Well... Smart Response is... Better than nothing. I actually use it alongside a RAID-1 (mirroring) volume that I have as my "users" partition, and it helps in many cases, but it isn't ideal in many others (and does NOT have performance like a real SSD installation).
Smart Response, unlike Apple's Fusion Drive system, is essentially a fancy cache. It has two major "components", read and write caching.
For write caching, in the normal "enhanced" mode (which is really all you should use as the Max mode has other issues), it is a write-through cache. This means that any data written to the volume must hit both the SSD cache AND the HDD itself before the write is released. This has some benefits, as changes are guaranteed to always be written to the HDD in full, so if the SSD fails, your volume isn't toast (just not accelerated anymore). However, this obviously heavily negates one of the main benefits of an SSD! All writes still go to the HDD, and are still subject to regular HDD latency. There's really very limited benefit from Smart Response directly when it comes to writes. The only reason it actually writes to the SSD at all is to pre-build the cache for future Read operations (since the write obviously touches a "recently used" block, it would want to cache it anyway for future read purposes).
Smart Response is primarily a read-caching system. Now, of course, having that Read cache can, itself, help with writes, because reads happening simultaneously can (if they've been cached) hit the SSD alone and not touch the spinning disk. As
AnandTech puts it: "Note that you may still see an improvement in write performance vs. a non-cached hard drive because the SSD offloading read requests can free up your hard drive to better fulfill write requests."
But, to be clear, write performance in an Enhanced-mode Smart Response setup
is limited by the performance of the spinning disk.
The read caching system is much easier to understand... I'll let Anand explain again:
Intel's SRT functions like an actual cache. Rather than caching individual files, Intel focuses on frequently accessed LBAs (logical block addresses). Read a block enough times or write to it enough times and those accesses will get pulled into the SSD cache until it's full. When full, the least recently used data gets evicted making room for new data.
That's pretty straightforward, and works pretty well. It basically caches frequently accessed logical blocks to the SSD, which then can be read directly from the SSD on future reads. That works pretty well, though it is limited to 64GB.
What this means is, that for common use-cases, Smart Response will make read-heavy operations that you do frequently (primarily launching a commonly used application) feel nearly indistinguishable from an SSD.
However, write-heavy operations (which often includes booting and shutting windows down, and certainly includes accessing MC's Library) have essentially no performance enhancement. The problem is, that the way we usually use our computers doesn't translate well to benchmarks of single applications. When multi-tasking, like a normal human, the system is constantly doing reads and writes and switching contexts. This is the biggest benefit of an SSD (random access performance is an order of magnitude better than on a spinning disk), but Intel's Smart Response cache doesn't provide much of these benefits because it is limited by the write performance (and Random Access IOps) of your HDD.
To combat this, you can set up Smart Response in Maximum mode, which allows certain writes to only hit the SSD, and then be later written to the HDD in the background. This seems like it would solve the problem. But, there are some issues... Because it is still a cache, and not really "part of" the "true volume" (as far as Windows is concerned) if something bad happens before that background write to the HDD happens (maybe you lose power), your volume can get in a very bad state. Basically, the HDD is the "truth" and the SSD is considered ephemeral. For this reason, Intel is pretty aggressive even in Max mode about pushing changes out to the HDD. They're not technically "blocking" for the HDD write anymore, but it doesn't delay for long, which means in write-intensive situations, the HDD is still kept pretty busy. This limits the overall performance benefit of the system, when it comes to writes, and in benchmarks, the write-performance benefit for enabling Max mode is often pretty limited. And, in fact, it often
hurts read performance (because you lose some of that magic of freeing up the HDD for simultaneous reads).
Benchmarks show mixed results when using Max mode, and even in best case the differences are relatively small.
Incidentally, this is in contrast to Apple's Fusion Drive system, which actually "unifies" the SSD and the HDD into a single volume. The full SSD and the full HDD are used simultaneously, almost like a RAID-0 (striped) array. So, if you use Fusion Drive on a 1TB HDD and a 128GB SSD, you actually end up with 1.1TB of usable storage (the combination of the two). And, unlike Intel's system, ALL writes to the volume are written to the SSD, and never touch the HDD. The system then later moves infrequently accessed data to the HDD, but since the SSD is part of the "truth" this can be done minutes or hours (or days) later when the system is truly idle.
And, of course, it can and will fill the entire SSD up with data. If you "fuse" a 1TB SSD and a 1TB HDD (giving 2TB of total disk space), and only have 500GB of data written to the volume?
The HDD will be totally BLANK, and the SSD will have all of the data. Since the SSD is part of the truth, you don't need to worry about constantly swapping data between the two disks to keep updating the truth. It isn't a cache, it is an expanded volume, that is smart about performance characteristics. In practice, Apple's Fusion Drive system feels much, much closer in all scenarios to a true SSD.
Of course, the big downside to this is that, like a striped RAID volume, if you lose one or the other, the whole thing is toast. So, this requires a good backup strategy.
None of that matters, but it just drives me nuts how people think Apple's Fusion drive was just Smart Response rebranded. It isn't. The technologies are entirely different.
So...
I just figured I'd mention all of this and explain. If you have the cash, you'd still see a decent performance bump from going to a 250GB SSD and putting your entire OS partition on there. Nowhere near the bump you got from going to the Smart Response cache in the first place, but it is different. It is primarily different when:
1. Multitasking.
2. Using database heavy applications (and none of us ever do this, right).
3. Boot and shutdown times, especially when you include the time it takes to make the system "usable" (including all of the tray icon applications starting up and whatnot).
Author
Topic: AppCrash using MC19.0.118 (No Anti-Virus) (Read 6437 times)