In this week’s blog post, I’m going to share how I came up with this new feature and how the original idea was transformed into reality. You are going to have access to never-before-seen prototypes, initial screenshots, and our development cycle. Enjoy!

What is TuneUp Program Deactivator?

Windows users know this age-old theory; the more software you install on your PC, the slower it gets. Now, instead of uninstalling programs (you might need them later, right?), we figured out a way to disable programs rather than get rid of them entirely.

So what does that mean? TuneUp Program Deactivator turns off all of the parts of a program that slow down your machine the most: start-up processes, services, and automatic background tasks. The program is still there, but it’s just “sleeping”. There is no memory consumption, processor usage, annoying balloon tips, splash screens during startup, or hard disk activity involved. The cool thing is that the moment you start the program (just double-click on its launch shortcut), TuneUp Program Deactivator enables it again on the fly!

From idea to reality

Do you remember our 200 programs experiment from earlier this year? (Read more in Part 1, Part 2 and Part 3) This marked the birth of TuneUp Program Deactivator.

Back in January, we tested how much a typical PC suffers under the load of having 200 programs installed on it. The result: IT horror! Boot time took more than seven minutes, and the performance of the 2009 notebook was comparable to a vintage machine from 1999—it produced error messages and froze on a regular basis. The test bed was a prime example of a PC that a user never cared to maintain or didn’t know how to optimize.

This sparked an idea! What I wanted to know was how much of an effect TuneUp Utilities 2010 would have on this horribly bogged down machine. I used our suite to optimize the PC as thoroughly as I could; I disabled unnecessary services, turned off start-up entries, and defragmented the disk. I also tried optimizing the system with the current versions of all major competitive products. And while performance improved a bit after I used our own as well as other tuning suites, it did not reach the level of a clean and fresh Windows PC.

For me, this situation was just not acceptable— all of the tuning suites on the market (including our own) couldn’t really solve PC performance problems caused by third-party programs. This haunted me for weeks—I simply had to find a solution to this problem.

And no, simply uninstalling these programs wasn’t a solution—at least if users wanted to use these programs again in the future. Many users don’t like to uninstall software because they don’t know how to reinstall it should they need it again. In some cases, they don’t know where they downloaded the software from, they don’t have the CD anymore, or they don’t know where the product key is, among many other reasons.

The incredible loss in performance is because many programs claw or “hook” themselves in way too many areas of Windows to keep track of. They don’t just run at startup (which is rather easy to disable); they also install services or create so-called “scheduled tasks” that get started regularly while the PC is running, and in many cases they even use several of these methods at the same time. Combine all this and multiply it by the amount of installed programs, and you’re pretty much facing a ruined PC.

I needed a way to get rid of all those “hooks” of a program (at once) and restore the performance of even the most bogged down Windows PC—all the way back to its original state. My first step was to determine if this huge undertaking was even possible—could performance get back to normal, even with hundreds of programs installed? I proved this theory in February by manually disabling each and every hook entry using Sysinternals Autorun and shutting down all third-party processes.

It was mind-bogglingly complex because even I as an expert sometimes couldn’t figure out which programs these entries belonged to, but it worked. Without all of these entries, all of the test PCs eventually performed as well as they did on the very first day they were used.

Obviously, 99% of all Windows users out there don’t a) have the time or b) the knowledge to find every single entry of a program and safely turn them off. We needed a solution—a tool to automatically do this. This is how the idea of TuneUp Program Deactivator was born!

One problem, two solutions

To create a solution to this performance problem, I locked myself in a room with two of my most skilled developers to theorize about possible solutions and test them. Our goal was to figure out a way to deal with all of the hooks programs could use and thus speed up Windows again. We eventually ended up with two solutions:

• The creation of a virtual environment, or a “sandbox”, that’s separate from Windows. We needed to figure out a way to put all of these programs inside that environment. While a good idea in theory, we realized that this is too complicated and that there are many issues users would face, one of them being bad performance through the virtualization overhead.

• The disabling of all of the hooks. We needed to figure out a way to detect all relevant resource-draining hooks a program had created and turn them off safely.

Despite the fact that finding all of these hooks is a complex task, we went with the second route. We wanted to give users the least complicated way to see which programs slow down their PCs, and give them full control by allowing them to turn programs on or off with just one click.

This is TuneUp Program Deactivator in its earliest form. We used this command line prototype to experiment with the deactivation of programs

During the initial research phase, we faced many problems. Disabling some hooks within the operating system caused error messages and led to stability problems. There was a need to care about dependencies between services, scheduled tasks, and start-up entries; disabling program A should not lead to problems with an enabled and active program B! In March 2010, we created early prototypes, which were basic command line tools that disabled all of the hooks of a program and re-enabled them.

Using these prototypes, we experimented with hundreds of programs and developed several methods to determine how programs integrate into Windows systems, such as reading out Windows Installer tables, uninstall keys, version information blocks of all executable files, and other relevant entry points in the Windows registry.

This is a German mockup of one of our first user interface concepts for TuneUp Program Deactivator, which was named “Sandman”

This was finally implemented into an early German beta version a couple of weeks later

This screenshot shows how TuneUp Program Deactivator finally ended up in our optimization software suite

Main goal achieved

May 1 marked the day of our proof-of-concept. We had successfully proven that it was possible to disable essentially all of the hooks of a program with our own software, and we were in the stages of defining programs, especially security software and drivers, that we “whitelisted” and thus protected from accidental deactivation. We then moved the early prototypes into the next stage of development, such as tweaking the graphical user interface and fixing bugs. Next on the agenda: a name! During the entire development phase, we had several potential ones for this feature; these included: “Sandman”, “Program Reliever”, “Program Standby Switch”, “Programs-on-Demand Switch”, “Performance Drain Switch”, “Slowdown Eliminator”, “Program Unhooker”, and “Program Silencer”.

