Toshiba has a wide range of storage solutions for every computing need, in both HDD and SSDs, for client and enterprise applications. From bulk HDD storage to performance HDDs, and the increasingly popular SSD segment, Toshiba features solutions for every workload. A long history of OEM integration has given Toshiba plenty of experience delivering reliable solutions, and their extensive history in NAND production affords them deep knowledge of flash products.
The SSD field has fragmented into purpose-built SSDs; from high-end 12GB/s SAS to 6GB/s SATA, there is a different offering with specific features for specific applications. The perfect example of a purpose built SSD is the Toshiba THNSNJ512GCSU HG6, but it is based on a surprisingly modular architecture.
The HG6 addresses relatively light workloads such as server boot drives, read-intensive server applications, workstations, thin clients, and notebook PCs. The THNSNJ architecture comes in a variety of form factors, with 7.2mm and 9.5mm Z-heights for traditional 2.5" applications, and mSATA and M.2 single and double-sided versions also available.
The HG6 comes in capacities of 60, 128, 256, and 512 GB, signaling that these are SSDs with zero overprovisioning. The maximum amount of NAND is user-addressable, which is a desirable characteristic for the value market. This usually leads to lower endurance and random performance specifications. The HG6 features up to 90,000 / 30,000 IOPS read/write. The HG6 also offers sequential read/write speeds of 534/482 MB/s, respectively.
The HG6 features Adaptive SLC Write cache technology. This allocates a portion of normal MLC as SLC NAND during use. This significantly speeds random write performance, particularly in bursty operating system environments. SLC cache also increases endurance by writing cached random data as sequential data when flushing to the MLC NAND layer. The adaptive portion allows the dynamic adjustment of the size of the SLC layer depending upon available capacity. The HG6 provides 1 DWPD of endurance for five years, which is more endurance than competing solutions in our test pool. We test with sustained workloads for enterprise testing, so readers must keep in mind that in bursty environments, the SLC layer can provide significant performance enhancements. This is the first enterprise SSD we have tested with an adaptive SLC layer.
The HG6 also offers end-to-end data protection and TCG/OPAL 2.0 compliant encryption, along with Wipe Technology on the SED SKUs, which are important for use in business laptops. Advanced Power Management (APM) helps regulate power in tandem with Devsleep. Deterministic Zeroing TRIM is supported, and is particularly useful in operating system environments. The SSD offers an expected three-year warranty, which is fine for many OEM deployments, but is a bit lacking for the retail market. Toshiba's proprietary QSBC (Quadruple Swing by Code), an enhanced ECC algorithm, protects user data.
Some of the optimizations for smaller form factors, such as M.2 and mSATA, are present on the 2.5" drive as well. The THNSNJ512GCSU does not feature a DRAM cache module, which typically speeds access to the all-important LBA tables. These tables, or maps, allow the controller and FTL (Flash Translation Layer) to quickly and effectively route data to the correct location on the NAND packages. Eliminating the DRAM cache chip is ideal for space-constrained implementations such as mSATA and M.2 designs, and has the tertiary benefits of reducing power consumption and component cost.
Another advantage of forgoing DRAM cache modules is an enhanced resilience to host power-loss events. The majority of power capacitor arrangements, which provide power-fail protection, primarily flush data from the DRAM cache to the NAND. With no DRAM cache, a capacitor-less design such as the HG6 does not have to cope with flushing the volatile DRAM cache during power loss. Firmware enhancements can detect and flush data prior to total power loss. Typical designs flush data to the "fast" NAND pages, which works well in tandem with the SLC layer present in the HG6.
The negative impact of forgoing DRAM caching is usually felt in performance. Slower access to the LBA tables can slow the performance in some scenarios. Competing DRAM-less designs use data compression algorithms to speed access to the LBA tables held on the NAND media. We are not sure if Toshiba implements this type of feature to speed access to the data tables, so let's take a closer look to see how the HG6 fares in our testing.
Toshiba HG6 Internals and Specifications
Toshiba HG6 Internals
The Toshiba THNSNJ512GCSU HG6 comes in the 2.5" form factor with a 7.2mm and 9.5mm z-height, with Toshiba branding. We can spot the thermal pads peeking out through the holes in the bottom of the case.
An additional four internal fasteners hold the HG6's PCB in place.
The PCB is fastened to a riser, which provides the correct gap between the PCB and thermal pads. These thick pads transfer heat from the components into the case to keep the drive cool in demanding environments.
The bottom of the PCB is bare, with all eight Toshiba A19nm Toggle 2.0 MLC NAND packages on top. We also spot the bare mounting pad for a DRAM package, which is not included in this model of the drive.
