Performance Tweaking & Results

[kjx]

I am an avid Windows performance tweaker and I consider myself to be somewhat an expert at it (>10 years, with a nearly 6 year old application dedicated to performance improvement - DiskMax - it's free btw, and Download.com editors like it).

Ever since Windows 7 released, everyone's been going on about how leaving it alone is the best idea, and that tweaking or 'spring cleaning' would never improve its performance and how indulging in that may well even diminish its performance. While I agree that Windows 7 manages performance far better than any of its predecessors, and that trying to improve performance without knowing what is being done is quite foolhardy, I always maintained that there was a lot of scope for improvement.

But I did have my doubts since I used to always implement tweaks and run DiskMax within a day of installing Windows/apps/updates, and people used to tell me that Windows 7 would take a while to come up to speed. Fair enough, so I thought I'll try it out. I recently upgraded my laptop from 3GB to 6GB of RAM and needed to install the 64-bit version of Windows 7.

It's been about 2 weeks now since the install, and for the whole duration, I resisted the temptation to do anything performance related except removing unnecessary startup items. On fresh installs with my tweaks, I could bring startups to a *fully usable* desktop to sub-30seconds consistently and my personal best was 24 seconds. But even after close to 2 weeks, Windows 7 could come nowhere close to that on its own - with a best case record of 1 min 8 secs, and sometimes as bad as 1 min 45 secs. I gave it all the time in the world to do its thing leaving the laptop on all night several times (defrag ran on schedule - Wed. 1am). Nothing really came out of that.

Today, I decided to perform some tweaks one by one and measure the difference in startup performance and RAM usage.

Tweaking

1) Stock measurements:

min: 1 min 8 sec. max: 1 min 50 secs. ram: 1.2 - 1.5 GB at idle.

2) After a complete scan with Diskmax (my own program; you can find it here: http://download.cnet.com/DiskMax/3000-18512_4-10912587.html or http://www.koshyjohn.com/software/diskmax.html):

min: 52 sec. max: 1 min 24 sec (at step 3). ram: 1.1 GB

diskmax report:

Files deleted: 1150

Space freed: 1024.13 MB

Files queued for deletion: 53

Space to be recovered: 86.78 MB

Total: 1110.91 MB (1203 files)

3) Cleared prefetch (for improvements by DiskMax to show through):

- first boot (for Windows to "learn")

1 min 24 sec. ram: 800 MB

- second boot (to see how it makes use of what its "learnt")

39 sec. ram: 1 GB

4) Services optimization:

39 sec. ram: 1 GB

5) Cleared prefetch (for improvements due to services optimization to show through):

- first boot

34 sec. ram: 1 GB

-second boot

37 sec. ram: 1 G

6) Eusing Free Registry Cleaner (the fact that it has now reached the top of the Google search results for 'registry cleaner' one indication of its quality)

1237 problems found - I verified the entries - ~1400 of them referred to files I had deleted.

startup: 33 sec. ram: 1 GB

There are more things that can be done, including disabling some startup items that I usually remove but I've left enabled (the IDT Audio tray app, Synaptics touch pad app - I use a wireless mouse but I left this for scrolling to work on the touchpad in the rare instances that I actually use it). I usually defrag my registry leading to a another ~5 seconds being shaved off but I'm not certain right now about what application to use with the 64-bit version of Windows 7.

Other things I've not measured but can comment on based on the feel of things: shutdowns are noticably faster, applications launch much faster - particularly browsers (instantaneous - but I imagine Firefox will still take longer (waiting for the final version of Firefox 4); Chrome and IE are top-notch). Explorer windows open instanteously and thumbnails generate *way* faster. Side note: Do NOT set 'Windows Live Photo Gallery' as the default (if you install it), thumbnail generation is faster with the default 'Windows Photo Viewer' (or you can try IrfanView, but the difference is marginal, and the icons look ugly).

Conclusions

1. Windows 7 can do with tweaking. My personal experience this time around clocks in at a 51% decrease in startup time (16 secs from DiskMax, another 13 from *clearing prefetch*, 6 secs from services optimization + registry cleaning).

2. Prefetch does not manage itself well. If you notice that startups are becoming slower, a reset will work wonders. This is something that will have many of you raising your pitchforks, but I'm calling it like I see it. I've always experienced this but this is the first time I've quantified it separately.