This is an early concept design of the performance graph in German. We thought about different ways of showing how TuneUp Program Deactivator affects users' PCs

But in the end, the name TuneUp Program Deactivator was chosen. Then we set two new goals to make the feature even more unique. First of all, we wanted a performance meter that shows the load of each installed program both before and after users disable it. We came up with several ideas to show our users how each disabled program affected boot-up performance, day-to-day performance and shutdown performance.

Our second goal: We didn’t want to force users to constantly go in and out of Program Deactivator to turn on programs they needed. We knew that would be just too much of a hassle. We needed a way to automatically re-enable programs “on the fly”. That’s why we designed TuneUp Program Deactivator to detect when the user tries to launch a disabled program again, so it can automatically reactivate all of the necessary components (the processes, background tasks and services). Typically when a feature of a program is missing, the program will try to launch its self-repair mechanism or produce an error message. To protect our users from these types of problems without slowing down the PC through constant polling or monitoring, we used a little-known redirection mechanism provided by Windows itself. This means when users double-click on a disabled program (say “itunes.exe”), Windows launches a small part of TuneUp Program Deactivator instead, which first re-enables all of the program’s components automatically and then starts the original program itself. The program will never know it was disabled in the first place!

The finish line

Once we achieved all of our goals, we went into the beta and release candidate phases in late summer and early autumn of 2010. Our thousands of internal and external testers helped us fine-tune TuneUp Program Deactivator and fix remaining bugs. In total, we had eight beta versions and one final release candidate before we launched TuneUp Utilities 2011, with our new key feature TuneUp Program Deactivator.

All in all, we’ve made speeding up and stabilizing much easier for all of our users. Rather than asking users to uninstall (and then probably reinstall) their programs regularly to keep their PCs at peak performance, they can now just turn them on and off with a single click. I’m proud that, over the past couple of months, we’ve developed a hassle-free solution that achieves all the positive effects of many startup, service, and scheduled task managers combined—while still being much easier to use. Actually it’s a lot of fun to just flip those little on/off switches and see your PC’s performance go up instantly, especially on PCs with a large number of installed programs!

Readers, now it’s your turn: Please leave a comment, and tell us what you think about TuneUp Program Deactivator. Do you have any more wishes for the feature or overall software suite in the future?

The truth: Can an optimization suite solve all Windows issues?

The TuneUp Blog team—with our flagship product, TuneUp Utilities, in mind—made a bold choice and put our very own tool to the test. We wanted to see if and how much our tool could restore the PC’s performance and get rid of the problems crippling this computer.

Readers, please note that while the notebook’s performance improved after many of the benchmark tests, some of the problems could not be resolved by TuneUp Utilities. So, let’s talk about how we tried to improve the computer.

Getting the tuning process started

We used the “Restored PC” (also known as the system after the 200 programs were uninstalled) for more than two weeks. It ran nine hours a day and we used it like we normally would—to edit photos, surf the Web, check our e-mail via Outlook, play games, watch movies, and listen to music. Over the course of these couple of weeks, we repeatedly used TuneUp Utilities and performed the following actions:

• We defragmented our hard disk a couple of times. The “Thorough Defragmentation” option was used in each and every instance.
• We removed 131 invalid short cuts like Start menu and desktop shortcuts, and invalid “Last used” entries that are often found in various programs.
• We corrected about 2200 registry entries. The first time we ran TuneUp Registry Cleaner, it came up with 917 errors. We then rescanned the registry after a couple of restarts and corrected hundreds of additional entries.
• We defragged the registry and were able to reduce it by 12 percent.
• We disabled 14 startup programs.
• We removed almost 12GBytes of unnecessary files.
• We disabled the hibernation feature and gained 2 GBytes of additional hard disk space.
• We disabled the Windows Search feature.
• We enabled Turbo Mode permanently to turn off a couple of unnecessary services, automatic tasks, and background processes.

Below is a screenshot gallery, so that you can see how we cleaned up all of the junk that came with the 200 programs.

Finding out the results

After the two weeks, we figured it was time to go through our series of benchmark tests again; these new results are listed under the “Optimized PC” column. As a quick reminder, “Clean PC” is the fresh install, “Junk PC” is the notebook with the 200 programs installed, and “Restored PC” is the notebook after the 200 programs were removed.

As usual, we made sure that SuperFetch adapted to the programs we used on a regular basis; we also performed each of the tests three times.

Boot-up performance

Due to all of the disabled startup entries and services, Windows 7 launched faster than ever before. It took little more than a minute for Windows to finish booting. While 15 seconds might not seem like much, it is a noticeable jump in performance.

Shutdown performance

Neither the removal of the 200 programs nor the use of TuneUp Utilities made a difference—the notebook still needed 11 seconds to shut down. This came as a bit of a surprise, with Windows 7 shutting down less services and processes after the tuning process.

Virus scan performance

This hard-disk-intensive task of scanning 1.5 GBytes of compressed RAR files took 2 seconds less on the “Optimized PC”. This probably has something to do with the hard disk being defragged and the CPU being not as busy as before.