The custom Toshiba T635879BXBG controller powers the HG6.
Toshiba HG6 Specifications
The THNSNJ512GCSU also comes in a SED model, signified by the THNSNJ512GBSU part number.
Test System and Methodology
Our approach to storage testing targets long-term performance with a high level of granularity. Many testing methods record peak and average measurements during the test period. These average values give a basic understanding of performance, but fall short in providing the clearest view possible of I/O QoS (Quality of Service).
While under load, all storage solutions deliver variable levels of performance. "Average" results do little to indicate performance variability experienced during actual deployment. The degree of variability is especially pertinent, as many applications can hang or lag as they wait for I/O requests to complete. While this fluctuation is normal, the degree of variability is what separates enterprise storage solutions from typical client-side hardware.
Providing ongoing measurements from our workloads with one-second reporting intervals illustrates product differentiation in relation to I/O QoS. Scatter charts give readers a basic understanding of I/O latency distribution, without directly observing numerous graphs. This testing methodology illustrates performance variability, and includes average measurements during the measurement window.
IOPS data that ignores latency is useless. Consistent latency is the goal of every storage solution, and measurements such as Maximum Latency only illuminate the single longest I/O received during testing. This can be misleading, as a single "outlying I/O" can skew the view of an otherwise superb solution. Standard Deviation measurements consider latency distribution, but do not always effectively illustrate I/O distribution with enough granularity to provide a clear picture of system performance. We utilize high-granularity I/O latency charts to illuminate performance during our test runs.
Our testing regimen follows SNIA principles to ensure consistent, repeatable testing, and utilizes multi-threaded workloads found in typical production environments. We measure power consumption during precondition runs. This provides measurements in time-based fashion, with results every second, to illuminate the behavior of power consumption in steady state conditions. We also present IOPS-to-Watts measurements to highlight efficiency.
The SSDs in the test pool feature the following capacities: Toshiba THNSNJ512GCSU HG6 is 512GB, and the 845DC EVO and Toshiba HK3R are 480GB. The SSDs are tested over their full LBA range to highlight performance at maximum utilization. The first page of results will provide the "key" to understanding and interpreting our test methodology.
Benchmarks - 4k Random Read/Write
4k Random Read/Write
We precondition the 512GB Toshiba THNSNJ512GCSU HG6 for 9,000 seconds, or two and a half hours, receiving performance reports every second. We plot this data to illustrate the drives' descent into steady state.
This dual-axis chart consists of 18,000 data points, with the IOPS on the left and the latency on the right. The red dots signify IOPS, and the grey dots are latency measurements during the test. We place latency data in a logarithmic scale to bring it into comparison range. The lines through the data scatter are the average during the test. This type of testing presents standard deviation and maximum/minimum I/O in a visual manner.
Note that the IOPS and latency figures are nearly mirror images of each other. This illustrated high-granularity testing can give our readers a good feel for latency distribution by viewing IOPS at one-second intervals. This should be in mind when viewing our test results below. This downward slope of performance only occurs during the first few hours of use, and we present precondition results only to confirm steady state convergence.
Each level tested includes 300 data points (five minutes of one second reports) to illustrate performance variability. The line for each OIO (Outstanding I/O) depth represents the average speed reported during the five-minute interval. 4k random speed measurements are an important metric when comparing drive performance, as the hardest type of file access for any storage solution to master is small-file random. 4k random performance is a heavily marketed figure, and is one of the most sought-after performance specifications.
The Toshiba HG6 averages 67,693 IOPS, and the Toshiba HK3R averages an impressive 83,266 IOPS at 256 OIO. The Toshiba HK3R is second only to the Samsung 845DC EVO, which delivers 85,155 IOPS at 256 OIO. The HG6 random read performance isn't chart-topping, but it is sufficient for most moderate workloads.
Our Latency vs IOPS charts compare the amount of performance attained from each solution at specific latency measurements. Many applications have specific latency requirements. These charts present relevant metrics in an easy-to-read manner for readers who are familiar with their application requirements.
The HG6's incrementally slower random read performance results in 67,030 IOPS at .1ms; the HK3R provides 83,000 IOPS at .1ms, and the 845DC EVO delivers 85,000 IOPS.
Garbage collection routines are more pronounced in heavy write workloads, leading to performance variability.
Under a sustained heavy 4k write workload, the Toshiba HG6 scores 7,544 IOPS at 260 OIO, the HK3R averages 17,437 IOPS, and the 845DC EVO averages 13,841 IOPS. The HG6 has lower speed in this sustained workload, but in bursty operating system environments, this will not be as much of a concern. It does exhibit a nice, tight performance profile that is a welcome sight. This type of consistent performance yields tremendous gains in RAID environments.
