Intel's Smart Response; SSD Caching Tested

[Rob Williams]

Interested in faster-than-mechanical performance but don't want to splurge on a big SSD? Intel has a tech for that, and it meets you half-way by utilizing a modestly sized SSD for caching purposes, used in conjunction with any HDD you want for total storage. Read on as we cover both the setup process and our real-world test results.

Read through Ryan's in-depth look at Intel's Smart Response Technology and then discuss it here!

 

[marfig]

Fantastic technology and would benefit me greatly. I'm a software developer and file access is our bread and butter. Understandably, you didn't include the testing of Visual Studio, but I have no doubts whatsoever SSD caching would fit like a glove on these type of applications. And particularly people like me; whose budget is more agreeable with the $110 required here.

A shame though that this calls for a new chipset. At this point it would be unacceptable to upgrade a computer bought just 3 months ago. But will keep an interested watch on the technology and how it develops in the nearby future.

 

[Rob Williams]

Fantastic technology and would benefit me greatly. I'm a software developer and file access is our bread and butter. Understandably, you didn't include the testing of Visual Studio, but I have no doubts whatsoever SSD caching would fit like a glove on these type of applications. And particularly people like me; whose budget is more agreeable with the $110 required here.

Welcome to the forums

To be honest, I am not too sure that having an SSD would speed up a coding environment much at all. There are a lot of little files accessed all the time, but it'd be rare when a bunch of them are accessed at the exact same time. In the event of compiling, speed benefits might be seen, but I think the gains would be negligible. The CPU would be the bottleneck there for obvious reasons, not the storage device.

The only way I could picture a compile process being sped up because of an SSD is if the project tends to be small in file size, but there are thousands of files that need to be processed. A Web browser would be a good example of this.

 

[marfig]

Thanks

I'm thinking more in terms of the process of development as a whole. Not so much, specific tasks directly related to programming. You eventually have some code editor opened, external documentation, auxiliary applications of all sorts, possibly even a local database... essentially we tend to end up with a working set of applications that is ripe with file IO operations going in the background to support them.

But depending on your type of development, I suppose you may also gain directly much from this technology. For instance, there's a lot of IO going on when coding for data access using editors that can connect directly to databases, including remote databases. I necessarily agree though, that for the specific task of coding and building there wouldn't be much of a benefit.

 

[Rob Williams]

I'm thinking more in terms of the process of development as a whole. Not so much, specific tasks directly related to programming. You eventually have some code editor opened, external documentation, auxiliary applications of all sorts, possibly even a local database... essentially we tend to end up with a working set of applications that is ripe with file IO operations going in the background to support them.

Oh, whoops; I misunderstood

I wholeheartedly agree that the general workflow will be much better when working with an SSD. That doesn't only apply for coders, but any developer or heavy multi-tasker. Once you become spoiled to a near-lack of waiting for a file to open or program to load, it's really, really difficult to go back.

I upgraded to an SSD a couple of months ago (I was long overdue, thanks to laziness) and in general still can't get over the super-fast response of the OS as a whole. In the time I used to load an OS and a single application I am already raring to go with five or six applications loaded on an SSD. It's impossible to consider running an OS on an SSD again ;-)

If it fits your budget, I wouldn't hesitate to recommend upgrading to an SSD to improve workflow.

 

[OriginalJoeCool]

I so want an SSD. I'm planning on getting one when I finally build that gaming PC of mine. High-capacity SSD, ATI multi-display (at least 1 x 3 config, maybe something more if I have the money/room for it haha!), infinite speed quantum CPU! Ok, maybe not the last one.

 

[Rob Williams]

I so want an SSD. I'm planning on getting one when I finally build that gaming PC of mine. High-capacity SSD, ATI multi-display (at least 1 x 3 config, maybe something more if I have the money/room for it haha!), infinite speed quantum CPU! Ok, maybe not the last one.

Sounds like a great build to me! It's AMD though, not ATI ;-)

 

[Optix]

Hey Joe, I dare you to call ASUS, GIGABYTE or NVIDIA Asus Gigabyte or Nvidia.

In all honesty, there's no way I could go back to a mechanical drive for my OS. I'd sell an organ before I do so. Even my wife, who is far from a hardware junkie wants one after using my system for the first time.

Getting back to RST though, I'm really impressed how well it worked. I don't really have any general use benchmarks aside from start up and application load times so people will just have to believe me when I say that everything just flows better when accelerated.

