Why you can trust Tom's Hardware Our expert reviewers spend hours testing and comparing products and services so you can choose the best for you. Find out more about how we test.
Comparison Products
The Lexar NM1090 Pro, in essence, competes with the original line of high-end PCIe 5.0 SSDs built on Phison’s E26 SSD controller. Although that controller had three “levels” of performance, we’re looking at the top end: the Corsair MP700 Pro SE is a good representative. The NM1090 Pro is more likely to compete with the higher-end options, such as the Sandisk WD_Black SN8100 and Samsung 9100 Pro, on the merits of price rather than performance or even power efficiency. Once Phison moves past the E26 with E28-based upgrades, we expect the drive to also contend with newer drives, particularly on the cost front.
Thanks to its DRAM, though, the NM1090 Pro will have an easy time against E31T-based drives like the Crucial P510, even when they have newer flash. In some ways, it might have to compete with DRAM-less drives, such as the Biwin Black Opal X570, which use Maxio’s unique eight-channel solution, since such drives have high performance but may use older flash to bring the price down. We don’t have enough data to go on for a full comparison there yet, but we expect that DRAM does give the NM1090 Pro the advantage. This means that it will fill a role similar to older E18-based drives – like the Sabrent Rocket 4 Plus-G – which were fast but not quite Black SN850X or 990 Pro fast. This makes the most sense at higher capacities, where you want the storage space and bandwidth but don’t necessarily need something slightly faster that costs more.
Lastly, we wanted to see how far Lexar has come from its super-popular NM790. That drive was always a winner at 4TB for its cost, and the NM1090 Pro can, in some ways, fill the same role. It’s not exactly the same, but if you’re looking for a 4TB drive and are caught between PCIe 4.0 and PCIe 5.0, the comparison here will show you just how much you have to give up. Not shown is the NM1090 Pro’s predecessor, the NM980 Pro, mainly because the InnoGrit IG5236 controller on that drive has since gained a poor reputation. We think the NM1090 Pro’s SMI SM2508 should prove to be more reliable.
Trace Testing — 3DMark Storage Benchmark
Built for gamers, 3DMark’s Storage Benchmark focuses on real-world gaming performance. Each round in this benchmark stresses storage based on gaming activities including loading games, saving progress, installing game files, and recording gameplay video streams. Future gaming benchmarks will be DirectStorage-inclusive and we also include notes about which drives may be future-proofed.
The Lexar NM1090 Pro should match the Acer Predator GM9000, and it does. These two drives are clearly above older PCIe 4.0 drives, as well as the mid-level PCIe 5.0 ones. They feel somewhat like the Phison E18-based drives – such as the Sabrent Rocket 4 Plus-G – did when facing the WD Black SN850X and Samsung 990 Pro. That is, they are high-end but don’t quite feel like the fastest drive out there. Drives like the NM1090 Pro have a bullseye more on the backs of Phison E26-based drives, like the Corsair MP700 Pro SE, because it can now offer a comparable experience with much better power efficiency.
We do think the Sandisk WD_Black SN8100 remains the king for load times as it has the lowest latency in multiple tests, including this one. The BiCS8 TLC flash works excellently with SMI’s SM2508 controller. But you don’t really need a drive of that caliber for games, and if you still want PCIe 5.0 bandwidth, you can settle for something like the NM1090 Pro and get that 4TB of space at a lower price point.
Trace Testing — PCMark 10 Storage Benchmark
PCMark 10 is a trace-based benchmark that uses a wide-ranging set of real-world traces from popular applications and everyday tasks to measure the performance of storage devices. The results are particularly useful when analyzing drives for their use as primary/boot storage devices and in work environments.
The NM1090 Pro fares a bit better than the GM9000 in PCMark 10, but not by enough to make a real difference. The take-home from these results is similar to what we got from the 3DMark benchmark: this drive beats everything that came before it technologically, and it fits the role of being a competitor to the original wave of high-end PCIe 5.0 drives. It doesn’t quite match the MP700 Pro SE here, but it’s fast enough that with its power efficiency and potentially better selling price that it could be a reasonable choice depending on your needs.