The HG6 is clearly not intended for sustained write workloads, and falls well behind the competing SSDs in this test. Even though there was a tight performance profile with the IOPS output, the latency exhibits some variability.
Our write percentage testing illustrates the varying performance of each solution with mixed workloads. The 100% column to the right is a pure 4k write workload, and 0% represents a pure 4k read workload.
The HG6 makes a nice gain in performance consistency in this test, albeit at the sacrifice of some performance.
We record power consumption measurements during our precondition run. We calculate the stated average results after the device has settled into steady state during the last five minutes of the test.
The low power consumption of 2.95 watts is an excellent characteristic for an SSD headed into varying applications, including mobile. The HK3R averages 3.85 watts, and the 845DC EVO averages 3.55 watts.
IOPS-to-Watts measurements are generated from data recorded during our precondition run, and the stated average is from the last five minutes of the test.
We test power consumption and efficiency from write workloads, which is not the intended workload for the HG6. The low power consumption does not entirely offset the lower write performance, and the HG6 scores 1,862 IOPS per watt. Write efficiency metrics are not going to be a huge concern for a SSD headed into light read workload environments. The HK3R averages 4,249 IOPS per watt, and the 845DC EVO averages 3,973 IOPS per watt.
Benchmarks - 8k Random Read/Write
8k Random Read/Write
Many server workloads rely heavily upon 8k performance, and we include this as a standard with each evaluation. Many of our server workloads also test 8k performance with various mixed read/write distributions.
The average 8K random read speed of the Toshiba HG6 is 42,511 IOPS, the Toshiba HK3R is 47,315 IOPS, and the Samsung 845DC EVO is 52,731 IOPS.
The HG6 exhibits an expected latency profile considering its performance characteristics.
The HG6 averages 4,371 IOPS with the same consistent performance profile. The HK3R averages 9,656 IOPS, and the 845DC EVO averages 7,112 IOPS.
The HG6 latency measurements fall behind due to the more robust write performance of the other SSDs in the test pool.
The HG6 falls behind in performance, but once again posts a very solid and consistent band of performance.
Power consumption for the HG6 averages 3.04 watts, the HK3R averages 3.91 watts, and the 845DC EVO averages 3.54 watts.
The HG6 averages 1,108 IOPS per watt, the HK3R averages 3,112 IOPS per watt, and the 845DC EVO averages 1,994 IOPS per watt.
Benchmarks - 128k Sequential Read/Write
128k Sequential Read/Write
128k sequential speed reflects the maximum sequential throughput of the SSD, and is indicative of performance in OLAP, batch processing, streaming, content delivery applications, and backup scenarios. The Toshiba HG6 pulls off an impressive 523 MB/s, the HK3R averages 514 MB/s, and the Samsung 845DC EVO averages 526MB/s at 256 OIO. The HG6 features impressive sequential read performance.
The HG6 goes neck and neck with the 845DC EVO in sequential read latency performance.
Sequential write performance is important in tasks such as caching, replication, HPC, and database logging. The Toshiba HG6 averages 445 MB/s; the HK3R is also excellent in this test with a blistering 456 MB/s. The 845DC EVO trails with 427 MB/s.
The HG6 trails its sibling, the HK3R.
The HG6 experiences significant variability during the mixed sequential workload testing.
The HG6 continues to deliver excellent low power performance with an average of 3.42 watts. The HK3R averages 4.12 watts, and the 845 DC EVO averages 4.14 watts.
The Toshiba HG6 averages 130 MB/s. The HK3R scores 111 MB/s per watt, and the 845DC EVO averages 99 MB/s per watt.
Benchmarks - Database/OLTP and Web Server
This test consists of Database and On-Line Transaction Processing (OLTP) workloads. OLTP is the processing of transactions such as credit cards and high frequency trading in the financial sector. Databases are the bread and butter of many enterprise deployments. These demanding 8k random workloads with a 66 percent read and 33 percent write distribution bring even the best solutions down to earth.
The HG6 trails the other SSDs, but offers a very consistent 13,787 IOPS at 256 OIO. The HK3R averages 20,332 IOPS, but suffers extensive variability. The Samsung 845DC EVO has a very consistent average of 19,678 IOPS at 256 OIO.
The HG6 offers steady latency performance, and improves at 256 OIO.
The HG6 averages 3.13 watts, the HK3R averages 3.52 watts, and the 845DC EVO averages 3.55 watts.
