The OCZ Trion 150 SSD Review
by Billy Tallis on April 1, 2016 8:00 AM ESTFinal Words
Unlike for processors, we've come to expect bad things from a process shrink for flash. It helps lower costs but often hurts endurance and performance. Until recently the performance penalty was mostly a matter of reduced parallelism from higher-capacity NAND flash dies, and as such the penalty could be offset by simply getting a bigger drive. With the value SSD market now dominated by TLC, die shrinks also bring an increased reliance on error correction–and more advanced error correcting schemes like LDPC are much slower.
Given the above, we're very glad to see that the Trion 150 only performs below the Trion 100 on a few tests (most notably, random read speed). I suspect that Toshiba's 15nm flash process was designed with very careful attention to mitigating the disadvantages of TLC flash where possible, but the drive's firmware also deserves a lot of credit. The steady-state performance consistency behavior of the Trion 150 is completely different, vastly better than the Trion 100 and better than a lot of mid-range SATA drives. Most other tests show at least moderate performance improvement relative to the Trion 100. Power efficiency has also improved, though not enough to prevent the improved performance from pushing overall power consumption over 5W at times. Overall the Trion 150 has no trouble proving its worth as an upgrade from a hard drive, and it's better-suited to that purpose than its predecessor.
Several tests showed a marked difference in behavior between the 240GB Trion 150 and the larger capacities, with the 240GB drive sometimes dramatically outperforming or underperforming the othe two sizes. These differences mostly washed out and the results on our AnandTech Storage Bench tests of real-world access patterns showed no such surprises. It would be interesting to know what causes the different behaviors, but none of those results are cause for concern.
When it was first announced, we expected the Trion 150 to be the end of the road for Toshiba's planar TLC SSDs. The race to the bottom has fortunately not kept up quite the same pace with this latest product cycle. While Toshiba is still certainly trying to get their 3D NAND out the door as soon as possible, the Trion 150 shows that their 15nm TLC is not as unsatisfying as we expected. We've also seen Samsung introduce a low-end planar TLC SSD as a cheaper alternative to their 3D NAND options, and companies are continuing their planar NAND R&D efforts alongside 3D NAND development. If another die shrink can be pulled off like the 15nm transition, we might see one more generation of mainstream SSDs using planar flash, though only in certain market segments.
However, for all that the Trion 150 didn't live up to our fears and turned out to be pretty good for a sub-20nm planar TLC drive, it also did nothing to significantly close the performance gap with MLC drives. This means the price still needs to be going down to create a meaningful separation in price tiers between TLC and cheap MLC drives. Aside from a $20 rebate for the 480GB Trion 150 on Newegg, it's not currently priced aggressively, but it's definitely a drive to watch. Any time a sale makes it the cheapest option, it would be the best buy among low-end TLC drives. Against competitors like the ADATA SP550 or PNY CS1311, it can only command a few dollars premium.
| Value SSD Price Comparison | ||||
| Drive | 960GB | 480GB | 240GB | 120GB |
| ADATA SP550 | $217.99 | $109.99 | $58.99 | $38.49 |
| PNY CS1311 | $219.99 | $114.99 | $59.99 | $39.99 |
| OCZ Trion 100 | $199.99 | $139.99 | $59.99 | $39.99 |
| OCZ Trion 150 | $229.99 | $117.49 | $61.99 | $45.99 |
| Crucial BX200 | $239.99 | $119.99 | $63.88 | |
| SanDisk Ultra II | $199.99 | $124.25 | $74.99 | $52.90 |
Facebook
Twitter
RSS
79 Comments
View All Comments
Arnulf - Friday, April 1, 2016 - link
So not only is this a low-end drive, it's an OCZ low-end drive.Thanks for the article and giving consumers heads-up so that we don't get burned!