 

[marfig]

Yeah. My problem here is the cost associated with SSDs. If you think I'm currently occupying 500Mb of my 1TB mechanical drive, any SSD solution is expensive at this point. Even assuming I could use a smaller SSD to keep up the operating system and some critical applications, while retaining a mechanical drive, I'd still be looking at a system that is needlessly more complicated to manage with two Programs folders and the problems that can bring to legacy applications. But disregarding that argument, I would still have to face the fact that only a portion of my application base would benefit from SSD performance. The rest -- whatever was installed on the mechanical drive -- would remain... slow.

RST however proposes to bring whatever is in a mechanical drive up to the next level. For a more modest price. One that could fit my wallet on a good day. Certainly not as optimal as a full SSD solution, but surely much better of what can be offered by modern fast mechanical drives. For $110, i'd be immediately sold if my system currently supported it.

 

[Rob Williams]

RST however proposes to bring whatever is in a mechanical drive up to the next level. For a more modest price. One that could fit my wallet on a good day. Certainly not as optimal as a full SSD solution, but surely much better of what can be offered by modern fast mechanical drives. For $110, i'd be immediately sold if my system currently supported it.

I couldn't agree more. In the case that someone is Z68-less, there is still the similar solution from Seagate available in the form of the Momentus XT, but unfortunately that only caches up to 4GB worth of data on NAND flash, which is rather paltry in comparison. And still, technically it shouldn't be quite as impressive as Intel's solution here, because Seagate wouldn't be able to offer the same level of NAND performance due to the tight space it'd be working with.

Just to throw it out there, not all SSDs are super-expensive. Take Kingston's 64GB for example:

http://www.newegg.com/Product/Product.aspx?Item=N82E16820139421

It's $125 - not much more expensive than Intel's 20GB SLC - and supports TRIM and fast speeds (250MB/s read, 145Mb/s write). It's only 64GB, but at that price, it seems like a great way to jump in and go. And for what it's worth, if you don't store everything on the SSD, 64GB is totally manageable.

I'm running a 160GB SSD split into two sections, and at any given time I am using about 40% of the space for my Linux OS and about 75% for the Windows OS. I could scale down each even further than this though, so what I'm saying is, where there's a will, there's a way, and anything you have to sacrifice is going to become a non-issue due to the faster speeds you're enjoying.

 

[marfig]

so what I'm saying is, where there's a will, there's a way, and anything you have to sacrifice is going to become a non-issue due to the faster speeds you're enjoying.

That's definitely a good point!

SSDs have always been on the back of my mind; i.e. I think about them for sure, but haven't really made much of an effort into rational thinking, basing most of my decisions on budget considerations without truly evaluating cost-benefit relationships. Or, more importantly perhaps, without doing a sound judgment of my current storage practices and whether it pays to adhere to a new, hmm, paradigm.

hmm... I really have to go that way...

 

[Optix]

To further what Rob said, I'll be doing additional testing using a 64GB MLC Kingston SSD to see just how it stacks up to Intel's 311 Larson Creek offering.

I just read an article from GIGABYTE that says they are due to launch of Z68 boards with onboard mSATA SSDs. I didn't have time to dig in to see if they are specifically going to be used for disk caching but why else would they do it?

My personal setup is in my signature with the 640GB WD drive split up into a 200GB partition closest to the spindle for games and the rest for general storage. Even on such a small SSD I have about 20GB of space left and my install is really quite light compared to most people.

 

[Bravo]

intel SRT - tests vs reality

Hi guys,

I just wanted to say that I really liked your tests and your layout, I feel like I got some solid information out of yoru testing style, and it was one of the main reasons for going to an SSD. However I feel like there is some improvements that you could have to yoru SRT vs SSD test.

I've just installed SRT and found that it extended my boot times by about 7 seconds as it goes through the pre OS loading stage. Your test was quoting speeds between when the windows screen was shown to the time you could use windows. You should have compared from when the computer was turned on to when windows appeared. Then you can compare SSD boot to SRT boot to HDD boot. The extra overhead of the BIOS stage loading is important if you are comparing times to get into windows.

Not saying bad things about the rest of your test in fact it was after reading the actual useful information in here that I got an SSD, just thought that you might want to improve this part of it.

Also I note that you do a Lightroom test in here, if that's cause you actually use it, maybe it's worth posting to this link:

http://www.computer-darkroom.com/blog/will-an-ssd-improve-adobe-lightroom-performance/

I found this to be invlauable in deciding where to spend $$$.