The HG6 averages 3,068 IOPS per watt, the HK3R averages 6,065 IOPS per watt, and the 845DC EVO averages 5,465 IOPS per watt.
The Web Server workload is read-only with a wide range of file sizes. Web servers are responsible for generating content users view over the Internet, much like the very page you are reading. The speed of the underlying storage system has a massive impact on the speed and responsiveness of the server hosting the website.
The HG6 offers impressive performance of 26,003 IOPS at 256 OIO, the HK3R averages 28,114 IOPS, and the 845DC EVO averages 28,054 IOPS.
The 845DC EVO and HK3R run neck and neck in latency vs. IOPS testing.
The HG6 averages 3.02 watts, the HK3R averages 3.88 watts, and the 845DC EVO averages 3.55 watts.
The HG6 averages 573 IOPS per watt, the HK3R averages 1,125 IOPS per watt, and the 845DC EVO scores 914 IOPS per watt.
Benchmarks - Email Server
The email server workload is a demanding 8K test with a 50% read and 50% write distribution. This application is indicative of the performance in heavy write workloads.
The HG6 averages 7,728 IOPS at 256 OIO, the HK3R averages 17,342 IOPS, and the 845DC EVO averages 14,041 IOPS. This heavy write workload is clearly not the HG6's forte.
The HG6 averages 3.07 watts, the HK3R averages 3.79 watts, and the 845DC EVO averages 3.54 watts.
The HG6 averages 2,008 IOPS per watt, the HK3R averages 4,487 IOPS per watt, and the 845DC EVO averages 3,815 IOPS per watt.
The Toshiba THNSNJ512GCSU is a purpose-built solution for the value-conscious who need to address read-centric and operating system workloads, such as boot drives in servers. The DRAM-less design is advantageous to reducing component cost, and though it isn't ensured, without volatile DRAM cache, there is less chance of losing data due to a host power-loss event.
The lack of a DRAM cache also incurs a performance penalty, but for read-centric environments, and even mobile applications, the fastest SSD isn't always a steadfast requirement. For those with a good knowledge of their workload requirements, the Toshiba HG6 can provide more than enough performance, and the zero overprovisioning design offers maximum usable capacity.
One encouraging sign from the HG6 was its consistent performance profile during our testing. The consistent performance will pay off in RAID environments and application performance, even though it comes at a sacrifice of some speed.
The HG6 isn't going to win any top performance awards, but it isn't designed to. In random read/write workloads, it offered solid performance, but trailed the other entrants. It also trailed in mixed random workload testing, but delivered a solid consistent performance profile that would be suitable for most low-intensity applications.
Sequential performance was a bright spot, with solid read/write performance that matches, or beats, the competition. The HG6 suffered more performance variability when we moved into mixed sequential workloads. In our server workloads, the HG6 continued the trend of offering steady, consistent performance that didn't top the charts. Many of our server workloads are write-intensive; in the read-centric web server tests the HG6 was very competitive with the other SSDs in the test pool.
Another bright spot was the very conservative power consumption of the HG6. It consistently beat the competition with excellent power consumption metrics. SSDs often encounter extensive idle time in operating system environments; the HG6 offers Devsleep functionality to accommodate that reality, effectively shutting down the SSD while waiting for commands.
The HG6 also did not get to display the benefits of its adaptive SLC cache layer in our testing. This layer of SLC cache catches the performance-inhibiting small random writes, and flushes them to the MLC layer later. This delivers a big performance boost in operating system environments, but is not displayed in our sustained workload testing. The SLC layer also boosts endurance, which is listed at 1 DWPD for five years. The HG6 also offers TRIM support for operating system environments.
The HG6 is designed to meet a specific need, and a complex mix of factors always affects the purchasing decision. The HG6 offers multiple enterprise features, such as end-to-end data protection and SED options. If priced competitively enough, the HG6 can be the answer for those looking for an economical way to boost performance while maintaining a low power threshold.
|Quality, Design, Build and Warranty||90%|
|Power Consumption and Efficiency||92%|
The Bottom Line: Toshiba's THNSNJ architecture is flexible enough to find itself into many winning OEM solutions. The HG6 is a purpose-built solution for read-centric applications, and offers end-to-end data protection, Devsleep, encryption, SED, and other enterprise -specific features.
PRICING: You can find products similar to this one for sale below.
United States: Find other tech and computer products like this over at Amazon.com
United Kingdom: Find other tech and computer products like this over at Amazon.co.uk
Australia: Find other tech and computer products like this over at Amazon.com.au
Canada: Find other tech and computer products like this over at Amazon.ca
Deutschland: Finde andere Technik- und Computerprodukte wie dieses auf Amazon.de