Application performance

TuneUp Utilities was able to shave two seconds off of the original time it took to start Outlook 2010 Beta—the program never launched faster on this machine! Start-up times for Google Chrome finally went back to normal, however, the slow start-up issue associated with Windows Media Player was not resolved. Like we explained in our last article, this is probably due to a codec that was damaged or altered by the 200 programs we installed and then uninstalled.

Processor, graphics, and memory performance

We were pleasantly surprised with this benchmark. Instead of taking nearly 3 minutes like the “Clean PC”, the “Optimized PC” needed only 2 minutes and 10 seconds to compress the three files. In addition, the frequently occurring hiccups while playing an HD file were gone.

Photo editing performance

Thanks to an optimized Windows, the notebook was able to apply the “Enhance” effect in 21 seconds. Again, due to a combination of all of the tuning techniques, the processor and graphics card were able to process the image just as quickly as before.

Cinebench R10 performance

We were really intrigued by the results of Cinebench, which is designed to test how well the processor and the graphics chip perform. TuneUp Utilities helped increase the performance of the graphics chip and especially the single CPU render test.

S.T.A.L.K.E.R. Call of Pripyat performance

A bit of a mixed bag here—after TuneUp Utilities finished its job, gaming performance increased immensely compared to the “Junk PC” and “Restored PC” benchmarks. But, it did not quite get back to the original levels. (Note: We were able to increase the performance to exactly the original levels when we followed the steps detailed in our gaming blog series.)

PCMark Vantage performance

After conducting the hour-and-a-half-long PCMark benchmark, we noticed that performance was almost back to normal. A level of 2544 points (in comparison with the original 2624 points) is absolutely fantastic, especially considering that we destroyed Windows with the 200 programs.

All in all, the optimization software suite restored the notebook’s performance back to the original levels—and, in some cases, it was even better. In a few cases, however, TuneUp Utilities missed this mark and was not able to speed up the system.

Resolving all of the problems

After we uninstalled the 200 programs, the TuneUp Blog team noticed some major issues with the machine. We were really anxious to see not only how performance improved but also what TuneUp Utilities could do about these issues.

• Strange pauses during file copy – This completely disappeared. We copied a lot of videos and photos from a DSLR camera, camcorder, USB thumb drive, and external hard disk. All of the pauses were gone, as if they had never existed!
• Garbage in “Documents” and “Program Files” – TuneUp Utilities does not offer a feature to get rid of orphaned or empty program and document folders. However, we should also note that cleaning up this mess was done manually in about 5 minutes.

• Remaining Start Menu folders – TuneUp Utilities removed all of the invalid shortcuts within the Start menu folder, but it did not delete the empty folders itself. We needed to do that by hand as well.
• Empty Control Panel – Perfect score! TuneUp Utilities fixed this error: The Windows 7 Control Panel worked as expected, no matter how often we launched it.
• Weird noises during audio playback – Although we are still not entirely sure what caused these weird audio hiccups, we noticed that (thanks to the Windows tuning) these issues were a thing of the past.
• Orphaned Startup Programs – Interestingly, after the optimization process with TuneUp Utilities some of the invalid start-up entries were gone (for example, the InstallShield entry), while others, like the ZoneAlarm client or updater, remained. These invalid entries do not in any shape or form impact the boot-up performance; they are merely part of the junk that was left over by the 200 programs.
  

• Huge Temporary Files – Gone! We were not only able to delete the 3.5 GBytes of temporary data but also a couple of GBytes worth of unnecessary files and folders. As our test bed had only a relatively small, 120-GByte hard disk, we found ourselves very happy with the disk space we were able to recover.
• Start page changed to Yahoo – The start page was still set to Yahoo. But TuneUp Utilities has no way of knowing if we changed this deliberately or if a freeware tool modified the start page. So we just made the change manually.
• A new “Temp” Folder on the hard disk – The folder still exists, despite optimization with TuneUp Utilities However, it is actually good that the optimization suite does not delete folders containing installation files that you might need later. We ended up manually deleting it.
• Spamihilator still active – Not anymore! Spamihilator was disabled by TuneUp Utilities and we couldn’t be happier.

Although some small, residual folders and entries remained, TuneUp Utilities effectively solved the biggest problems and helped clean up the computer. It got the machine’s performance pretty much back to normal.

We’ve been testing this theory and working with this PC since December 2009. During the entire testing period, the TuneUp Blog team could not wait for the day when we finally would be able to just wipe the hard disk and reinstall Windows 7. What we did not expect—the test notebook was finally fun to work with (yet again) after we removed all of the programs and optimized the system. Its performance has been snappy, and we have not encountered a single problem since—and are now using the machine on a daily basis.

Written by Christoph Laumann and Sandro Villinger

Over the last couple of weeks, we have discussed the fact that installing too many programs will destroy the performance of your computer. Check out our two original blog posts and see for yourself:

• In Part 1, we documented and benchmarked what happened to a clean, Windows 7 machine when we loaded it with 200 programs. The performance had decreased significantly in nearly every test we conducted.
• In Part 2, we uninstalled the 200 programs and benchmarked the PC again. We found out that, in some cases, the original performance had been restored, yet noticed a dramatic decrease in speed in other tests.

For the past week, we have used the same PC on a daily basis and encountered a lot of problems that we had not noticed before. Note: We did not change the system configuration or install any additional programs or drivers that might have caused some of these issues to happen.