3. Services optimization works. One of those online guides will get you started, but once you know what every service does and what its dependencies are, there's a lot of potential to be unleashed.

4. I was expecting to report that registry cleaning results in lower RAM usage. But I guess my registry wasn't messy enough; and I hadn't defragmented it.

Discussion

I know a lot of people will be itching to disagree with me. But constructive replies with reasons for your stance would benefit everyone.

To save everyone time: Simply parroting others without knowing what you are talking about, or taking articles by Microsoft and reputable authors out of context is not helpful. And if you are going to say that you've tried tweaking Windows and things have gotten worse, it's simply because you are not doing it right.

[xdot.tk]

It's all in your head.

[hdood]

Could I ask what you are using to measure boot and application launch times? WPT?

When you get to differences like 33-34-37 seconds, there might also be random variables at play.

[stevember]

Slightly off topic, I have just tried your search program I am absolutely amazed how simple and well it performs.

Thank you, I would have paid for it.

[chrismgtis]

It's all in your head.

:sleep:

He provided sources.

Provide yours.

[+BudMan MVC]

If you looking to track your bootup and what takes what time, etc. you should check out http://www.soluto.com/ Looks like they have disabled download until next release.. but its pretty slick in tracking bootup time, etc.

edit: to be honest I really don't care all that much, since my main desktop is pretty much on all the time - just have it put spin down the disk and standby the monitor - so when I need to do something its like .3 seconds til full access ;) Only time I reboot it is on patches, etc. So does not really matter if on cold reboot if it takes 35 seconds or 59 seconds ;)

[Skulltrail]

I agree with BudMan, I could care less how long it takes to boot as long a it boots properly and with all the necessary services and programs I need to have running when the OS settles. I used to use ASO and TuneUp but I never really saw a notable difference in performance.

I restored my notebook to factory settings a few hours ago due to a problem with my audio. The problem persisted (and seemed to have stopped after flashing/updating the BIOS) but I think it stopped now. My i5 notebook runs well as is, so I won't be using any tweaking solution anytime soon as long as the performance remains this way.

All you have to do is clean the registry every now and then, and defrag the hard drive a few weeks. Thats all. Everything else, keep it the way Microsoft initially made it unless the features are superfluous or indeed causing a notable decrease in performance.

[hdood]

All you have to do is clean the registry every now and then, and defrag the hard drive a few weeks. Thats all.

There is no need to touch the registry. Having keys and values that do nothing doesn't affect performance. Putting invalid data in certain system keys can, but the answer to that is to stop breaking the system. Ironically, registry cleaners and optimizers cause more problems than they solve.

[Phenom II]

I have had Windows installed on here for a couple months now, no tweaks or anything done to improve performance or boot time, no overclock for this test

BEFORE running DiskMAX

AFTER running DiskMAX

Your App actually slowed my machine down :laugh:

But my machine was ready to be used well before this boot timer finished its stuff so take another 10 or so seconds off that - and that is how long it takes for me to have a fully useable desktop

Tweaks are all a myth, except the obvious startup programs of course

Hardware is the key to fast machines

[kjx]

Could I ask what you are using to measure boot and application launch times? WPT?

When you get to differences like 33-34-37 seconds, there might also be random variables at play.

There are four ways of measuring startup time as far as I am concerned:

1. Using a stop watch from HDD boot to taskbar appearance. This is subject to user errors and is not reliable. And it does not actually measure time to a usable desktop.

2. Using a stop watch from HDD boot to a "usable desktop". Not reliable and "usable desktop" is subject to user perception.

3. Measuring the time taken for the system to startup up and for the userinit.exe process to disappear. This measures the whole startup process but does not really account for the fact that an usable desktop is achieved way before the userinit.exe process disappears.

4. Windows actually logs the time since the boot process started, and that is what is used by the uptime counters. I use a light desktop gadget called 'Uptime' that is actually used to measure uptime. But I've found that the time it displays as it first appears is a sufficiently reliable measure of time to a "usable desktop". The time it displays has been independently verified using a stop watch, and it always appears after other startup items have loaded. Looking at hard-disk activity and processor usage at the instant of its appearance, it was just about always at a bridging point from a period of high activity to a "usable desktop". It is not the most scientific but it serves the purpose. The biggest plus point as far as I am concerned is that I can monitor every single startup and immediately take note if there is a significant and sustained decrease in performance.

