Superfetch and Readyboost in Vista

[sram]

These two are great features implemented in Vista to speed up the system. Now is it worth to say:

1- Use more ram to improve the effectiveness of Superfetch. For example, increase the ram from 4GB to 8GB to make the OS better at learning my habits?

2-Buy a 16 GB ( The maximum capacity one can afford nowadays) flash drive and dedicate it for Readyboost? What I mean : connect it to one free USB port and leave it connected there forever??

[bbz_Ghost]

If you have 4GB of RAM already, you're doing fine. Unless you're actually using it with applications to excess - meaning you regularly use 3GB or more of it just running applications - then SuperPrefetch is using your RAM just fine. Adding more is fine if you want (and with today's prices, why the hell not and be future-proof for RAM for a year or two) but it's not an absolute requirement unless you're using the RAM with the applications. That means some rather power stuff I'd say, like rendering applications for 3D work that suck up RAM like housekeepers with vacuums, or running VMs with several at once. Those situations will definitely benefit from more RAM themselves and not specifically SuperPrefetch.

As for ReadyBoost, it was designed to work at ratios of your physical chip RAM, like 2 to 1, or 1 to 1. So if you've got 4GB of RAM now, a 4GB USB stick would help out but wouldn't make as much difference as it would on a system with 1 or 2GB of RAM. See how that works? There is a point where ReadyBoost simply has no effect anymore on the system's performance because SuperPrefetch will cache stuff in physical RAM which is still far far faster than any USB stick with Flash-RAM will ever be.

ReadyBoost was designed primarily as a benefit to systems without the ability to increase the RAM past a certain point, or mid-range boxes that the owners don't want to spend more money on for RAM, basically. It's a "helping hand," as it were, not a replacement for more RAM which is always the best solution. If you can afford more, especially with the low pricing we have these days, do it and never look back.

Hope this helps...

[finalgt]

I have yet to see any evidence that either of these actually increase a system's responsiveness. In fact, ArsTechnica (or was it Anandtech?) actually did some benchmarks that showed that Readyboost uniformly slowed a system down, instead of speeding it up.

[bbz_Ghost]

Ever since ReadyBoost was announced I've been wondering "how the hell could you actually test that sort of thing, or benchmark the performance differences?" and after nearly two years, I still don't have a clue. I've been building PCs for decades, tweaking them since day 1, and I've used pretty much every testing application, suite, benchmark, or performance monitor that's ever been put into code form, but I simply have no clue as to how to test for things like ReadyBoost.

As for SuperPrefetch, well, I've said this a lot lately and the problem lies in the benchmarking methodology of these sites, including [H]ardOCP if they choose to test Vista against <competition>, mostly XP.

Vista is a self-tuning OS, and it takes roughly 2-3 weeks of daily usage of the machine and the applications installed on it for the real performance spikes to begin to show up. After that, after the tuning period (different for each machine, each platform, each configuration, and each software configuration also) Vista should - and in my experience with Vista IT DOES - outperform XP on the same hardware, both in the 32 bit editions of Vista as well as the 64 bit editions.

The problem with this is that these websites, including [H]ardOCP and all the others, don't seem to fuckin' realize this aspect and they fire up a machine, install Vista on it, benchmark it for a day or two with their already skewed towards XP benchmark suites and applications and then report "Vista underperforms when compared to XP."

Well no shit, Sherlock, of course it does. It's nearly 10x larger, has 4x as many services at the first boot, takes up 10x more space on the hard drive, consumes more RAM and has to fill it with cached data during the actual boot process and beyond, etc.

And they do this testing within minutes of installing Vista? Give me a break, that's so obviously lopsided it's not even funny.

Get a machine, put XP on it, benchmark it from here till Sunday and back, get your baseline for comparison.

Get Vista installed on that same machine, use it for 3 weeks daily, using the applications and everything on it, then benchmark it from there till the following Sunday, get your final results and I promise you Vista will outperform XP on the exact same hardware.

Be nice if some hardware site would figure this shit out before I have to create a site myself that does benchmarks the right way. Everyone automagically spits out Anand, Ars, [H]ardOCP, and other sites like they were "the word of the hardware Gods" and I simply don't go for it. I do my own testing, for my own purposes which are primarily learning and then I tell people what the results are.

It's up to them to keep being sheeple and listen to the "experts" or listen to some of us that actually do this stuff not only for a living but have been doing it for decades.

'Nuff typed.

[Finn]

Readyboost is a marketing gimmick with no real world value. So is superfetch.

[Kowan]

Good point about a fresh install vs. one that has "learned" the user's habits/software useage.

From my limited understanding, ReadyBoost is really only a boost for those with smaller amounts of system memory. Like in the 1 gig range.

[Finn]

Good point about a fresh install vs. one that has "learned" the user's habits/software useage.

