Designed to replace the DC1000M for mixed-use workloads in data centre environments, Kingston's DC1500M uses a Gen 3.0 x4 NVMe PCIe interface to deliver high throughput with low latency via a U.2 connection. The DC1500M comes in four capacities to cover most data centre needs; 960GB, 1.92TB, 3.84TB (the drive we are reviewing here) and the flagship 7.68TB model. The drive uses a Silicon Motion SMI 2270 16-channel controller combined with BiCS4 96-Layer 3D TLC NAND. Kingston quote performance figures for the DC1500M series as up to 3,100MB/s for Sequential reads for the 960GB, 3.84TB and 7,68TB drives with the 1.92TB drive the fastest in the range at 3,300MB/s. Sequential writes are listed as up to 2,700MB/s across the range with the exception of the 960GB drive which is rated at 1,700MB/s. 4K Steady State read performance is listed as up to 440,000 IOPS for the 960GB drive, 510,000 IOPS for the 1.92TB model and up to 480,000 IOPS and 420,000 IOPS for the 3.84TB and 7.68TB models respectively. As for random writes, the 960GB drive is the slowest at 150,000 IOPS, next comes the 7.68TB drive at up to 200,000 IOPS, the 3.84TB drive is rated at up to 210,000 IOPS and the fastest is the 1.92TB drive at up to 220,000 IOPS. Power consumption for the 3.84TB drive is quoted as 6.80W idle, 6.40W for average reads (6.63W maximum reads) and 14.20W for average writes (16W maximum writes). As you should expect from a drive aimed at data centre usage, it comes with enterprise-class power loss protection in the shape of power capacitors. The drive has an endurance rating of 1 DWPD (Drive Writes Per Day) over the length of the 5-year warranty Kingston backs the drive with. Physical Specifications: Usable Capacities: 3.84TB. NAND Components: BiCS4 96-Layer 3D TLC NAND. NAND Controller: Silicon Motion SMI 2270 16-channel. Cache: DDR4. Interface: NVMe PCIe Gen 3.0 x4 / U.2. Form Factor: 2.5″ x 15mm Dimensions: 100.09 x 69.84 x 14.75mm. Drive Weight: 155g. Firmware Version: S67F0103. Many companies deliver their data centre/enterprise drives in plain vanilla boxes, not so Kingston, they ship theirs in blister packs just like their retail consumer drives. The DC1500M’s blister pack has the drive’s capacity clearly labelled on the front along with which market segment (s) the drive is aimed at and a logo that displays the 5-year warranty that Kingston backs the drive with. The rear of the packaging has multilingual marketing and warranty notes on it. The drive is built on a 2.5in 15mm format and comes with a metal enclosure which also serves as a pretty good heatsink. At the heart of the DC1500M lies a Silicon Motion SM2270 16-channel controller which is combined with Kioxia BiCS4 96-layer 3D TLC NAND. The Silicon Motion SM2270 is a high-performance PCIe NVMe controller designed for data centre drives. The 16 channel controller features triple ARM Cortex R5 Dual-core CPUs and has a PCIe Gen3 x8 host interface and dual 32 bit DRAM data bus. The controller's firmware has been updated to now support up to 64 namespaces. The drive uses a U.2 (PCIe NVMe Gen3 x4 ) interface which is hot-pluggable and compatible with U.2 server and data storage backplanes. Kingston’s SSD management software utility is simply called SSD Manager. With it, you can monitor the health of the drive and how it’s being used, check the drive’s SMART data (including reliability tracking, usage statistics, life remaining, wear levelling and temperature) and update the firmware. For testing, the drives are all wiped and reset to factory settings by HDDerase V4. We try to use free or easily available programs and some real-world testing so you can compare our findings against your own system. This is a good way to measure potential upgrade benefits. Main system: Intel Core i7-7700K with 16GB of DDR4-3200 