Console Testing — PlayStation 5 Transfers
The PlayStation 5 is capable of taking one additional PCIe 4.0 or faster SSD for extra game storage. While any 4.0 drive will technically work, Sony recommends drives that can deliver at least 5,500 MB/s of sequential read bandwidth for optimal performance. In our testing, PCIe 5.0 SSDs don’t bring much to the table and generally shouldn’t be used in the PS5, especially as they may require additional cooling. Check our Best PS5 SSDs article for more information.
Our testing utilizes the PS5’s internal storage test and manual read/write tests with over 192GB of data both from and to the internal storage. Throttling is prevented where possible to see how each drive operates under ideal conditions. While game load times should not deviate much from drive to drive, our results can indicate which drives may be more responsive in long-term use.
Right now, we’re still looking at PCIe 4.0 for console applications, which means that any PCIe 5.0 drive would be overkill. However, not everybody buys a drive to put it in and leave it forever. If you’re buying a drive that will be repurposed later, the NM1090 Pro could potentially make sense, and it is more efficient than the original high-end PCIe 5.0 drives, so it should run cool in this scenario.
Transfer Rates — DiskBench
We use the DiskBench storage benchmarking tool to test file transfer performance with a custom, 50GB dataset. We write 31,227 files of various types, such as pictures, PDFs, and videos to the test drive, then make a copy of that data to a new folder, and follow up with a reading test of a newly-written 6.5GB zip file. This is a real world type workload that fits into the cache of most drives.
DiskBench is one benchmark where we first start to see issues with the NM1090 Pro, specifically focusing on its capacity. The Predator GM9000 has identical hardware at 2TB and doesn’t have the same issues that the 4TB NM1090 Pro does. In theory, these should write at the same speed with both having large enough caches and similar peak write speeds. The NM1090 Pro is the larger drive and will have a larger cache, though, which could impact write consistency.
For larger transfers, the NM1090 Pro should probably perform just fine, but the results here set off some warning bells that force us to analyze the data more closely in our write saturation test. While it’s possible the grade of flash is different on these drives, we suspect instead that Lexar has a different pSLC caching pattern that may or may not be due to the capacity difference. After all, 4TB means 32 dies versus the Predator GM9000’s 16, which means there could be a gap with TLC speeds beyond the cache as well.
Synthetic Testing — ATTO / CrystalDiskMark
ATTO and CrystalDiskMark (CDM) are free and easy-to-use storage benchmarking tools that SSD vendors commonly use to assign performance specifications to their products. Both of these tools give us insight into how each device handles different file sizes and at different queue depths for both sequential and random workloads.
ATTO is useful for pointing out performance nuances of the drive. Here, the NM1090 Pro largely performs as expected with a small dip for 512KiB reads – the same is true for the Predator GM9000 and the MP700 Pro SE. There’s also a large dip at 2MiB for reads, which we also see for the MP700 Pro SE, which uses a different controller but the same flash, but not so much for the Predator GM9000.
It’s not unusual for drives to struggle here due to the nature of newer flash, which has more planes than back in the day, and how things are parallelized with channels and more dies. 4TB is just a harder capacity to hit with Micron flash – we see the same dip with the 4TB Crucial T500 – and that might be related to the reason we see multiplexors on the T500 and the Crucial P510. Signal integrity at higher speeds can be an issue, and that is one reason Phison had to optimize its E26 controller to reliably hit over 14 GB/s. The E28 has no such issue, but in our testing, it’s been using Kioxia/SanDisk BiCS8 flash instead. We would state that Micron’s flash favors capacity and bandwidth more generally, though.
The read problems persist in CrystalDiskMark when we examine QD1 sequential reads. The result isn’t great with the 1MB block size, as the NM1090 Pro falls behind even the mid-range P510. With more queue depth or with writes, there are no problems. 4KB QD1 random read and write latencies fare better, as the NM1090 puts up a good showing, with the main winners being the E28 engineering sample and the SN8100. Those two drives use the aforementioned BiCS8 TLC flash, which has proven to provide excellent latency. We’ve speculated in the past that this is likely due to optimization for the four-plane, rather than six-plane, flash, making it more agile than YMTC’s and Micron’s.
