Random Performance – Iometer (IOPS/latency), YAPT (random)
We are trying something different here. Folks tend to not like to click through pages and pages of benchmarks, so I'm going to weed out those that show little to no delta across different units (PCMark). I'm also going to group results performance trait tested. Here are the random access results:
Iometer:
Iometer is an I/O subsystem measurement and characterization tool for single and clustered systems. It was originally developed by the Intel Corporation and announced at the Intel Developers Forum (IDF) on February 17, 1998 – since then it got wide spread within the industry. Intel later discontinued work on Iometer and passed it onto the Open Source Development Lab (OSDL). In November 2001, code was dropped on SourceForge.net. Since the relaunch in February 2003, the project is driven by an international group of individuals who are continuesly improving, porting and extend the product.
Iometer – IOPS
The M6e's do great in our Iometer testing. Of particular note is the impressive random read performance shown in our Web Server workload. Latencies are also very good at lower queue depths. The only thing to best them here was a RAID-0 pair of 850 EVOs, until their cache filled up and write speeds dropped (as seen mid way through the Database test).
Iometer – Average Transaction Time
For SSD reviews, HDD results are removed here as they throw the scale too far to tell any meaningful difference in the results. Queue depth has been reduced to 8 to further clarify the results (especially as typical consumer workloads rarely exceed QD=8). Some notes for interpreting results:
- Times measured at QD=1 can double as a value of seek time (in HDD terms, that is).
- A 'flatter' line means that drive will scale better and ramp up its IOPS when hit with multiple requests simultaneously, especially if that line falls lower than competing units.
Impresively low latencies seen by the pair of M6e units.
YAPT (random)
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.
The M6e Black sits high in the SATA pack on writes and easily surpasses all single SATA devices on random reads.
Would be nice to see what
Would be nice to see what real world gains you get from so fast drive .
Agreed. I have asked Allyn
Agreed. I have asked Allyn in the past to start looking at real world benchmarks on SSD’s. While the synthetic benchmarks are good at creating an objective analysis on the strengths and weaknesses between SSD models and manufactures, it begs the question, “So, what? How does that benefit me?”
As a data storage engineer for workstation and HPC environments, I have argued that consumers should first look at their drive size requirements for buying an SSD (due to high cost/GB) and THEN look at specific features and performance (power protection, onboard encryption, high IOPS, etc).
Right now you could buy an inexpensive and high end SSD and they will perform very similar in real world examples: OS boot times and application loading.
I believe it would benefit readers to show that while synthetic benchmarks can show improvements or deficiencies in performance, “fast” is fast enough and that drive size, features and pricing should be the real conclusive factor.
“Plextor also made the M6e
“Plextor also made the M6e available with a half-height PCIe interposer”
On the first page; I don’t think interposer is the right word. It is just an adaptor card. Is plextor calling it an interposer? Interposer has a specific meaning which I don’t think is (or should be) confused with package or pcb.
“An interposer is an
"An interposer is an electrical interface routing between one socket or connection to another."
That's exactly what it is doing – electrically connecting M.2 PCIe to a desktop PCIe socket. I believe a lot of people just want to call it an adapter, but some also call it an interposer. I started calling it that when Intel kept referring to mSATA to SATA adapters as interposers.
Thank you for clarifying.
Thank you for clarifying. There are always a lot of semantic difficulties that I do not wish to needlessly multiply. Do you take interposer to specifically mean pass-through? That is, not bridging or translating different interfaces.
I thought I heard you say in
I thought I heard you say in the podcast that it has a backplate, could you add a photo of the backplate? Thanks.
Oh and nice work, Allyn
No back plate, but maybe I
No back plate, but maybe I was referring to the 'black' PCB?
Heh, yeah you probably was,
Heh, yeah you probably was, my bad.
I always go for the aesthetic look of PCI-e cards, so I want a gorgeous backplate! lol. Right now, ASUS ROG is the most cool looking one although the logo is upside down, bah.
Aside from topic, but Allyn, you might be able to send the message. 😉 I have a gripe on PCI-e cards. I believe most people have tower case(s) and won’t see the card’s face anyway. Even if you have horizontal case, you have to choose which card to show (being the closest to the window). Manufacturers beginning to spice up the tops but they also need to focus more on the backplate! Just sayin.
I’m right there with you, and
I'm right there with you, and that's why I personally use one of those 'inverted' ATX design cases (all the cards are face up).
I’d sacrifice most of that
I’d sacrifice most of that speed for 2x or 4x the capacity.
So who exactly would buy this
So who exactly would buy this over doing a RAID 0 setup? Unless you are running a server requiring high I/O I’m not getting the value add for the enthusiast market…
For this particular Marvell
For this particular Marvell PCIe controller, it's sort of a wash. That's why I included a RAID-0 pair of SSDs as a comparison point. We need NVMe or faster AHCI PCIe SSDs to make it worthwhile over a simple SATA RAID.