RAM, Sapphire R9 390 Nitro and an Asus Prime Z270-A motherboard. Other drives Intel Optane SSD DC P4800X 750GB. Intel Optane SSD DC P5800X 800GB. Intel SSD DCP4510 8TB. Kingston DCP1000 1.2TB (RAID 0). Kingston DCP1000 1.2TB (RAID 1). Kingston DC1000B 480GB. Kingston DC1000M 1.92TB. Kingston KC1000 960GB. Software: Atto Disk Benchmark 4. IOMeter. SPECworkstation 3.1 All our results were achieved by running each test five times with every configuration this ensures that any glitches are removed from the results. Trim is confirmed as running by typing fsutil behaviour query disabledeletenotify into the command line. A response of disabledeletenotify =0 confirms TRIM is active. To aid in the testing of the drive, Intel kindly supplied us with a U.2 adaptor and the appropriate data/power cabling. The ATTO Disk Benchmark performance measurement tool is compatible with Microsoft Windows. Measure your storage systems performance with various transfer sizes and test lengths for reads and writes. Several options are available to customize your performance measurement including queue depth, overlapped I/O and even a comparison mode with the option to run continuously. Use ATTO Disk Benchmark to test any manufacturers RAID controllers, storage controllers, host adapters, hard drives and SSD drives and notice that ATTO products will consistently provide the highest level of performance to your storage. Kingston rate the Sequential performance of the 3.84TB version of the DC1500M as up to 3,100MB/s and up to 2,700MB/s for Sequential read and writes respectively. Using the ATTO benchmark we could confirm the maximum read figure with a test result of 3,110MB/s, while writes were a little shy of the official maximum at 2,560MB/s. We could confirm and indeed better both the official Sequential figures of up to 3,100MB/s and up to 2,700MB/s for read and writes respectively with our own tests. The drive produced peak read test results of 3,485MB/s using 4-threads and 3,480MB/s for a single thread. Sequential write performance peaked at 2,855MB/s (QD4) using a single thread and 2,881MB/s (QD1) using four threads, both confirming the official figures. From QD4 both reads and writes showed pretty consistent performance. Kingston rate the random 4K read performance of the drive as up to 480,000 IOPS. Using our 4-threaded tests, we couldn't quite hit that maximum figure, the best we saw was 456,920 IOPS (QD64). Our tests also confirmed the low latencies that the drive operates at throughout the tested queue depths. The DC1500M is rated at up 210,000 IOPS for random writes with the drive in a steady state. With our FOB 4-threaded testing we got a peak IOPS figure of 229,483 IOPS at QD of 128. As the DC1500M has been designed for use in data centres we also tested the drive using 8K data. In our read test, the drive climbed steadily through the tested queue depths, topping out at 262,93 IOPS (QD32) before slowly dropping back to finish the test run at 243,615 IOPS at QD128. Once again the low latencies are worthy of note. In the 8K write test, the DC1500M peaked at the end of the test run at 119,266 IOPS at QD128. The DC1500M has been designed to deliver high I/O performance at low latencies for mixed workloads. Our mixed load testing confirmed both of these aims, particularly at lower queue depths. In our read throughput test, the DC1500M peaked at 2,663MB/s at the 16MB block mark, short of the official maximum of 3,100MB/s. In our write throughput test, the drive peaked at 2,435MB/s again short of the official maximum of 2,700MB/s but closer to the official figure than the read performance. We also tested the drive in a number of workload scenarios that it might face in real life. The IOMeter settings we used to test with are listed below:- Database Server