Issue 1 – Weird pauses during file copy

When copying large video files from an HD camcorder and pictures from a DSLR camera, Windows 7 became basically unusable. The system froze for at least one minute during the copy operation; we could only move around the windows and click on the Start orb. Nothing else would work, such as opening a new tab in our Google Chrome browser.

This did not happen in the original configuration. We even performed a full rollback to the original clean image from December 2009 to confirm that this was the case.

Issue 2 – “Documents” and “Program Files” garbage

Even though we uninstalled all of the 200 programs properly, we found ten folders in our “Documents library” that did not belong there. The situation in the “Program Files” folder was even worse, with 51 folders of previously uninstalled programs still there. While there may be some settings left when you uninstall a program, in most cases, you don’t need these settings and residuals left on your hard disk. (A plea to all developers—give us the choice to fully uninstall programs, including all settings and folders.) Some programs do this, but as you can see, a quarter of all of the software we installed left unnecessary garbage on the hard disk.

Issue 3 – Remaining Start Menu folders

We were shocked when we opened “All Programs” in the Start Menu and found more than 20 leftover folders from the 200 programs; these were either empty or still filled up with orphaned shortcuts.

Issue 4 – Empty Control Panel

What we found was just plain weird: When we opened the Control Panel, the window stayed empty, something that had never happened before, with it only occurring in about one out of every five times we used the Control Panel. This was definitely a result of one of the 200 programs messing things up in the Windows registry.

Issue 5 – Weird noises during audio playback

Every now and then, while listening to music, we noticed that the audio would skip for two seconds. When this happened, we could hear a distorted, very loud, and unpleasant noise. Since we listen to a lot of podcasts and audiobooks on this specific machine, we can confirm with 100% certainty that this issue was not there before. Even installing the latest Realtek HD Audio driver did not make things better.

Issue 6 – Orphaned startup programs

When we opened “msconfig,” a built-in Windows tool to enable and disable Startup programs, we found several active entries of previously uninstalled programs such as ZoneAlarm, Explorer View, ICQ, and a mysterious “InstallShield” application that tries to start a “Setup.msi,” which didn’t even exist anymore.

Issue 7 – Huge temporary files

By accident, we noticed several very large temporary files (all ranging from a couple of kilobytes to several hundred megabytes) located in the “Windows\Temp” folder. We could not figure out where they came from but suspect a couple of program installers put them there. These files amounted to about 3.5 GByte worth of data.

Issue 8 – Start page changed to Yahoo

Some of the program installers changed Internet Explorer’s start page to Yahoo.com, which was not our original homepage.

Issue 9 – A new “Temp” folder on the hard disk

The program “ABBY FineReader 9.0″ put a folder called “temp” onto our C drive. This contains all of the files necessary to install FineReader. However, even though we removed FineReader 9 completely from our system, this folder still remained on the system drive.

Issue 10 – Spamihiliator still active

It was only after seven days that we discovered we had missed uninstalling one program called “Spamihiliator”. The simple reason is that this program did not put an entry in the “Programs” section of the Control Panel. The only entry with an “S” we found was the screenshot program Snagit.

In only seven days, we noticed all of these quirks and leftovers on the machine. What will we find next week? Next month? Next year? One thing has become clear—not only did the machine’s performance suffer quite a bit, but both the stability and disk space were ruined by the 200 programs. Also, thanks to all of the leftover folders, the machine felt “cluttered.” It was like living in an apartment that no one had cleaned for several weeks! Don’t forget to check back and read our evaluation of how well optimization suites can salvage the performance situation!

Written by Christoph Laumann and Sandro Villinger

Read the next part of our 200 programs experiment – The truth: Can an optimization suite solve all Windows issues?

The experiment also left the TuneUp Blog team and our readers with more questions. Is it possible to get this machine running again by removing all of these unnecessary applications? Can the original performance be restored? Or, will this machine be plagued forever with error messages, crashes, and an underwhelming performance?

See all of these programs and pop-ups? We got rid of the 200 programs we installed in a previous blog post!

Coming to the machine’s rescue

If your machine is in this kind of state, the obvious recommendation would be to wipe the hard disk and reinstall Windows! But, this is not an option for folks who have set up their operating system and programs just the way they like them. Backing up and restoring settings, e-mail accounts, contacts, program configurations, etc. is a tough chore and, often takes more effort than trying to clean up the original Windows installation. So, what happens if we uninstall all of these programs again? Do you think this will make the system usable again?

The uninstallation process

This is basically one of the dullest jobs anyone could possibly think of—we opened up the Control Panel and uninstalled the 200 programs one by one. Since we didn’t want to skew the results, we avoided simultaneous uninstallations, which also could cause conflicts with the system files and the Windows registry.

The entire process took us more than a day to complete. Since the computer was extremely sluggish, a simple program often took up to 10 minutes to be completely removed. In the cases of huge software products, such as the Nero 9 Suite, we waited about 20 minutes.

Benchmarks, benchmarks, benchmarks

Just as we were on the verge of losing all feeling in our fingers from clicking “Yes” on every “Do you really want to remove the selected application?” dialog box, the job was finally finished. Then, we started on the tests. We performed all of the benchmarks conducted during our last run. As with the previous performance check, we used the computer on a regular basis for a couple of days and carried out each test three times to get rid of all of the discrepancies and variations. To give you an exact comparison, we included the benchmark results of our last run in the following tables below.

