Intel SSD 750 Series 1.2TB PCIe and 2.5" SFF Review - NVMe for the Consumer
Sequential Performance - HDTach, HDTune, File Copy, YAPT (sequential)
We have shifted over to combining our results into two groupings for consumer reviews. First up is sequential performance:
HD Tach will test the sequential read, random access and interface burst speeds of your attached storage device (hard drive, flash drive, removable drive, etc). All drive technologies such as SCSI, IDE/ATA, 1394, USB, SATA and RAID are supported. HDTach tests sequential performance by issuing reads in a manner that was optimized more for HDD access, but this unique method has proven useful in evaluating the sequential response time of SSDs. The accesses are relatively small in size (2k), and are issued with a single working thread (QD=1). The end result is that devices with relatively IO high latency will not reach their ultimate rated speed.
Looks like Intel just punched everyone else in the gut to start off this fight. HDTach is not an ideal sequential read test for an SSD, as it performs very small read commands that are issued sequentially (rather than queued). Despite this, the SSD 750 turns in the highest figures we've ever seen in this test.
HDTune tests a similar level of features as compared with HDTach, but with a different access pattern. Thus provides us with an additional set of benchmark numbers to compare between storage configurations. CPU utilization has proven negligible with modern processing horsepower, and is no longer included. Additionally, we do not include write performance due to HDTune's write access pattern not playing nicely with most SSDs we have tested it on.
We're not sure why HDTune has become as inconsistent as it has with these faster SSDs, but it appears to not mesh well with NVMe devices. Speeds are still good, but we know the SSD 750 is much more capable than what this test reports. Included here mostly as a data point, but we will be discontinuing use of this test in the near future.
PCPer File Copy Test
Our custom PCPer-FC test does some fairly simple file creation and copy routines in order to test the storage system for speed. The script creates a set of files of varying sizes, times the creation process, then copies the same files to another partition on the same hard drive and times the copy process. There are four file sizes that we used to try and find any strong or weak points in the hardware: 10 files @ 1000 MB each, 100 files @ 100 MB each, 500 files @ 10 MB each and 1000 files at 1 MB each.
Intel cleans house on file creations right until the 1000 x 1MB file creation, where it slows a bit. This is because its enterprise pedigree heavily optimizes for 4k and higher random access, and writing very small files with our tool means a lot of <4k file table updates, slowing things down a bit.
Overall the fastest we've ever seen the copy test complete. Watching the batch file run was like listing a large directory at the command line. It simply flew by. Since writes during a copy operation can be cached by Windows, the small file creation slow down we saw above no longer had an impact on the SSD 750.
YAPT (yet another performance test) is a benchmark recommended by a pair of drive manufacturers and was incredibly difficult to locate as it hasn't been updated or used in quite some time. That doesn't make it irrelevant by any means though, as the benchmark is quite useful. It creates a test file of about 100 MB in size and runs both random and sequential read and write tests with it while changing the data I/O size in the process. The misaligned nature of this test exposes the read-modify-write performance of SSDs and Advanced Format HDDs.
YAPT has always done a great job of maxing out SSDs, and the same applies here. The SSD 750 takes the crown with nearly 2.6 GB/sec reads...
...but falls short of the Phoenix Blade on sequential writes. This is because Intel was more conservative in its write performance in the interest of lower per-IO latency. This means that overall IOPS of the SSD 750 is higher in mixed workloads as it is not saturating itself with writes.