Slightly off topic, I have just tried your search program I am absolutely amazed how simple and well it performs.

Thank you, I would have paid for it.

Thank you. :) You can always donate whatever you think it is worth to you. It helps cover my hosting bills.

If you looking to track your bootup and what takes what time, etc. you should check out http://www.soluto.com/ Looks like they have disabled download until next release.. but its pretty slick in tracking bootup time, etc.

edit: to be honest I really don't care all that much, since my main desktop is pretty much on all the time - just have it put spin down the disk and standby the monitor - so when I need to do something its like .3 seconds til full access ;) Only time I reboot it is on patches, etc. So does not really matter if on cold reboot if it takes 35 seconds or 59 seconds ;)

I remember taking note of Soluto when it first released. I'll go back and get it. If I see something of real value, you can bet I'll make something better and put it in DiskMax or some other app.

I agree with you about not having to worry about bootup time. My laptop is also just about always on (or on standby), and I have the battery acting as a UPS in the extremely unlikely event of a power failure. Boot up times mean even lesser for those who have SSDs (that'll be my next upgrade when cost/GB falls to something reasonable). But the thing is there are users who constantly suffer through 2 mins+ startup times. And shockingly, they think that it's time to buy a new computer (they attribute the slowdown to the computer's "age").

And while I have only measured startup times and idle RAM usage here, all my apps start faster too. I've already mentioned browsers and explorer windows. Microsoft Word 2010 (applies also to Excel & Powerpoint) starts up in 0.5 secs (hard to measure, used the split timer on my cell phone) - down from about 2 sec. Microsoft Expression Web 4 starts up in 3 to 5 seconds, down from around 8 to 15 seconds. Windows Media Player loads fully in 1.3 seconds (not sure how much it used to be before but only marginally more, no mind blowing improvements). Paint.net loads in 2.21 sec (first start), 1.45 sec (second and subsequent). All other applications mentioned before were measured for their first starts on a cold reboot (to negate any improvements from Windows caching stuff in RAM).

I agree with BudMan, I could care less how long it takes to boot as long a it boots properly and with all the necessary services and programs I need to have running when the OS settles. I used to use ASO and TuneUp but I never really saw a notable difference in performance.

I restored my notebook to factory settings a few hours ago due to a problem with my audio. The problem persisted (and seemed to have stopped after flashing/updating the BIOS) but I think it stopped now. My i5 notebook runs well as is, so I won't be using any tweaking solution anytime soon as long as the performance remains this way.

All you have to do is clean the registry every now and then, and defrag the hard drive a few weeks. Thats all. Everything else, keep it the way Microsoft initially made it unless the features are superfluous or indeed causing a notable decrease in performance.

As much as I want this to sound unbiased, it won't because I make an application that in some ways competes with TuneUp Utilities. But for what it's worth, I've had users telling me that they've seen much better performance improvements from running DiskMax than TuneUp Utilities. I could not publish that without trying it out myself. So some time in the last month, I installed the trial version of TuneUp Utitilies 2010. And I was underwhelmed considering it's a $50 utility. It is certainly good, possibly for people who have never maintained their system. And there is also an element of human psychology at play: on one hand you have an app made by a computer science student costing nothing, and on the other you have a company backed application costing $50. People who can afford the $50 will, more often than not, think, 'if they are charging $50 for it, it must be good'.

I have an obsession with optimising performance, both of computers and of algorithms (you can see it in my software - like neoSearch which in unbelivably light weight and quick for a desktop search application - it's biggest plus point is that it allows you to index networks you are connected to - something that Windows 7 cannot do). Some of the things I've learnt will not be of that much value to those with powerful hardware, but few people have that good fortune.

I have had Windows installed on here for a couple months now, no tweaks or anything done to improve performance or boot time, no overclock for this test.

Tweaks are all a myth, except the obvious startup programs of course

Hardware is the key to fast machines

I'm going to assume you ran the Complete Scan with both the defrag and prefetch optimization process at the end.

Could you delete the contents of your prefetch folder, and then reboot your system? Let it idle for about 5 minutes for ReadyBoot to properly populate. Then, reboot again and measure the difference.

Also could you tell me what BootTimer app you are using? Just so that I can see how it works on my machine..