As a reference guide, “Clean PC” represents the machine’s original state. Windows 7 had been freshly installed, and only a handful of programs were added. The machine was at its best possible performance. “Junk PC” represents the performance of the computer after the 200 programs had been installed, while “Restored PC” represents the performance of the computer after the 200 programs had been uninstalled.

Boot-up performance

This benchmark measured the amount of time it took the machine to load the last automatically running program on the startup list.

It was almost back to normal. Windows 7 took 1 minute and 36 seconds to boot up.

This came to us as a bit of a surprise. After we removed the 200 programs, the boot-up time was almost completely restored to its original state. It took only 14 seconds longer than usual to fully load Windows. We figured it was because of the high degree of disk fragmentation. But, let’s be clear about something—what difference does 14 seconds really make, especially when the 200 programs made the PC boot up in more than seven minutes, right? Big thumbs up for Windows 7 here, as it did not let the programs ruin the boot-up process too much.

Shutdown performance

We then measured the time it took from clicking the “Shut Down” button to actually powering down.

The shutdown process returned to its original state.

Windows 7 took exactly 11 seconds to power down the machine, and that is exactly the way it was before we installed all of those programs. Fantastic!

Virus scan performance

We used Microsoft Security Essentials (MSE) to test how long it took to scan 1.5 GByte worth of RAR archives.

The results went back to a normal level yet again. We were happy to see that the 200 programs did not have an impact on scanning performance. This was expected, as MSE’s performance entirely depended on the hard disk speed and processor power. And, since all of the unnecessary background processes from those 200 programs had been removed, there was no interference.

Application performance

The TuneUp Blog team also measured the performance of selected applications. We started with the time it took for Outlook 2010 Beta to load and display the first e-mail. We then started Google Chrome with eight tabs. Like the last run, we also tested how long it took our computer to fire up Windows Media Player and start playing a 1080p full HD video file.

The results were fascinating. No matter how often we tried, Google Chrome never returned to its original performance. Again, we think this is due to hard disk fragmentation. (Or, did any of these programs actually interfere with Google’s browser, so that it now takes 50% longer to start up?)

Windows Media Player needed much more time to load our test video file.

Windows Media Player produced another weird result. After we removed the 200 programs one by one, it needed nine seconds to load up the HD video file—six seconds longer than under a clean install. This is inexcusable and probably due to the many codecs that were added, replaced, and removed with some of the 200 programs. Don’t worry; we made sure that the Windows Prefetch feature optimized Windows Media Player, to make it launch faster.

Processor, graphics and memory performance

We watched WinRAR compress our three test files and frankly could not believe our eyes. Instead of taking two or three minutes (like we were used to), WinRAR needed more than four minutes to compress the files, which took even longer than when we loaded the PC with the 200 programs. There is simply no other explanation than that fragmentation really slowed down things. Even when we compressed the three files, deleted the ZIP file, restarted the PC, and repeated these steps five times, the results stayed the same.

During the compression, the CPU was never utilized more than 20%. That means hard disk performance really suffered from the 200 programs we installed.

Also, during the 1080p full HD playback benchmark, we noticed that every 10 to 15 minutes, the movie started to stutter for at least 10 seconds. This did not happen under the original and freshly installed Windows configuration.

Photo editing performance

Using Corel PhotoImpact X3, we applied the “Enhance” filter to a 30 MB image file. This test should have proved how fast both the processor and the graphics chip are able to render images.

Again, this task was slowed down by the 200 programs that created so much background activity. After removing all of the applications, everything went back to normal. Just like before the installation, PhotoImpact was able to enhance the picture in exactly 21 seconds.

Cinebench R10 performance

Cinebench is still one of our favorite benchmarks, as it clearly shows your PC’s processor and graphics chip’s capabilities. The more points you get, the faster your PC is.

Cinebench performance dropped as well, but the difference was quite minimal. It took just a bit longer to render the sample image on the screen, which resulted in a slightly lower score for our test laptop.

S.T.A.L.K.E.R. Call of Pripyat performance

This benchmark is based on the game S.T.A.L.K.E.R. Call of Pripyat, which was released in Q4 2009.

What a disaster! Gaming performance did not return to the original levels; it basically stayed the same (expect for the SunShafts test from S.T.A.L.K.E.R.). Again, we blame the high level of fragmentation. The game was not able to load all of the textures fast enough from the hard disk to ensure a continuous frame rate.

PCMark Vantage performance

PCMark Vantage was the last benchmark on our list. Again, this is an all around program that tests the PC in different areas, such as music editing, and gaming. The more points PCMark spits out in the end, the faster your PC is.

It was interesting to note that the machine was not able to achieve its original performance this time around. The total scores for our cleaned up system were better than the scores obtained with the 200 programs installed, which was expected, but we were unable to reach 2624 points, even though we performed the tests several times.

Overall performance and stability

In this last section of our test, we wanted to express our personal feelings—how responsive did we find the system and how we thought Windows 7 performed. We used the system for a full week for traveling, business, and a bit of gaming, as well. We edited photos and even hooked up an external Blu-ray drive to watch a couple of HD films. Overall, Windows 7 performed faster without the 200 programs installed, but we felt that the performance suffered quite a bit compared to the original state. Some of our programs did not start up as fast as they used to, menus took a bit longer to show up, and movies stuttered here and there. It was as if we were using a 1.4 GHz (rather than our 1.86 GHz) machine. The TuneUp Blog team believes the system’s performance was “good” when we first installed Windows 7, whereas now it’s only “okay”.

