TurboWrite: MLC Performance on a TLC Drive

All NAND trends towards lower performance as we move down to smaller process geometries. Clever architectural tricks are what keep overall SSD performance increasing each generation, but if you look at Crucial's M500 you'll see that it's not always possible to do. Historically, whenever a level of the memory hierarchy got too slow, the industry would more or less agree to insert another level above it to help hide latency. The problem is exascerbated once you start talking about TLC NAND. Samsung's mitigation to the problem is to dedicate a small portion of each TLC NAND die as an SLC write buffer. The feature is called TurboWrite. Initial writes hit the TurboWrite buffer at very low latency and are quickly written back to the rest of the TLC NAND array.

Since the amount of spare area available on the EVO varies depending on capacity, TurboWrite buffer size varies with capacity. The smallest size is around 3GB while the largest is 12GB on the 1TB EVO:

Samsung SSD 840 EVO TurboWrite Buffer Size vs. Capacity
  120GB 250GB 500GB 750GB 1TB
TurboWrite Buffer Size 3GB 3GB 6GB 9GB 12GB

I spent some time poking at the TurboWrite buffer and it pretty much works the way you'd expect it to. Initial writes hit the buffer first, and as long as they don't exceed the size of the buffer the performance you get is quite good. If your writes stop before exceeding the buffer size, the buffer will write itself out to the TLC NAND array. You need a little bit of idle time for this copy to happen, but it tends to go pretty quickly as it's just a sequential move of data internally (we're talking about a matter of 15 - 30 seconds). Even before the TurboWrite buffer is completely emptied, you can stream new writes into the buffer. It all works surprisingly well. For most light use cases I can see TurboWrite being a great way to deliver more of an MLC experience but on a TLC drive.

TurboWrite's impact is best felt on the lower capacity drives that don't have as many NAND die to stripe requests across (thus further hiding long program latencies). The chart below shows sequential write performance vs. time for all of the EVO capacities. The sharp drop in performance on each curve is when the TurboWrite buffer is exceeded and sequential writes start streaming to the TLC NAND array instead:

On the 120GB drive the delta between TurboWrite and standard performance is huge. On the larger drives the drop isn't as big and the TurboWrite buffer is also larger, the combination of the two is why the impact isn't felt as muchon those drives. It's this TurboWrite buffer that gives the EVO its improvement in max sequential write speed over last year's vanilla SSD 840.

Endurance: Not a Problem Even at 19nm RAPID: PCIe-like Performance from a SATA SSD
POST A COMMENT

131 Comments

View All Comments

  • halbhh2 - Saturday, July 27, 2013 - link

    Exactly. I found that even moving up from 8GB to 16GB had a great effect for me with an old Samsung F3 hard drive. The difference: after just 30 minutes from boot, loading an often used program like iTunes (for podcasts) was very similar to the speed on my laptop which has an 830 SSD and only 4GB. Both load in about 4 seconds, and the 16GB desktop loads so fast because it has had time to cache a lot of iTunes. Before the ram upgrade, that load time for iTunes on the desktop computer was about 14 seconds. Quite a difference due to windows 7 caching. The extra improvement I'd get from installing an SSD onto the desktop computer now would be modest, since I usually only need to reboot once or twice in a week. Still, the sweet spot of price/performance for me is approaching, probably around $60-$70, and that won't be long. Reply
  • Klimax - Sunday, July 28, 2013 - link

    It's in wrong place. Unlike OS level caching (at least in Windows), which is in cooperation between memory manager, cache manager and file system driver, this is too low in the chain and sees only requests but nothing else and also takes memory from OS and takes too few. Reply
  • Coup27 - Thursday, July 25, 2013 - link

    Typo: "although I wouldn't recommend deploying the EVO in a write heavy serve Microsoft's eDrive standard isn't supported at launch"

    Excellent article. Samsung continue to push SSDs and I'm really excited about RAPID. Is the 840 Pro due for a successor any time soon? I am selling my current ATX Sandy Bridge + 830 and getting a mITX Haswell + (840 Pro?) and want the fastest Samsung consumer SSD available and I'd be gutted to buy an 840 Pro to see it's successor released a few weeks later.
    Reply
  • vLsL2VnDmWjoTByaVLxb - Thursday, July 25, 2013 - link

    Another typo last page:
    "Even though its performnace wasn't class leading, it was honestly good enough to make the recommendation a no-brainer. "
    Reply
  • JDG1980 - Thursday, July 25, 2013 - link

    Will there be an 840 EVO Pro coming out later? To me, TLC is still a deal-breaker.
    By the way, what happens if power goes out during a TurboWrite (before the data has been written to the normal storage space)? Does this result in data loss, or, worse, bricking? I'd suspect Samsung at least avoided the latter, but I'd like to see some confirmation on this.
    Reply
  • sherlockwing - Thursday, July 25, 2013 - link

    I guess you didn't read the Endurance part of the review? Even if you write 100GiB a day all of those drives last longer than their warranty( 3 years), that's more than enough endurance. Reply
  • Coup27 - Thursday, July 25, 2013 - link

    Some people just don't want to accept the facts. TLC could get to 99.9% of MLC endurance and people would still want MLC. I've been deploying 840's in a light duty enterprise environment and they've been fine. The only reason I use MLC at home is because I want the absolute fastest performance and I can afford it, not that I actually need it. Reply
  • Oxford Guy - Thursday, July 25, 2013 - link

    The SanDisk Extreme 240 was just on sale for $150. TLC NAND still seems like a solution in need of a problem. Reply
  • Spunjji - Friday, July 26, 2013 - link

    You can approach TLC pricing with an MLC drive in a sale, but the fact remains that when it comes to actual sustainable production pricing TLC NAND has a 50% density (and thus manufacturing cost advantage) over MLC. Given that NAND price determines drive cost and drive costs are the primary barrier of entry to SSDs, I'm fairly sure it has a problem to solve.

    FWIW I have not seen any drive touch the 120GB 840's price here in the UK, on sale or otherwise.
    Reply
  • Oxford Guy - Friday, July 26, 2013 - link

    However, there is also the problem of increasing latency and lifespan from node shrinkage. Reply

Log in

Don't have an account? Sign up now