On this storage review, we are taking a good look at the HP EX900. An M.2 PCI-E SSD featuring HBM technology first seen on AMD’s high end graphics cards. The EX900 also comes with HP’s own controller which is interesting and worth checking out on this review.
The HP EX900 we’ve got is the 120GB version. While it is the lowest capacity out of the bunch, it comes at a sub 60 USD price point. That’s stupidly good especially if we’re taking the specifications into account.
|Interface||PCI-E x4, NVME 1.3|
|Max. Transfer Rate||1900MB/s (Read), 650MB/s (Write)|
|System Requirements||Windows 7, 8.1, 10|
|Feature||TRIM, ECC, SMART, NCQ, DIS|
|Warranty Period||3 Years / 100 TBW|
Our unit is the more relaxed version of the EX900 at 120GB, with a 1900MB/s read and 650 MB/s write performance. Having read Tom’s Hardware review of the 500GB version, we had ourselves excited over the leaner model’s performance.
PACKAGING AND ACCESSORIES:
The HP EX900 is packed inside the usual M.2 cradle. That said, we’ve got an external carton here and a two part internal shell that houses the drive itself.
Motherboards nowadays with M.2 slots are required to supply their own screws. HP never bothered to include such, but they insisted to provide a simple multilingual user’s manual instead.
DESIGN, LAYOUT AND CONNECTIVITY:
The HP EX900 M.2 SSD is a simple looking fellow. It is, since it resembles your typical 2280 form factor M.2 SSD.
The SM2263XT controller is close to the M.2 connector and it doesn’t feature any heatsinks. Next to it are the four NAND flashes, with 32GB of capacity each.
What I like about this setup though is the single sided design. This allows for better temperature management with motherboards that features a single sided M.2 cooling solution.
TEST SETUP AND METHODOLOGY:
Our storage device reviews revolves around the use of various storage benchmark tools and real world benchmarks. Our setup fills up the test drive to at least 50% of its capacity. This is done so to negate the FOB (Fresh Out the Box) performance of the drive; ensuring that we are testing the drive near its expected usage. The drive is formatted under NTFS and is attached to its natively supported interface. This is to ensure the system is at its optimal testing state.
|TEST SYSTEM SPECIFICATIONS|
|PROCESSOR||Intel Core i5-6600K|
|CPU COOLER||Cryorig C1 Top Flow|
|MEMORY KIT||Crucial Ballistix Tactical @ 2666MHz 4x4GB Kit|
|GRAPHICS CARD||ASUS GTX 1060 STRIX OC 6GB|
|INTERNAL STORAGE||Crucial BX200 480GB|
|POWER SUPPLY||CORSAIR RM850X 850W|
|DISPLAY||27″ DELL U2715H + LG 43UF680T 4K UHD TV|
|OPERATING SYSTEM||Microsoft Windows 10 Pro|
Throughput performance in MB/s is measured with CrystalDiskMark. First up on the test is the Sequential read and write performance, measured with a block size of 1MB, 1GB transfer size and 32 Queue Depth. This test is more in line with large file transfers; similar to watching a movie.
The second one is the Random 4K read and write performance, measured with a random block size of 4KB, 1GB transfer size and 32 Queue Depth. This test is more in line with small file transfers; similar to transferring installation files and reading game data.
The Sequential transfer rate of the HP EX900 is excellent for both read and write – able to achieve 1861 MB/s for the read and 642 MB/s for the write. Random 4K performance is also good at 526 MB/s for the write while the read at 227 MB/s could see an improvement.
Input/Output Operations per Second is is measured with AS SSD. The Random 4K-64Thrd read and write benchmark is used for this test. Performance is measured with a random block size of 4KB, a 1GB transfer size and 64-thread IO requests. This tests the storage medium’s ability to use Native Command Queuing (NCQ) at higher Queue Depth. A Useful metric for server side applications.
IOPS performance is generally good at 147K IOPS for the write, while read performance hovers around the 103K IOPS mark.
ACCESS TIME PERFORMANCE:
The read and write latency is measured with AS SSD using a 512KB block size. Access Time is just as important as the throughput and IOPS performance of the drive; allowing us to peak into how fast or slow a storage medium can access a given data. Latency is measured in milliseconds.
Read latency is recorded at 0.022ms, while the write latency is at 0.045ms. The write performance is synonymous with SATA SSDs.
REAL WORLD PERFORMANCE:
Our real world performance test consists of 3 file folders containing 6GB worth of text files, images and videos each. The files are copied within the drive using TeraCopy to evaluate the storage medium’s performance. File copy performance is measured in seconds.
The HP EX900 struggles to provide us good results on the file copy performance test with its DRAM-less design. Results aren’t bad, just borderline decent.
WRAPPING IT UP:
The HP EX900 is a good alternative to those who are looking for a speedy exit from the usual SATA interface SSDs. Sequential speeds for example are up to HP’s claims. IOPS performance are also excellent for its price point, though we cannot say the same for the mixed load file transfer performance.
While we like how fast it is on paper, the HP EX900 sure had its own caveats as seen on our file transfer performance. Certainly, using this drive as a go to means to transfer movies and other sort of files within are not its strong points. Regardless, this is still an amazing SSD for your operating system or applications that requires speedy access to their own files.
The HP EX900 is now available locally, with our 120GB unit landing at a sweet $60 price point. It sure is not the fastest M.2 SSD around when it comes to mixed loads, but it is a definite upgrade from a SATA SSD without breaking the bank.
HP EX900 M.2 PCI-E SSD
The HP EX900 is now available locally, with our 120GB unit landing at a sweet $60 price point. It sure is not the fastest, but it is a definite upgrade from a SATA SSD without breaking the bank.