Leftovers and junk

Raw performance is one thing we wanted to take a look at; we were also very keen to see how many “leftovers” we had from the 200 programs, such as temporary files, unnecessary shortcuts, and empty and entire folders.

Let’s start with the desktop. Even though we got rid of all 200 programs, the desktop, especially the Start Menu, were still cluttered with dozens of invalid shortcuts and folders. The next thing we noticed was the number of files on the hard disk. On our clean PC, we had about 101,000 files (44.9 GByte in total). This number moved up to 190,000 files, which amounted to a total of 70.2 GByte, as we installed the 200 programs; this was a massive jump! The sad part was that, after we removed all of the programs, there were still 114.000 files left—and 53 GByte worth of data. There were two reasons for that. Firstly, uninstallers are not that effective; they leave traces behind! Secondly, the 200 programs created dozens of system restore points that amounted to a lot of unnecessary data.

Fragmentation had a big impact on performance. Before we started our rampage and installed the 200 programs, we regularly used TuneUp Utilities defragmentation feature to keep hard disk access times as fast as possible. So, the fragmentation level was usually 0-5% at most. After we installed and removed the 200 programs, TuneUp Drive Defrag told us that more than 54% of the hard disk was fragmented. This is probably the reason why the performance did not go back to normal levels.

Too much junk left

In some cases, performance could be restored by simply uninstalling the 200 programs. In other cases, and probably due to the mess we created, overall responsiveness and gaming tests in particular suffered immensely. The bottom line is that our test laptop wasn’t the same anymore. It was sluggish, and there was so much junk left on it, that the system was no fun to work with.

We were almost ready to back up the entire system and start from scratch. But, wait! That’s not fun either. How about we spice things up a bit? Curious? Just stay tuned for our next blog posts of this series.

Written by Christoph Laumann and Sandro Villinger

Read the next part of our 200 programs experiment – Performance and Stability Update.

This podcast is also available on TuneUp’s Mevio.com podcast channel here.

This podcast is also available on TuneUp’s Mevio.com podcast channel here.

More programs equals less performance?

What most users are unaware of: Installing too many programs can result in problems and a massive decline in performance and stability. And it is not their fault—how should they know? Windows by itself does not do a really good job of telling users to get rid of programs they rarely use and that may cause issues. It is mostly long after the fact that people start asking themselves, “Why is my PC so slow?” or yell out, “My Windows crashes all the time!”

So what’s the cause behind all this? The more programs you download from the Web or install from a DVD, the more services, background processes, files, startup items, and registry entries will be created. Some software products even install new drivers, for example virtual DVD drives. This obviously puts more and more of a strain on your machine, because it needs to assign a certain amount of resources to the newly installed programs, and this, in return, reduces performance.

The question is: Is this drop in performance noticeable? We’ll find out in this exclusive real-life test run!

Our testing method

The “Clean PC”: To test the theory, we measured the performance of a computer with only a minimum set of programs installed. We call it the “Clean PC”.

As you can see, we have Office 2010 Beta, Google Earth, TuneUp Utilities, Windows Live Essentials, Media Player Classic, Microsoft Security Essentials, Age of Empires III: Asian Dynasties, Sony Media Go and PhotoImpact X3 installed. All the drivers are up to date and all Windows 7 updates have been installed.

The “Junk PC”: Then we installed exactly 200 software products that are widely used. We selected these 200 based on the most used programs on Wakoopa (a software usage tracker), the top downloads at Downloads.com, from various “Top 10 Programs you need” lists and from our personal experience with what our readers use most often. It is an even mixture of very popular and huge software suites—such as Pinnacle Studio 14 or Nero 9 Full—and useful freeware applications such as Skype, WinZip and Opera. We even selected some utter junk (for example, Screensavers) that we very often find on typical PCs used by kids. A full list of products we used is available for download here. We call this the “Junk PC”.

Selecting, downloading, cataloging, and especially installing these applications took literally more than two days and over 50 restarts. There was tremendous click-work and lots of waiting involved, believe us. Also, a lot of free and even commercial programs come with additional software such as toolbars. So when we speak of 200 additional programs, in reality it might come close to 250 due to tons of toolbars and advertising programs that came with the programs. We ran each of these programs at least once, to ensure that they were properly set up and working.

The machine we used was an ultra-thin laptop (a MacBook Air running Windows 7 via BootCamp) with a 1.86 Ghz Intel Core 2 Duo L9400 processor, 2 Gigabyte of main memory, a GeForce 9400M graphics chipset and a 120 Gigabyte serial ATA hard drive.

The question is: Is there performance degradation? And if so, how much are we talking here? Do all the additional files, registry entries, background processes, and startup items of those 200 programs slow down the computer noticeably? Let’s find out!

And here they are… the results

So after almost going crazy from installing literally 200 applications, the fun part starts. Do a plethora of applications affect performance in any way? Here’s what we gathered:

Boot-up performance

The time it takes the machine once the power switch has been pressed to finally loading the last automatically running program on the startup list.

Windows 7 took over 5 ½ minutes longer to finally boot up

After the installation marathon, we basically restarted the computer at least 40 times and used the machine over a three day period to ensure that Windows 7 had optimized the startup process as best as it could. But still, it needed over 7 minutes to fully boot up and load up all the startup programs. Although the desktop was visible after about 2–3 minutes, due to the massive load it was impossible to use the machine.
Note: On a couple of boot up attempts, the machine simply froze a couple of seconds after the desktop displayed.

Shutdown performance

The time it takes from clicking the “Shut down” button to actual power down.

