Introduction

ADATA is Taiwan's largest manufacturer of flash storage and DRAM memory for computers. They have been at the forefront of SSD development for many years, bringing us famous SSDs such as the SX8200, Gammix S70 and Atom 50.



Back in August, we reviewed the ADATA Legend 970 2 TB, which was ADATA's first PCI-Express 5.0 solid-state-drive, built on the Phison E26 controller which is used on virtually all consumer Gen 5 SSDs on the market. Today we have the Legend 970 Pro on our test bench, which is designed around Innogrit's new PCIe Gen 5 controller. This makes ADATA one of the first vendors to feature multiple controller vendors in their Gen 5 lineup. Not long ago, we tested the Team Group GE Pro which uses the same IG5666 controller, but NAND flash from YMTC. On the ADATA Legend 970 Pro, 3D TLC NAND flash from Micron is used—the same 232-layer model as on most Phison E26 drives. As expected, a DRAM cache is included, too. What makes the Legend 970 Pro stand out is its cooling design. Virtually all Gen 5 SSDs come with an active fan-cooled cooling solution which requires external power, either through a SATA power connector or a PWM fan header from the motherboard. On the Legend 970 Pro, ADATA came up with a great alternative—the fan is powered by the SSD directly, which eliminates all the cable mess.

The ADATA Legend 970 Pro is available in capacities of 1 TB (price unknown), 2 TB ($300) and 4 TB (price unknown). The endurance is set to 740 TBW, 1480 TBW and 2960 TBW respectively. ADATA includes a five-year warranty with the Legend 970 Pro SSD.

Packaging

The Drive

The drive is designed for the M.2 2280 form factor, which makes it 22 mm wide and 80 mm long.


PCI-Express 5.0 x4 is used as the host interface to the rest of the system, which doubles the theoretical bandwidth compared to PCIe 4.0 x4.


On the PCB you'll find the controller and four flash chips, two DRAM cache chips are included, too.


The cooling solution is preinstalled, it comes with a little fan inside.


While most Gen 5 SSDs require external power, ADATA engineered a voltage conversion circuit on their PCB, which generates the 12 V needed for the fan from 3.3 V, using a step-up-controller.


The PCB is sandwiched between thermal tape on both sides to facilitate heat transfer.

Chip Component Analysis

The Innogrit IG5666 is the company's first client-focused PCI-Express Gen 5 controller. It comes with support for eight channels and is produced using a 12 nm process at TSMC Taiwan. Internally, the controller has eight Arm Cortex-R5 cores with NVMe 2.0 support.


The four flash chips are Micron 232-layer 3D TLC NAND. Each chip has a capacity of 512 GB.


Two Samsung DDR4-2666 chips provide a total of 2 GB of fast DRAM storage for the controller to store the mapping tables.

Test Setup

Synthetic Testing

• Tests are run with a 20-second-long warm-up time (result recording starts at second 21).
• Between each test, the drive is left idle for 60 seconds, to allow it to flush and reorganize its internal data.
• All write requests contain random, incompressible data.
• Disk cache is flushed between all tests.
• During these tests, M.2 drives are tested with additional active fan-cooling, to ensure thermal throttling can't happen

Real-life Testing

• After initial configuration and installation, a disk image is created; it is used to test every drive.
• Automated updates are disabled for the OS and all programs. This ensures that—for every review—each drive uses the same settings, without interference from previous testing.
• Our disk image consumes around 700 GB—partitions are resized to fill all available space on the drive.
• All drives are filled with random data to 85% of their capacity. This is intentional, to run the drive in realistic operating conditions—nobody uses a nearly-empty SSD in their system. It also puts additional stress on the pseudo-SLC cache subsystem, because there is less free NAND space to work with.
• Partitions are aligned properly.
• Disk cache is flushed between all tests.
• In order to minimize random variation, each real-life performance test is run several times, with reboots between tests to minimize the impact of disk cache.
• All application benchmarks run the actual application and do not replay any disk traces.
• Our real-life testing data includes performance numbers for a typical high-performance HDD, using results from a Western Digital WD Black 1 TB 7200 RPM 3.5" SATA. HDDs are significantly slower than SSDs, which is why we're not putting the result in the chart, as that would break the scaling, making the SSDs indistinguishable in comparison. Instead, we've added the HDD performance numbers in the title of each test entry.
• During these tests, M.2 drives are tested with additional active fan-cooling, to ensure thermal throttling can't happen