[Phenom II]

Using this boot timer

http://www.planetsoft.org/utils/BootTimer.exe

and I will try the other suggestion tomorrow and post back (3:10am now)

[kjx]

Thanks for the link. It's quite an old application (2006), and there were discrepancies in its timing from Windows' own measurements (for eg. uptime from Task Manager; if there's an intentional 10 second delay between it's measurement and displaying the result, then it's ok - for eg: it'll say 31 seconds when Windows has already counted to 41 seconds). I'm assuming it's accurate anyway and here are my results after different stages.

1. Simple run of BootTimer after a day of regular usage since last optimization:

http://www.imgftw.net/img/874573302.png BootTimer: 51.932 s

2. After running DiskMax, clearing Prefetch, and rebooting:

http://www.imgftw.net/img/715453000.png BootTimer: 67.314 s

3. Rebooting again to measure the improvements:

http://www.imgftw.net/img/733152223.png <-- large image - full screenshot with Uptime gadget reading (27 s but captured at 31 s, my fault) also

http://www.imgftw.net/img/360570399.png BootTimer final reading: 30.778 s

[hdood]

There are four ways of measuring startup time as far as I am concerned:

It just seems a bit odd that you wouldn't take advantage of the fact that Windows supports event tracing, both for the kernel and applications. It provides valuable data.

Also, you didn't address how you deal with random variations. How many boots/launches do you average over? 10? 20? How many systems (I mean, it's not given that your improvements will apply to any system)? And do you measure the impact of your tweaking on other things than the boot process, particularly when it comes to the logical prefetcher/superfetch?

Don't get me wrong though, I'm not saying that people can't improve their boot time by doing your stuff, just that your testing seems a bit incomplete and simplistic. Your last post also seems to suggest that it's very temporary. By running it again after just a day, you cut the time down by 20 seconds again. A single day?

[Prt Scr]

How many posts are there on the interweb about tweaks in general

Always makes me laugh when people spend countless hours swapping registry settings and such crap. There have been occasions where people continued conversations about certain "tweaks" EVEN years after Microsoft debunked them as having ANY effect.

Theres only 5 ways to tweak things

1) New or faster CPU

2) More RAM

3) Faster HDD

4) Pagefile settings (and even this is marginal really)

5) Turning off unneeded services and GUI elements (just a little less marginal than pagefile tweaking)

Anything else is imagined

You will spend more time trying to detect any perceived improvement than the actual perceived improvement

As for the OP's advice about registry cleaners:

One of the most common things ive seen over a number of years is people who have come to me with non-functioning systems (ranging from fully non-working or hosed file associations where no file is associated with anything, or registry entries relating to critical system files being deleted) after running one of these "registry cleaners".

After reinstalling windows for them i give them all the same advice: If you see software on the internet that claims to speed up your PC, or "fix" your registry or speed up your computer by "fixing" your registry, do not install it. If you do, then please take your PC to someone else to fix next time.

And registry "cleaning" is completely different to defragmenting or compacting the registry. You would have to run your system for many months for defragging or compacting your registry files to notice even a slightest hint of difference.

Messing with prefetch (apart from clearing it once after you have installed all your apps after a fresh windows install) is a bit of a waste of time generally, mostly as it has its own functioning for purging items that arent used frequently

As for the OP's product, it is exactly the same as all the others, makes wild claims:

Quote from download.com (cnet):

Publisher's description

From KoshyJohn:

With DiskMax, your system will be peppier, applications will launch faster, files will open quicker, games will run better, and you will be happier. Running DiskMax from time to time is recommended to keep your system in shape. Even if you are an advanced user, DiskMax will be useful to you since it automates all the tasks required to keep your system clean and fast, leaving you to enjoy the whole new experience.

Will it really make me happier? Maybe v4.39 could solve world hunger as well? All in 864Kb

And exactly how will this make games run better?

[lordvj]

Have you tried disabling unnecessary non plug and play drivers, tasks from task scheduler , and startup event data collectors ?

[Phenom II]

It is nice to have a quick boot time - but to be honest, my PC very rarely gets rebooted, my laptop just gets closed and put to sleep - so its not really an issue

With older OS's yea - because leaving them running for longer and longer caused them to slow down and become a pain, and they regularly needed rebooting

But with windows 7, I am yet to see it slow down, even after being left on for months at a time without a reboot it stll responds just as fast as first boot day

[kjx]

It just seems a bit odd that you wouldn't take advantage of the fact that Windows supports event tracing, both for the kernel and applications. It provides valuable data.