During shutdown, Windows 7 simply froze up

Apparently, some of the new services and background processes caused Windows 7 to freeze during shutdown. We waited literally more than 20 minutes, but nothing happened. In only one of about a dozen shutdown attempts did the machine actually power down correctly—and that took exactly 2 minutes and 30 seconds.

Virus scan performance

We used Microsoft Security Essentials to test how long it takes to scan 1.5 GByte worth of RAR archives.

It took MSE nearly twice as long to complete the virus scan after the 200 programs had been installed on the machine.

Application performance

Here we measure the performance of selected applications. We start by measuring the time it takes for Outlook 2010 Beta to load and display the first e-mail. We move over and start Google Chrome with 8 tabs. We did not measure how long it took to actually display the Web sites, as that depends entirely on the Internet connection.
In the last run, we tested how long it took our test computer to fire up Windows Media Player and start playing a 1080p Full HD video file.

No surprise here, Outlook took almost twice as long to start up—same with the Google browser. We’ve also waited twice as long until the Full HD video clip started playing in Windows Media Player. This is absolutely intolerable. It felt like we were using Windows 7 on a 10 year old machine and not from a laptop bought in 2009.

Processor, graphics, and memory performance

We compress three video files (each about 350 MByte in size) to a single ZIP file, which took a significant amount of time—even before we “ruined” our machine with the massive amount of programs we put on this little machine. After that, it took almost 45 seconds longer—probably due to the massive load put on the CPU and the hard disk.

The video file wouldn’t even play smoothly for more than a couple of seconds. On one test run, Windows Media Player simply froze up on us.

Photo editing performance

Using PhotoImpact X3, we took a 30 MByte TIF graphic and applied the “Enhance” filter that is supposed to enhance colors and picture quality. We measured the exact time needed for the filter to take effect on the image.

Surprise, this rather GPU-intense task took only 3 seconds longer on the junked up configuration. Not a drastic performance reduction that we experienced in basically every other benchmark we ran.

Cinebench R10 performance

Cinebench is a leading benchmark for 3D animation performance based on the very popular CINEMA 4D animation software. The benchmark generates a score that represents the performance of a computer and that can be used to compare performances between different configurations. The higher the numbers, the better the performance of the PC.

It was no surprise either, that rendering the 3D animation took noticeably longer on the Junk PC configuration. This resulted in a significantly lower score on each test run of Cinebench R10.

S.T.A.L.K.E.R. Call of Pripyat performance

A benchmark based on the game S.T.A.L.K.E.R. Call of Pripyat released in Q4 2009.

Stalker suffered a great deal. The already pretty low-frame rate went down to an absolute unbearable 1–3 frames per second, depending on the scene and details on screen. It seems that games with massive textures and details suffer the most from a junked up computer.

PC Mark Vantage performance

PC Mark Vantage is a set of benchmark tools designed to test the speed of the CPU, and read/write speeds of the memory and hard disks. It also tests the machine by automatically performing tasks such as playing multiple video files, editing photos, or filling up WordPad with dozens of pages of text. At the end of the test, the benchmark calculates a score that is used to determine the performance of a computer. The higher the score, the faster your machine performs. We used the “default run” here.

It was shocking to see the total PC Mark Vantage score go down from 2624 to 1092 due to the massive amount of programs installed. We were, however, surprised that both the HDD test, the music and the gaming test were not affected that much from all this new background activity going on.

Overall performance and stability from the user perspective

Benchmarks test in pure numbers. What we test in this last “benchmark” is the performance from our point of view: The overall responsiveness when launching and switching between programs and Windows or when clicking on menus.
After the 200 programs were installed, you had to wait a few seconds until any window opened. It took about 25 seconds until the “Computer” and the “Control Panel” windows opened. Some programs did not even launch anymore or caused the computer to lock up. It took a few seconds until even the smallest context menus showed after a right-click. Granted, the system we used is not the most powerful high-end machine, but it is not out of date by any means. Still, everything seemed to run as slow as molasses.

Conclusion: massive performance hit! Advice: limit what you install

We expected a noticeable drop in performance before we started this experiment, but certainly not over 200% in some benchmarks. By installing that many programs, we simply ruined our machine to the point where it was absolutely no fun at all to work with it. Starting with the boot process: It was incredibly slow and after the logon over two dozen advertising windows, pop-up balloons, and program logos appeared, before we could click on anything. Even a couple of error messages showed up.

The performance of the poor machine became horrid, as it took over 30 seconds to start Windows Explorer even after the machine fully booted. Opening even a very lightweight application like Excel 2010 took almost a minute. It was simply embarrassingly slow.

We suggest a couple of rules

• Before you install any piece of software, think long and hard if you actually need it or plan to use it regularly. If you just need it once, remember to uninstall it.
• Go through the list of all your installed programs and decide if you still need all of these programs. Check out the TuneUp Blog Diary: Removing Unused Programs on My Mom’s Windows PC post.
• To further clean up your system, read the How To: Disable Unnecessary Features in Windows Vista and 7 post.
• Be sure that only the most necessary startup applications are loaded automatically when you turn your PC on. Read our post How to Reduce Windows Startup Time post for more information.
• Install new software within a virtual machine. That’s a simulated PC with an operating system! You can basically wreak havoc in this environment and test all kinds of programs without affecting your own PC. Try out free solutions from Sun or Microsoft. Note: You need a Windows XP CD, Vista DVD, or Windows 7 DVD and an additional license to install the virtual operating system.