When Kioxia/Sandisk backports this approach for 112-Layer BiCS5 with the upcoming BiCS9 flash, which would have more flexibility with the smaller 512Gb dies, we might see some surprising results. Traditionally, bit-cost scalable (BiCS) flash has been designed to achieve the best density per dollar, but the six-plane flash of Micron and YMTC favors capacity even more, with the extra planes also offering higher bandwidth potential, which is a factor if you also want to achieve top sequential speeds. The 1TB Crucial T710 specifications reinforce this last point. However, enthusiasts tend to favor low latency, and many markets want or need smaller drives, so there’s room here for both approaches. As far as the NM1090 Pro goes, it performs well enough for its niche, but there are definite weak points from its choice of flash, and that’s something to remember if you’re shopping PCIe 5.0 drives.
Sustained Write Performance and Cache Recovery
Official write specifications are only part of the performance picture. Most SSDs implement a write cache, which is a fast area of pseudo-SLC (single-bit) programmed flash that absorbs incoming data. Sustained write speeds can suffer tremendously once the workload spills outside of the cache and into the "native" TLC (three-bit) or QLC (four-bit) flash. Performance can suffer even more if the drive is forced to fold, which is the process of migrating data out of the cache in order to free up space for further incoming data.
We use Iometer to hammer the SSD with sequential writes for 15 minutes to measure both the size of the write cache and performance after the cache is saturated. We also monitor cache recovery via multiple idle rounds. This process shows the performance of the drive in various states as well as the steady state write performance.
The 4TB NM1090 Pro initially writes in the temporary and fast single-bit pSLC mode at an average of over 12.5 GB/s for 65 seconds. The cache is about 814GB in size, which is large but not the largest possible for a drive of this size. It’s more than ample. It’s also still small enough that the drive can hit a direct-to-TLC mode of over 6.6 GB/s, which is the fastest we’ve ever seen any drive write outside of pSLC, with the Sabrent Rocket 5 Plus coming in second. We note here that, technically, flash can operate in multiple modes, such as 2-bit pMLC, and the 4-bit QLC on the Micron 2600 also has a 3-bit pTLC mode. Likewise, 5-bit PLC could operate in these other modes, with Solidigm enterprise drives operating that flash as if it were QLC.
We point that out because the NM1090 Pro’s performance here could be mistaken for pMLC but, given the sudden drops to about 2.4 GB/s with a steady state also at that speed we can guess it is being forced to fold – that is, to copy data from the cache to the native flash to free up space in a way that slows down incoming writes – which plausibly allows Lexar to have this unique pattern on the 4TB SKU. Typically, you would expect something more like the Predator GM9000’s result, with an average write speed around 3.9+ GB/s as the ceiling for this flash. However, we know it can do more, with the 2TB Sabrent drive mentioned above managing 4.45 GB/s. The result here is faster but less consistent writes, which, actually, isn’t necessarily a bad thing. A 4TB drive of this caliber – one that’s fast, but not the fastest – could hold up very well with a large transfer.
Power Consumption and Temperature
We use the Quarch HD Programmable Power Module to gain a deeper understanding of power characteristics. Idle power consumption is an important aspect to consider, especially if you're looking for a laptop upgrade as even the best ultrabooks can have mediocre stock storage. Desktops may be more performance-oriented with less support for power-saving features, so we show the worst-case.
Some SSDs can consume watts of power at idle while better-suited ones sip just milliwatts. Average workload power consumption and max consumption are two other aspects of power consumption but performance-per-watt, or efficiency, is more important. A drive might consume more power during any given workload, but accomplishing a task faster allows the drive to drop into an idle state more quickly, ultimately saving energy.
For temperature recording we currently poll the drive’s primary composite sensor during testing with a ~22°C ambient. Our testing is rigorous enough to heat the drive to a realistic ceiling temperature.
The NM1090 Pro lives up to our expectations that the drive would be more efficient than earlier drives. In this case, the drive is almost 40% more efficient than the MP700 Pro SE, a considerable boost. It’s not as efficient as drives with Micron’s newest TLC flash – that would be the P510 on this chart, as well as the 2TB T710, which uses the same controller as the NM1090 Pro. Having more flash on a 4TB SKU can lower power efficiency in many cases, but we know the SM2508 controller can achieve better results, particularly with the BiCS8 flash on the SN8100. Still, we’d say this is “efficient enough,” especially with the improvements to idle power consumption.