From my limited understanding, ReadyBoost is really only a boost for those with smaller amounts of system memory. Like in the 1 gig range.

Readyboost can't boost anyone but people with antiquated pc's who have ram speeds in the range of USB / Flash operation. It could help a user who lacks enough ram to properly even run Vista to use it but boost? It's more like putting an athlete into an iron lung and tell his performance got boosted.

[SuperSubZero]

Vista is a self-tuning OS, and it takes roughly 2-3 weeks of daily usage of the machine and the applications installed on it for the real performance spikes to begin to show up. After that, after the tuning period (different for each machine, each platform, each configuration, and each software configuration also) Vista should - and in my experience with Vista IT DOES - outperform XP on the same hardware, both in the 32 bit editions of Vista as well as the 64 bit editions.
I've been running Vista x64 since Feaburary, and despite a fairly predictable application usage pattern (mostly Outlook, IE, Firefox, lately Virtualbox) I have seen no specific instance where I've been like "wow that loaded fast."

I have Vista x64 on a C2D 2.0ghz 2GB RAM laptop, 7200RPM hard drive. I recently copied a Virtualbox OS image from this to my C2D 2.4ghz 2GB, 7200RPM desktop. Now on Vista x64 I have been pretty much loading this exact same image every single day for the last few weeks, and using it for anywhere from an hour to the entire day. XP64, which has no pre-fetching of any kind, loaded the OS image faster the very first time I used it, than Vista does despite being in near-constant use for weeks.

On a smaller scale, Vista x64 doesn't seem to cache anything else I use with any regularity. Heck, Outlook is a *staple* application and is loaded as soon as I log on to the laptop. It loads no faster now than it did on day 1. Even if it loads .001% faster, it doesn't explain why I need to have several hundred megs "on hold" for this.

Is there any way to see what it's caching? I'm really curious what Vista "thinks" I may plan to use based on weeks of usage.

[DeaconFrost]

I know I'm only going on personal experience on this as well, but with Vista Enterprise x64, I can definitely see a difference in my primary apps loading as compared to the fresh install about 3 weeks ago. Outlook and Trillian load almost instantly for me now, and this is on a Pentium D 915, with 3 GB of DDR2-667 memory.

As for ReadyBoost, the general consensus, from what I've read is, if you have 1 GB of memory or less, it can make a small difference. Above that, and you won't notice anything.

[nobody_here]

my Ultimate 64bit install is much quicker now after two weeks or so of useage than it was the day after install

now...one thing that happened to me yesterday, i went to play with some photos for the first time and i simply right clicked on a jpg and chose edit, and it took it like 20 minutes to pull it up in the viewer.....granted it was a very high resolution photo, but file size was only 6Mb.....it nearly had the machine locked up, couldnt do anything

will see if that goes better next time...

[Finn]

Good point, don't you just hate how Vista takes MINUTES to generate those pesky little preview pictures? With XP you're done browsing all in all before Vista gets around showing the previews.

[ThreeDee]

even if you have 3-4gb of ram .. tho readyboost wont improve performance ..would it at least save on some hard drive thrashing here and there?

http://www.hardfolding.com/ftag1.php/mem/533.png (http://www.hardfolding.com?go=38&id=533&tm=33)
[F]old|[H]ard

[Uberbob102000]

I've been running Vista x64 since Feaburary, and despite a fairly predictable application usage pattern (mostly Outlook, IE, Firefox, lately Virtualbox) I have seen no specific instance where I've been like "wow that loaded fast."

I have Vista x64 on a C2D 2.0ghz 2GB RAM laptop, 7200RPM hard drive. I recently copied a Virtualbox OS image from this to my C2D 2.4ghz 2GB, 7200RPM desktop. Now on Vista x64 I have been pretty much loading this exact same image every single day for the last few weeks, and using it for anywhere from an hour to the entire day. XP64, which has no pre-fetching of any kind, loaded the OS image faster the very first time I used it, than Vista does despite being in near-constant use for weeks.

On a smaller scale, Vista x64 doesn't seem to cache anything else I use with any regularity. Heck, Outlook is a *staple* application and is loaded as soon as I log on to the laptop. It loads no faster now than it did on day 1. Even if it loads .001% faster, it doesn't explain why I need to have several hundred megs "on hold" for this.

Is there any way to see what it's caching? I'm really curious what Vista "thinks" I may plan to use based on weeks of usage.

Strange... It works much faster for me after ~2-3 weeks. On all 4 of my computers with Vista Ultimate 64

[MajorDomo]

I have a 2 gb ready boost on one of my older laptops with 1gb of RAM....it dropped my RAM usage from about 60% to around 35%. I tied it on the other laptop with 2gb and it made little to no diference in performance.