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
Despite having three CPU cores dedicated to background tasks, the single remaining core just doesn't scale to very high IOPS whenever writes are present (all but the Web Server test). This means that for multi-threaded workloads including writes, the Phison S10 controlled SSDs turn in results roughly half of the competing units.
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.
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.
This test has no regard for 4k alignment, and it brings many SSDs to their knees rather quickly. That said, the Phison S10 performed surprisingly well with misaligned IO – especially misaligned writes. Modern operating systems will align larger accesses to 4k boundaries, but smaller IO (file table writes, etc) may not be, and Phison's firmware appears to handle this better than others. This does come at some cost though, evidenced by the HDTach / HDTune results we saw earlier in this review.
“They differ in flash memory
“They differ in flash memory types used…”
Which reminds me why Kingston is still off of my consideration list.
Same here, nearly 90%
Same here, nearly 90% kingstons drive I have sold to this date have failed within 1hr of installation into clients machines. Its better to Avoid Kingston SSDs like a plague.
I have been using a pair of
I have been using a pair of Kingston HyperX 3K ssds in raid 0 as my primary drive and i havnt had a single issue with them.
Hey Allyn, i have a comment
Hey Allyn, i have a comment on your graphs. When using the graphs with a higher number of data sets, can you change up the shapes of the data points for similarly colored lines to better distinguish them? For example, the red line from the slightly-different-red line (Patriot vs Crucial), and the green line from the slightly-different-green line (Corsair vs Sandisk), etc.
Understood. I’ll try and see
Understood. I'll try and see if I can find a more contrasting / differing color for those two. What is now the white line used to be a *third* red/pink line :).
Allyn,
This isn’t so much a
Allyn,
This isn’t so much a complaint but a suggestion, as I’m certainly a fan of the critical work you do for PCper… and your dead-pan (sometimes pleasantly cheesy) sense of humor is *always* appreciated on the podcast.
Anyway, you should get Ken to brain storm idea for a new charting/color system as the current “Rainbow Edition” is a bit hard to follow sometimes, especially when the data are heavy populated. Thanks Allyn!
SATA SSDs don’t interest me
SATA SSDs don’t interest me as much since NVMe drives have become available.
It’s like the new hot neighbor that moves in next door. She has the looks AND the smarts. Now if only the price was right…
Sorry… eh… uh… thanks, Allyn! Nice article!
I wish I had a new, hot,
I wish I had a new, hot, neighbor, but I’ve always lived in the woods 🙁
It is unfortunately quite
It is unfortunately quite expensive, but there are sets of security bits that go down to T6.
http://www.walmart.com/ip/EAZYPOWER-86155-Security-Screwdriver-Bit-Set-30-Pc/40923636
Alternatively, it might be easier to just drill out a cheap T6 bit or use a screw extractor if you have the tools.
Allyn I see the Sandisk Ultra
Allyn I see the Sandisk Ultra 2 in your charts.
Have you reviewed that drive?
@chaitanya
90% failure means
@chaitanya
90% failure means to me something isn’t installed correctly.
Ever heard of “Devsleep”?
Bottom-line, if your power-supply has a fifth wire (usually orange)
going to the drive, it’ll go into a coma and never wake up.
I even found this hard to swallow, I mean, powering down the computer and restarting it would help, right?
Intel and Sandisk, inventors of the flawed technology,
quietly released a firmware update to deal with it (I’ve read).
Everyone else should use an adapter most motherboards include, that eliminates 3.3v.
Intel, al-la Queen-Biggie, said “let them eat signal instead of 3.3v”.
Don’t know if that’s your 90% in an hour, but it’s looking like a suspect.
You put corsair on title but
You put corsair on title but dont put corsair images on same text? thumbs down!