Temperature-wise, the NM1090 Pro uses multiple sensors that return different readings. Generally, you will have sensors to gauge at least the controller temperature and sometimes the NAND flash as well. A third reading could be more general for the PCB temperature, which could take DRAM into account, or it may just be an ambient reading. At peak, we saw 81°C, 74°C, and 78°C.
Our expectation is that the first is for the controller, and while that seems high, the ARM-based ASICs for consumer SSDs have a junction maximum around 125°C. However, other components have a narrower operation range, with consumer flash usually topping out around 75°, for instance. Flash can handle higher temperatures – we’ll avoid a deep discussion on how this impacts things – but this is given with consideration to the ambient. As a result, SSDs usually throttle based on a composite temperature, which in this case was probably the second number.
At 74°C, the NM1090 Pro only has about 10 degrees of headroom before throttling, so we would recommend a heatsink or other cooling solution. This is a double-sided drive with 4TB of flash, so it will run hotter. In a PCIe 4.0 slot, though, it should have no problems whatsoever. Our temperature monitoring comes from our sustained write test, which, to be fair, is realistic for usage on a high-end drive like this, but in normal use, this drive won’t overheat except in cases of high ambient temperatures.
Test Bench and Testing Notes
CPU | Row 0 - Cell 2 | |
Motherboard | Row 1 - Cell 2 | |
Memory | Row 2 - Cell 2 | |
Graphics | Intel Iris Xe UHD Graphics 770 | Row 3 - Cell 2 |
CPU Cooling | Row 4 - Cell 2 | |
Case | Row 5 - Cell 2 | |
Power Supply | Row 6 - Cell 2 | |
OS Storage | Row 7 - Cell 2 | |
Operating System | Row 8 - Cell 2 |
We use an Alder Lake platform with most background applications such as indexing, Windows updates, and anti-virus disabled in the OS to reduce run-to-run variability. Each SSD is prefilled to 50% capacity and tested as a secondary device. Unless noted, we use active cooling for all SSDs.
Lexar NM1090 Pro Bottom Line
The Lexar NM1090 Pro performs in many ways exactly like the Acer Predator GM9000, which is not exactly surprising given that both drives have the same hardware. There are some differences, though, and we speculate they derive from the difference in capacity. We don’t hold this against the NM1090 Pro as we have not tested many 4TB high-end PCIe 5.0 drives yet, but time marches on: newer drives are being announced every day, and this drive isn’t quite as impressive as the GM9000 felt even a few months ago. With Phison E28 drives on the horizon, we have even more difficulty in giving the NM180 Pro a high score, even though, on the whole, it does just fine for its spot in the market.
That spot would be a “budget” high-end PCIe 5.0 drive that excels at higher capacities and a lower price point than the somewhat faster competition. Given its mediocre power efficiency by today’s standards, it remains difficult to recommend it as a primary drive, though. The Sandisk WD_Black SN8100, for one, is just a better drive. If you’re going for a full-fledged PCIe 5.0 drive as your main drive, why not spend a little more for the best? It could make sense as a budget option if you’re coming from PCIe 3.0 on a new build, but we still think it makes the most sense as a secondary game and data drive. In that it succeeds as it offers more than lower-end PCIe 5.0 drives, including venerable ones based on Phison’s E26 controller like the Corsair MP700 Pro SE. This is perfectly fine if you have a new system that can handle lots of fast storage.
It’s true that the NM800 Pro filled a similar role in some ways, using the less expensive InnoGrit IG5236 controller, which allowed it to butt heads with E18-based drives like the Sabrent Rocket 4 Plus-G. However, in this case, the SM2508 is clearly superior to the E26 and the NM1090 Pro, instead lagging behind with its flash. This is a different sort of trade-off, particularly when we know what BiCS8 can do. You can get the raw sequential performance out of this drive, but it still feels half a generation behind in some ways. Given the expensive nature of drives in this class, this forces the drive into a semi-budget position, which, again, makes the performance gap an understandable and acceptable trade-off if you aren’t gunning for the very best.
MORE: Best SSDs
MORE: Best External SSDs