Transfer Size: 4K Reads: 67% Writes: 33% Random: 100% Boundary: 4K Outstanding IO: 1 - 128 Threads/Workers: 4 Exchange Email 4K Transfer Size: 4K Reads: 67% Writes: 33% Random: 100% Boundary: 4K Outstanding IO: 128 Threads/Workers: 4 Exchange Email 8K Transfer Size: 84K Reads: 67% Writes: 33% Random: 100% Boundary: 8K Outstanding IO: 64 Threads/Workers: 4 Exchange Email 32K Transfer Size: 32K Reads: 67% Writes: 33% Random: 100% Boundary: 4K Outstanding IO: 128 Threads/Workers: 4 Media Streaming Transfer Size: 64K Reads: 98% Writes: 2% Sequential: 100% Boundary: 4K Outstanding IO: 64 Threads/Workers: 8 Search Engine A Transfer Size: 4K Reads: 100% Writes: 0% Random: 100% Boundary: 4K Outstanding IO: 128 Threads/Workers:8 Search Engine B Transfer Size: 8K Reads: 100% Writes: 0% Random: 100% Boundary: 4K Outstanding IO: 128 Threads/Workers:8 Web File Server 4KB Transfer Size: 8K Reads: 95% Writes: 5% Random: 75% Boundary: 4K Outstanding IO: 32 Threads/Workers: 8 Web File Server 8KB Transfer Size: 8K Reads: 95% Writes: 5% Random: 75% Boundary: 4K Outstanding IO: 64 Threads/Workers: 8 Web File Server 64KB Transfer Size: 64K Reads: 95% Writes 5% Random: 75% Boundary: 4K Outstanding IO: 64 Threads/Workers: 8 Web Server Log Transfer Size: 8K Reads: 0% Writes: 100% Random: 100% Boundary: 4K Outstanding IO: 128 Threads/Workers: 4 Video On Demand Transfer Size: 512K Reads: 100% Writes: 0% Random: 100% Boundary: 4K Outstanding IO: 64 Threads/Workers: 8 As you might expect for a drive designed to handle most of these scenarios, the DC1500M tackles these tests very efficiently with some strong bandwidth figures. SPECworkstation 3 is a specialized test designed for benchmarking the key aspects of workstation performance; it uses over 30 workloads, containing nearly 140 tests to test CPU, graphics, I/O, and memory bandwidth. The workloads fall into seven categories; Media and Entertainment - 3D animation, rendering Product Development - CAD/CAM/CAE Life Sciences - medical, molecular Energy - oil and gas Financial Services, General Operations GPU Compute. We use the WPCstorage section of the benchmark to test drives which uses fifteen separate tests. In the SPECworkstation 3.1 test, the DC1500M outperforms the previous DC1000M drive in all but one test, the Energy Storage test. For the long term performance stability test, we set the drive up to run a 20-minute 4K random test with a 30% write, 70% read split, at a Queue Depth of 256 over the entire disk. The 3.84TB Kingston DC1500M averaged 59,272 IOPS for the test with a performance stability of 83%. We took note of the drive’s temperature during some of our benchmarking runs. The DC1500M average peak temperature from the 26 runs we measured was 47° C (23° C ambient) with the highest temperatures coming during the Database Server QD4 workload test run at 60° C. To test real life performance of a drive we use a mix of folder/file types and by using the FastCopy utility (which gives a time as well as MB/s result) we record the performance of drive reading from & writing to a 256GB Samsung SSD850 PRO. 100GB data file. 60GB iso image. 60GB Steam folder – 29,521 files. 50GB File folder – 28,523 files. 12GB Movie folder – (15 files - 8 @ .MKV, 4 @ .MOV, 3 @ MP4). 10GB Photo folder – (304 files - 171 @ .RAW, 105 @ JPG, 21 @ .CR2, 5 @ .DNG). 10GB Audio folder – (1,483 files - 1479 @ MP3, 4 @ .FLAC files). 5GB (1.5bn pixel) photo. BluRay Movie - 42GB. 21GB 8K Movie demos - (11 demos) 16GB 4K Raw Movie Clips - (9 MP4V files). 4.25GB 3D Printer File Folder - (166 files - 105 @ .STL, 38 @ .FBX, 11 @ .blend, 5 @ .lwo, 4 @ .OBJ, 3@ .3ds). 