16

Seagate has released a product called the Momentus XT Solid State Hybrid Drive. This looks exactly like what Windows ReadyBoost attempts to do with software at the OS level: Pairing the benefits of a large hard drive together with the performance of solid-state flash memory.

Does the Momentus XT out-perform a similar ad-hoc pairing of a decent hard drive with similar flash memory storage under Windows ReadyBoost?

Other than the obvious "a hardware implementation ought to be faster than a software implementation", why would ReadyBoost not be able to perform as well as such a hybrid device?

13

One major difference is that ReadyBoost is limited to USB 2.0 bandwidth (unless your computer has the ultra-rare and extremely bleeding edge USB 3.0), whereas the hard drive is on the much, much faster SATA interface.

Thus, putting fast flash memory on SATA alone is enough of a win to say definitively that it will be faster.

ReadyBoost is also designed around relatively slow I/O constraints, which limits the scope of what it can do, too.

The one review I found was quite positive. It does seem like, with the right algorithms, you could have the best of both worlds here -- the speed of a SSD (mostly) and the capacity and low price-per megabyte of a traditional HDD.

  • 2
    Can you not have ReadyBoost working from memory card readers on other controllers, eg. PCIe)? (Dunno what real world speeds they might actually get...) – Andy May 31 '10 at 3:13
  • just looked up fastest mem card speeds at toms hardware, and they seem way too slow... – Andy May 31 '10 at 3:23
  • Can you elaborate, or do you have a link on the specifics of the I/O constraints for ReadyBoost? (I wonder if it changed much Vista -> 7) – Andy May 31 '10 at 3:26
  • I was able to put ReadyBoost on my SSD as a test (connected via eSATA). Not sure how I would test any performance boost though. – Sun Sep 16 '14 at 15:53
2

I reckon it will not be the technology that's a winner here, it will be the algorithm used to decide what to store where. Seeing as we don't know the algo for Vista, Win7, or the hybrid, I guess it will take empirical evidence to get a reasonable answer. Having said that, the OS can run more complicated algorithms, look at usage patterns over longer periods, and understand the filesystem itself better, so perhaps there's more potential there. One possible slowdown to ReadyBoost is that it has to encrypt everything because it's assuming removable media, whereas the hybrid solution has no such constraint.

"a hardware implementation ought to be faster than a software implementation"

I'm not sure that has to be true, but you get the advantage of knowing that if your computer's under a heavy workload, the hard drive will still operate at optimal speed. Edit: also it keeps your data buses emptier.

An advantage I see to ReadyBoost is that you've separated the two storage technologies, so you can update them independently as prices decrease or technology improves.

1

One way the hybrid hard drive is superior is that it can speed up bootstrap, while ReadyBoost cannot. The HHD can cache files that are used in the boot, and can thus speed up the boot.

ReadyBoost can never do that because the OS doesn't trust the contents of the ReadyBoost drive across a boot. In fact, since the ReadyBoost session key is thrown away at shutdown and generated anew on every boot, the OS couldn't read the old ReadyBoost cache even if it wanted to.

-2

You are wrong with this assumption; ReadyBoost and hybrid drives are completely different. ReadyBoost was designed to "extend RAM" to help low-level machines, while hybrid drives are designed to improve disk performance for top level machines. The way how ReadyBoost works is: read data from HDD into RAM first, and if you out of RAM push data from RAM to flash drive. ReadyBoost can't improve performance if you have plenty of RAM (actually, it will rather slow your system down).

PS I'm surprised how many people are downvoting the only correct answer in this question. Just check it yourself and you'll see.

  • People are downvoting this because it's wrong. ReadyBoost gives more space for SuperFetch to work in. But it cannot be said to "extend RAM" because the storage in the ReadyBoost drive can't be used as ordinary RAM. – Jamie Hanrahan Mar 31 '17 at 5:29
  • @Jamie Hanrahan, "ReadyBoost gives more space for SuperFetch to work in." - that's completely wrong. – user626528 Mar 31 '17 at 16:54
  • Yeah? Turn off SuperFetch and see how much ReadyBoost does after that. (answer: Nothing.) The ReadyBoost cache is managed by the Store Manager service as an alternative "store" location that SuperFetch can use. If SuperFetch chooses to use ReadyBoost storage for something it wants to cache, then it doesn't have to use RAM for that, so it allows SuperFetch to be more effective even on systems that are short on RAM. But ReadyBoost can't be said to "extend RAM" because the ReadyBoost device cannot substitute for any other use of RAM. – Jamie Hanrahan Mar 31 '17 at 17:40
  • @Jamie Hanrahan, it's not about "more space", it's about more speed. And it doesn't give any benefit to a system that has enough RAM. – user626528 Mar 31 '17 at 21:52
  • "More speed" is the desired result. But the method RB uses to give more speed is by providing more space in which SF can cache things, which increases the chances of later cache hits - while not using RAM for the SF cache. I say "desired" result because if the disk blocks that RB has cached happen not to be used again, you don't get more speed, but more space is still being provided and used. Your second sentence is correct. RB also doesn't help if the system has far too little RAM. IME, for any given workload, there is a fairly narrow range of RAM size where RB will help noticeably. – Jamie Hanrahan Mar 31 '17 at 22:13

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.