My objective being to provoke discussion, I used methods that average users can somewhat easily replicate. Should I have gone for a fully scientific study? Yes, maybe I should have. But I really could not hold out any longer because a 1 min+ startup is positively sluggish vs. a 27 sec one. Windows tracing is not exempt from random variations, and the last time I used a boot trace was to investigate a serious problem was 3 years back when Windows consistently went for 2.5 min boots on a potential 30 sec system.

Also, you didn't address how you deal with random variations. How many boots/launches do you average over? 10? 20? How many systems (I mean, it's not given that your improvements will apply to any system)? And do you measure the impact of your tweaking on other things than the boot process, particularly when it comes to the logical prefetcher/superfetch?

My measurements are not scientific (actual measurements over single trials; since it validated unrecorded observations from 100s of boots, and given the scope of this thread, I didn't go into greater detail). I can fully appreciate the difference between what I've done and what you expect, but I felt this was sufficient for a discussion like this since it reflects value variations I see from the Uptime gadget in different scenarios for every single boot for 2-3 years now (I had it since Vista). I generally only help out with about 4 laptops per month and I do not measure differences in performance like I've done here. But all the owners have been extremely happy with the very significant improvements in performance, and I've been branded a 'genius' by many but honestly that's only because they couldn't understand what I did.

I did measure differences in application performance in one of my longer replies on the first page. But there can be variations of up to +/- 0.5 seconds since humans can only achieve such accuracy with a stopwatch no matter how alert they are.

There are random variations in my results, but unless there was an update or application install doing something during startup (and there was not), variations will not be in the +/- 15 seconds range. +/- 5 seconds is believable in the somewhat controlled conditions I created.

Don't get me wrong though, I'm not saying that people can't improve their boot time by doing your stuff, just that your testing seems a bit incomplete and simplistic. Your last post also seems to suggest that it's very temporary. By running it again after just a day, you cut the time down by 20 seconds again. A single day?

If an average user uses their computer for 1 hour a day, my system went through close to "2 weeks" of use in that day just by time period estimations alone. I do a lot of software development, web development, design, document management, massive amounts of web browsing (I'm on a 50Mbps+ university connection), and more. And I am an aggressive multitasker with virtual machines up and running at least 25% of the time (making near full use of my 6GB of RAM). All in all, I can effortlessly generate 1GB of 'junk' files every day. So calling it 'just a day' does not mean much really until you look at it in context.

How many posts are there on the interweb about tweaks in general

Always makes me laugh when people spend countless hours swapping registry settings and such crap. There have been occasions where people continued conversations about certain "tweaks" EVEN years after Microsoft debunked them as having ANY effect.

There are people who follow guides blindly to tweak stuff and there are people who know how to tweak properly based on accumulated knowledge. Generalizing the former category to include the latter is a huge disservice to the latter set.

Anything else is imagined

You will spend more time trying to detect any perceived improvement than the actual perceived improvement

I don't know what you are basing these statements off. Until atoms were discovered, scientists who attempted to find the building blocks of matter and failed could only conclude that they failed, and not that atoms (or their equivalent) did not exist. Similarly, being unable to noticable tweak Windows' performance does not mean that it is not possible, it only means that whoever tried and failed, failed.

As for the OP's advice about registry cleaners:

One of the most common things ive seen over a number of years is people who have come to me with non-functioning systems (ranging from fully non-working or hosed file associations where no file is associated with anything, or registry entries relating to critical system files being deleted) after running one of these "registry cleaners".

After reinstalling windows for them i give them all the same advice: If you see software on the internet that claims to speed up your PC, or "fix" your registry or speed up your computer by "fixing" your registry, do not install it. If you do, then please take your PC to someone else to fix next time.

Registry Cleaners can cause problems and often do for people who let them purge whatever they find without even inspecting the list. There are many bogus performance apps out there and I understand your point of view.

And registry "cleaning" is completely different to defragmenting or compacting the registry. You would have to run your system for many months for defragging or compacting your registry files to notice even a slightest hint of difference.