1.5GB AutoCAD File Folder (80 files - 60 @ .DWG and 20 @.DXF). In our seven real-life file transfer tests, the DC1500M averaged 422MB/s when writing (the fastest being the 519MB/s for the 8K Movie Scene folder) and 264MB/s when reading the data back, the fastest performance coming from the 3D Print Folder transfer. To get a measure of how much faster PCIe NVMe drives are than standard SATA SSD's we use the same files but transfer to and from a 512GB Toshiba OCZ RD400. Taking the SATA drive out of the equation saw, as you would expect, transfer speeds to rocket upwards with the drive averaging 1,299MB/s for writes and 1,160MB/s for reads. Designed as a replacement for the DC1000M (review here), the DC1500M features the same Silicon Motion SMI 2270 16-channel controller as the previous drive (but with updated firmware) and marries it up with BiCS4 96-Layer 3D TLC NAND instead of the DC1000M's BiCS3 64-layer NAND. The drive implements Kingston’s QoS (Quality of Service) requirements to ensure predictable random I/O performance and low latencies over a wide range of read and write data centre centric workloads. The drive uses a U.2 (PCIe NVMe Gen3 x4) interface which is hot-pluggable and compatible with U.2 backplanes of storage systems and enterprise servers. As you might expect for a data centre/enterprise focussed drive, the DC1500M comes with a host of features aimed at this market segment; reliability and usage statistics tracking, hardware power loss protection (PLP), telemetry monitoring, end-to-end data path protection and multiple namespace management. Talking of namespace management, the DC1500M may use the same controller as its predecessor, but the firmware has been updated so the drive now supports up 64 namespaces, a very useful addition for VM management. Kingston quotes Sequential read/write figures for the 3.84TB DC1500M as up to 3,100MB/s and 2,700MB/s for read and writes respectively. Incidentally, that read figure is the same for three out of the four drives, the 960GB, 3.84TB and 7.68TB. The odd man out is the 1.92TB drive which is the fastest at up to 3,300MB/s. The 1.92TB and 7.68TB have the same 2,700MB/s write rating as the 3,84TB drive with 960GB model down at 1,700MB/s. We confirmed both official figures with our own Sequential tests using a single thread with a peak read test result of 3,485MB/s with writes peaking at 2,855MB/s. Using four threads we got a best read result of 3,480MB/s with writes at 2,881MB/s. Random 4K performance for the 3.842TB drive is quoted as up to 480,000 IOPS for reads and up to 210,000 IOPS for writes with the drive in a steady-state. We tested the drive FOB using our standard 4-threaded tests and got a peak read figure of 454,358 IOPS (QD128). When it came to random write performance, the best we saw was 229,483 IOPS (QD128). Power-wise the DC1500M is a little hungrier than the previous 3.84TB DC1000M for reads; 6.80W idle (5.54W DC1000M), 6.40W for average reads rising to 6.63W maximum (5.31W and 6.10W maximum for the DC1000M). Average writes on the other hand see the latest drive being a bit more frugal than the DC1000M, averaging 14.20W as opposed to14.69W however, when it comes to maximum writes it swings back in favour of the older drive; 16W versus 15.5W. Kingston has given the 3.84TB DC1500M an endurance rating of 1 DWPD (Drive Writes Per Day) over the length of the 5-year warranty Kingston backs the drive with. You can buy the 3.84TB version of the DC1500M on Insight for £718.79 (inc VAT) HERE. Pros Overall Performance. Multiple Namespace support. Endurance. Cons Pricey. KitGuru says: Kingston's DC1500M isn't a leap forward from the previous DC1000M but more of a technology refresh with the move from 64-layer to 96-layer NAND and the addition of support for 64 namespaces while retaining features such as hardware power protection and end to end data protection.
KitGuru KitGuru.net – Tech News | Hardware News | Hardware Reviews | IOS | Mobile | Gaming | Graphics Cards
