the later, read-intensive web server benchmark, the Radeon R7 is among the front runners. Apparently, many manufacturers optimize their firmware for 4k access and then weaken with other block sizes.

Random writing

Very good results can be seen with random 4k writing, whereby the comparatively small drop under load can be seen. The other test subjects in the front field are all equipped with significantly more flash, which is advantageous due to the larger spare areas. The Radeon R7 is right in front in the closely spaced field in the short write bursts from AS SSD.

Web server, file server, workstation

The web server benchmark allows data of various block sizes to be read continuously and shows that the lower synthetic reading performance does not have as negative an effect in practice as the synthetic benchmarks suggest. In the entry-level segment, the ARC 100 only has to admit defeat to the Samsung EVO.

This clearly shows the strengths of the Barefoot 3 models from OCZ. The Radeon R7 clearly outperforms the competition in the two write-oriented tests. It even has a higher performance than the 1 TB variant of the M550 from Crucial, although it has a lead with significantly larger spare areas. The influence of the drive size becomes clear when you compare it with the measured value of the 256 GB version of the M550. OCZ's promise of long-lasting write performance under load can also be seen as fulfilled here.

HT4U-OpenOffice copy test

In the simple copy test, the test person runs along in the middle field without any abnormalities.

PCMark7 trace benchmarks

With the slight reading weakness and the strong writing behavior, the Radeon R7 cannot really set itself apart from the competition in the more reading-oriented practical tests. The priorities with this drive are the heavier loads.

Continuous load curves

When looking at the graph, you get the impression that the SSD is artificially withheld at a level of 73.000 to 75.000 IOPS, because at the beginning of the transition phase the drive increases again to over 80.000 IOPS. As with any SSD, the performance drops in the transition phase when there are no more free blocks and the garbage collection has to actively ensure free blocks even during writing. The Radeon R7 falls to an average of 20.000 IOPS, which puts it well ahead of the competition. The lead over the ARC 100 is probably due to the higher frequency of the controller. The M10 version of the Barefoot 3 on the ARC 100 clocks at 352 MHz, while the M00 version on the Radeon 7 runs at 397 MHz.

Furthermore, it can be stated that the controller clears some of the blocks via garbage collection during idle times. He showed that this cannot be taken for granted Test the Samsung 840 Pro, which therefore also did relatively poorly for a performance model in the load scenario in the sequential Iometer write test.

input

The idle values ​​are not very high with results of around 0,6 watts. Since additional energy-saving modes such as DevSleep are not supported, the energy consumption cannot be reduced any further.

For technological reasons, paperwork is the most power-hungry, as the cells then have to be charged and discharged and the controller has the most computing effort. Consumption of 5 watts sounds like a lot for SSDs, but you shouldn't lose sight of the fact that in most cases these are idle, and when something needs to be done, it's mostly read access. You should therefore always see these results in relation to your own use of the SSD.

Summary

Let's summarize how the drive fares in the individual categories: While the simple performance data for sequential read and write rates do not stand out from the competition, the Radeon R7 SSD clearly leaves the competition products behind under load . On the other hand, it weakens when reading randomly in comparison with the other models from the performance segment, but it is even more noticeable when writing randomly.

Marketing places the Radeon R7 SSD as a drive for gamers. This SSD is no less suitable for this than the competing products. The measurable reading weakness is only measurable in practice, but not noticeable. The Radeon R7 is noticeable because of its much better performance in load scenarios than the competition. In other words, the area where the wheat is still separated from the chaff these days.

In addition, their price has fallen massively since it was launched. In the field of performance models there are cheaper candidates like the Crucial M550 with 256 GB. However, this has lower write rates and even less performance under load. While the Radeon R7 can hardly stand out from the competition in systems with little load, nothing speaks against a bundle in a gamer's computer. But its price-performance ratio is really unbeatable as a cheap SSD in workstation use.

[ri], February 14, 2015