Messing with prefetch (apart from clearing it once after you have installed all your apps after a fresh windows install) is a bit of a waste of time generally, mostly as it has its own functioning for purging items that arent used frequently

The registry can get fragmented in 2 ways:

1. Internal fragmentation - Entries that are deleted are not 'deleted' - the space that used to be occupied by those deleted entries remain in the registry files. Internal fragmentation theoretically causes more RAM to be occupied than required. Entries that serve no purpose can be considered to contribute to internal fragmentation in a sense.

2. External fragmentation - the registry files get physically fragmented on the hard disk because they grew bigger than the largest contiguous free block from the end of the file. External fragmentation causes bootups to become slow.

Then there are wrong entries in the registry - like those that point to a file that no longer exists, or ask an application to try and load something that no longer exists. Registry cleaners also help identify useless entries left behind by less-than-thorough uninstallers. And there are those MRU lists. What many people fail to realize when they look at the split between the 'Cleaner' and 'Registry' areas in CCleaner is that the many of the options in the former category actually clear out stuff from the registry. You probably knew that but most don't.

About prefetch: Did you miss the whole point of the original post? One of my contentions was that prefetch could not manage itself well enough. What you are stating is something I've read lots about but it's something that still does not work like intended in actual practice.

As for the OP's product, it is exactly the same as all the others, makes wild claims:

Quote from download.com (cnet):

Will it really make me happier? Maybe v4.39 could solve world hunger as well? All in 864Kb

And exactly how will this make games run better?

How do you know they are wild claims? Did you try it out?

Will it make you happier? I don't know. But the vast majority of my users love it. And if it wasn't good, there wouldn't be hundreds of sites on the internet reviewing it favourably. Here's a screenshot of a small selection of feedback I've received this month: http://www.imgftw.net/img/122610335.png - if they are happy, that's all that matters. Views such as yours based solely on assumptions are of little value to anyone.

You seem to think being as small as 864Kb is some sort of a handicap. Being that small and efficient is something that takes a lot of work to achieve. Oh but bigger is better? I'll create a 100MB version of DiskMax for you.

Games will load faster and run better on an optimized Windows system. The difference may not be noticable on powerful systems but it'll make a noticable difference on those with less powerful hardware.

Have you tried disabling unnecessary non plug and play drivers, tasks from task scheduler , and startup event data collectors ?

Drivers - no because the only thing I don't need is my modem. Task Scheduler - yes but only those that have triggers set to launch at Windows startup, and those pesky Customer Experience Improvement things. Startup event data collectors - no, but from what I understand those traces can manage 20,000 events a second with less than 5% processor usage (on 2005-era processors). Have you noticed any significant improvements from disabling them?

[kjx]

It is nice to have a quick boot time - but to be honest, my PC very rarely gets rebooted, my laptop just gets closed and put to sleep - so its not really an issue

With older OS's yea - because leaving them running for longer and longer caused them to slow down and become a pain, and they regularly needed rebooting

But with windows 7, I am yet to see it slow down, even after being left on for months at a time without a reboot it stll responds just as fast as first boot day

I don't disagree with what you've stated. From personal experience though, depending on what applications I've run and closed, I often see my idle RAM usage climbing and often doubling from what I'd see from a clean start. I try to trim processes' working set and the system cache but there are instances where the recovery does not yield the expected results. And in those situations, snappy restarts just make me smile.

[soldier1st]

to i agree what budman said about the bootup part, as long as it boots within a reasonable time(not over 3 mins as that would be too long) and cleaning the registry could break things, you can defragment the registry as that will not remove anything but compact it and optimizes it as that can help a bit but as far as the services go, there is no need to change the defaults but you can change the 3rd party services if they are causing a problems sometimes disabling a 3rd party service could yield some improvement but i said could but not definate. now the situation would change if your circumstances changed like if your moving around and going from a public spot like a cafe then you could temporarily change some of the networking services to help keep you more secure but that would be your choice and it would depend on which windows version you have. you don't have to do that with Linux.

[Phenom II]

I don't disagree with what you've stated. From personal experience though, depending on what applications I've run and closed, I often see my idle RAM usage climbing and often doubling from what I'd see from a clean start. I try to trim processes' working set and the system cache but there are instances where the recovery does not yield the expected results. And in those situations, snappy restarts just make me smile.