Arnulf - Friday, April 1, 2016 - link
Ugh, EDIT function pretty please ...So not only is this a low-end drive, it's an overpriced OCZ low-end drive.
close - Friday, April 1, 2016 - link
I wouldn't call this a "burn" but it's definitely a warm beer after a hot summer day.LB-ID - Friday, April 1, 2016 - link
<Wanders in, sees another OCZ drive review><Reads the review, sees nothing has really changed>
<Wanders out to buy from a reputable brand>
Samus - Saturday, April 2, 2016 - link
I now feel like Toshiba is a Trojan horse, because all the OCZ drives that were in the buffer during bankruptcy and released after Toshiba initially bought them, in particular the exceptional ARC 100, were a sign they were turning things around.Then Toshiba releases this crap under the OCZ name. WTF are they thinking OCZ is a performance focused brand and these drives are clearly marketed at grandma's old Dell laptop.
xrror - Wednesday, April 6, 2016 - link
Yea but I think Toshiba's plan was to use OCZ for ALL of their customer facing drives. Before they bought OCZ they simply didn't offer retail drives at all!As far as a Toshiba Trojan horse... Toshiba actually had a competent drive itself right before it bought OCZ.
The Toshiba THNSNH drives were actually good (they don't bench awesome, but they're good) but since they were OEM only you've never heard of them. THNSNH was also a big jump from Toshiba's prior drives that seemed mostly made for industrial (performance wasn't great, but HDD wouldn't handle vibration/shock).
Bonus that Toshiba kept OCZ's in development models, so you had the ARC series come out. (vs. something ... not so nice like continuing Octane). But those were Marvell (I think) based with Indilinx firmware? Nothing wrong at all with that, but I imagine Toshiba's shareholders would question why they weren't using native Toshiba tech...
I'm actually really glad Toshiba of anyone (also, a surprise!) was able to snag the OCZ engineering team. Yea everyone loves to hate on OCZ, but OCZ really was the first company to so extremely aggressively bring SSD drives to the consumer market. Everyone likes to hate on Ryan Petersen but damnit, he really DID push to make SSD drives an affordable "thing" much MUCH sooner than just Intel alone would have.
They had a really rough time pounding out the bugs with SandForce yes... but seriously - who could have predicted that? (no, SERIOUSLY - before the kneejerk reaction, imagine being in OCZ's development shoes - that had to be really tough cause each time they thought they had it fixed, and it was always something very subtle). I always find it interesting that everyone forgets that their Indilinx controllers just worked - but they weren't the "sexy" fastest.
Saying that, I totally understand people who... after forking out their $600+ on a drive and having failure after RMA failure after "no this firmware really fixes it honest" failure upon failure would yea... swear off OCZ forever. But today's OCZ isn't that. Different time, place, era.
This comment has gone way too long. And I've lost my point lol.
And so you all don't think I'm a forgive and forget person. I'll never buy another Kingston product after the asshattery with the V300. When a company basically tells Anand himself to piss off when they bait and switch the NAND used...
http://www.anandtech.com/show/7763/an-update-to-ki...
Hey, at least Ryan Peterson and OCZ back then was super responsive to Anand directly when they got ripped a new one ;)
xrror - Wednesday, April 6, 2016 - link
Ugh, and I just realized I didn't address the part of using the OCZ name for non-performance parts.Basically, unless you're Samsung right now nobody is a performance part. And so that doesn't sound like a hipster "well, duh" answer - Samsung went crazy (awesome!) insane mad R&D into controllers and 3D NAND. Holy moly they even kicked Intel's butt enough so that Intel announces vaporware crosspoint memory.
I mean, halo effect in SATA market at least is crazy! SanDisk has some very extremely competitive models in their high end - but because they're maybe 3% slower in some benchmarks? Nobody on the street cares. That's nuts.
I dunno, I guess my point is that unless Toshiba/OCZ can somehow squeeze 12Gb/s through a SATA 3 link (I joke) they're not going to be able to market as a performance product these days.
Actually, the killer solution I think is to make an SSD controller that is both power efficient and makes TLC NAND not suck (so much). And that's not easy. At all. Cause small process planar TLC really a physics nightmare. I mean that's taking crappy flash drive memory and getting it to perform. Not easy.
leexgx - Sunday, April 24, 2016 - link
maybe i been unlucky i had to many sandisk SSDs (3) and SD (more than 5) cards fail (failing as what is in my phone right now about to funny enough change it for a 32GB EVO+)i do wish they would make more SSDs like the BX100 its the most power effect SSD i have ever seen (the BX200 and MX200 the very bad) BX100 is not very fast for a SSD but in laptops its very low power
jasonelmore - Saturday, April 2, 2016 - link
Still mad about that OCZ rebate scandal i see.Leyawiin - Sunday, April 3, 2016 - link
No, its a Toshiba low end. Do try and keep up.