Windows 7 is designed to use up as much ram as possible though, and when another app requests it, to free it up - so yea, seeing high ram usage in Windows 7 means it is working properly

If you want to see lower ram usage, open something that uses up a shed load of ram, and then close it again - your ram useage will drop massively, and then if its not getting used, the system will start allocating it out to different things until its needed again

[+BeLGaRaTh Subscriber¹]

That Soluto software is amazing Budman, just grabb ed it and tried it. Boot time went down from 3:17 to 2:12 with it. Awesome stuff. Unsure what the Ongoing Frustration Research option is for though?

[+BudMan MVC]

I have just played with the software a bit, having it running cuz it seems to be pretty neat stuff ;) But it seems to me they are trying to put together a database -- I have updated some stuff for it.. When you look at your bootup sequence, and it shows some.exe running without a description you can put details into their wiki for them about what the process is, etc.

So once they have this database off all the things people are running, when they run, etc. And it keeps track of your machine booting and installing stuff -- you can click that your frustrated with your machine, and they look to the database for others having same sort of issue, running the same software, etc. Atleast that is my guess on it - again have not spent a lot of time on it.. Came across it a while back -- and their graphics are pretty slick, and it tracks your boot time and what is needed, what is not, etc.. Seems like a easy way for users to see all the nonsense they end up loading on boot up over a few months of installing this and that, and they all freaking seem to have tray icons, and helper services running, etc.

Like I said I could give a crap if my machine boots in 35 or 59 seconds since its always on anyway.. But I have seen some users boxes take like 5 minutes to load ;) So this is kewl software that you could point them too that even the most simple minded user could understand ;)

BTW -- just another 2 cents on the whole bootup tweaking... Lets say you save your machine 10 seconds in bootup.. How long did you spend on shaving off that 10 seconds -- how many boots is it going to take 10 seconds at a time to get that time back?? ;) Something to think about for all those obsessed trying to shave a second here a second there.. I like the comment about getting faster hardware -- I agree, if your not happy with your 1 min boot time get a FREAKING SSD... or don't shut down your machine so much :rofl:

Edited August 15, 2010 by BudMan

[kjx]

That Soluto software is amazing Budman, just grabb ed it and tried it. Boot time went down from 3:17 to 2:12 with it. Awesome stuff. Unsure what the Ongoing Frustration Research option is for though?

3:17? 3 minutes, 17 seconds?? What hardware are you on? If you are interested, I'll try and help you get it to below a minute.

BTW -- just another 2 cents on the whole bootup tweaking... Lets say you save your machine 10 seconds in bootup.. How long did you spend on shaving off that 10 seconds -- how many boots is it going to take 10 seconds at a time to get that time back?? ;) Something to think about for all those obsessed trying to shave a second here a second there.. I like the comment about getting faster hardware -- I agree, if your not happy with your 1 min boot time get a FREAKING SSD... or don't shut down your machine so much :rofl:

If it includes manually measuring improvements like I did in the first post, I agree there's really no point. But if it's for distilling the knowledge obtained from experimenting into a program like DiskMax, all the time-cost is transferred to the person writing the program. You can run it whenever, even while you continue with whatever else you are doing. And the performance benefits come to you without any effort beyond actually starting the program up.

As for other tweaks like for services, it's very rarely that you need to do it more than just the once per Windows install. That'll take about 3 minutes if you are experienced (15 minutes if you need to refer to a guide, and a whole day if you screw something up).

And not everyone can afford the hardware of their dreams, as simplistic as that solution is.

[lordvj]

Soluto modifies stuff at the kernel level , is what i've heard. Never gonna use it.

And yes disabling drivers did speed up boot , amdxata.sys was starting up at boot even on my intel based system , some hyper-v drivers [Disk Virtual Machine Bus Acceleration Filter Driver] , 2 drivers for NetBIOS - who uses that anyways , set the sandboxie driver to demand start , msahci - i use iastor.sys , set TCP/IP Registry Compatibility driver demand start - its for legacy applications anyways and disabled the offline files driver.

you could check this , its a little extreme , but meh.

http://www.graphixanstuff.com/Forum/index.php?showtopic=15194

[+BudMan MVC]

"Soluto modifies stuff at the kernel level"

Where did you get that it modifies anything?? From this statement?

"Soluto employs innovative low-level Windows kernel technologies to identify what users are asking their PC to do, and what their PC does in return"

So what antivirus do you run? What software firewall? Just the stuff from MS? If not -- guess what most likely they are also using "low-level Windows kernel technologies" ;)