Stick to our advice and watch out which and how many programs you install on your machine. Yes, this test was an experiment, but we have actually seen many computers in the last couple of months that have been messed up like that. In an upcoming blog post, we will show you the best way to handle even these chaotic machines without actually having to reinstall them. Stay tuned!

Written by Christoph Laumann and Sandro Villinger

Read the next part of our 200 programs experiment – Can the original performance be restored by removing all of these unnecessary applications?

Installing the latest Catalyst driver

This guide is based on the latest Catalyst 11.1, which was published on January 26th, 2011. To download the latest driver, go to http://support.amd.com , and select the driver that’s most applicable to your system. For example:

Selecting the Right Driver for Your Windows Operating System and Graphics Card

Click on “View Results”, then download and install the full package (Option 1), which features the Display Driver, the ATI WDM Integrated Driver and the Catalyst Control Center.

ATI Catalyst 11.1

Tweaking for the best performance in the ATI Control Center

To open up the ATI Control Center, right-click on your desktop, and select “ATI Control Center”. In the “Welcome” screen, click on “Graphics”, and select “3D”. To make the right choices in each section, carefully read the following:

Anti-aliasing (AA)

Description: AA reduces the pixelation of edges and objects in games thereby creating a smoother looking environment. The higher the setting is set to, the less jagged all objects in games will appear. However, the graphics card has to calculate many more pixels, which can lead to a massive drop in performance. This will certainly be the case if you select higher AA settings, such as 8x or 16x; this also depends on what your graphics card supports.
Our Advice: Stick to 2x or 4x AA, which will allow for smoother images, with only a minimal reduction in performance. However, if you own a weaker ATI card, disable AA to get more frames per second (FPS).

Adaptive Anti-aliasing (AAA)

Description: This setting improves AA for transparent textures.
Our Advice: We recommend enabling this feature. In our tests, with a Radeon HD 3650 and a Radeon 4850, games only dropped about 1–2 FPS.

Anisotropic filtering (AF)

Description: The higher this setting is set to, the sharper distant textures will appear. For example, when you use the maximum AF setting in a racing game, the textures of the street in the distance will appear sharper. However, this increases the processing load on your graphics card.
Our Advice: Anisotropic filtering does not reduce performance noticeably when it’s set to 2x or 4x. For a faster game play, stick to these lower settings—especially in modern games with more detailed textures.

Catalyst A.I. (A.I.)

Description: This setting intelligently and automatically reduces image sharpness and the level of detail thereby noticeably improving performance. In most games, even with the “Advanced” setting, which aggressively reduces image quality, we did not notice any visual differences.
Our Advice: The setting will boost performance, so make sure it is enabled. Keep in mind that A.I. can interfere with 3D benchmarks, such as 3DMark Vantage. To get an actual benchmark result, turn off A.I. before running the tests.

Mipmap

Description: This slider controls the visual quality of all textures in games. This article describes in detail how mipmaps work.
Our Advice: We recommend leaving this slider on the standard (performance) position. Reducing or increasing the visual quality may result in glitches on some graphics cards.

ATI Control Center

You can achieve the right balance of performance and visual quality with nearly all of the settings described in this post. There is no right way to configure your ATI Control Center—it depends on what games you play and the power of your ATI graphics card. If you have a relatively new graphics card but play games that are a couple of years old, we recommend adjusting all of the options to the maximum settings (except Mipmap). For more modern games, such as Crysis 2 or Call of Duty Black Ops, moving to the maximum settings can result in a noticeable performance hit, since only the most advanced ATI cards (such as the new HD 6000 series) are able to handle these settings.

The myth

Unfortunately, some technical books and articles discussed the “Dr. Watson” program that apparently logs all of the activities on your Windows XP, Vista, and 7 machines and sends them to Microsoft. This was rumored to not only compromise your privacy but also decrease the performance of your PC.

To disable the “Dr. Watson” program, various sources recommended opening the registry key “HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows NT\CurrentVersion\AeDebug”, double-clicking on “Auto” and typing in “0″. But does this really help?

What is the “Dr. Watson” program?

“Dr. Watson” is a debugging tool for applications. If a program crashes, “Dr. Watson” is able to jump in, collect data and, if you choose to do so, manually transfer this data to support personnel to diagnose the problem. The data is stored in the “Drwtson32.log” file and can be immediately sent to Microsoft. Some additional background information on “Dr. Watson” within modern Windows operating systems:

• Microsoft developed an advanced technology based on “Dr. Watson” that was introduced in Windows XP. Known as “Windows Error Reporting” (WER), this technology automatically offers to send debugging information to Microsoft, thus helping the company to solve most of the problems in the operating system. So, “Dr. Watson” is almost obsolete in the operating system but is still there if someone needs it.
  

• “Dr. Watson” is only available in Windows XP and earlier Microsoft operating systems. It was replaced with the “Problem Reports and Solutions” applet of the Windows Vista Control Panel, which is called “Problem reporting” in the Windows 7’s Action Center.
  

As you can see, “Dr. Watson” is nothing more than one of the many legacy tools in Windows XP.

Debunking the myth

First of all, “Dr. Watson” does not run at any time during regular Windows XP (SP3) usage, which we’ve monitored for long periods of time. Even during a couple of application crashes, “Dr. Watson” does not magically turn itself on; instead, WER appeared.

To further debunk this myth, we conducted a small series of benchmarks using PCMark Vantage (To see how we conduct benchmarks at the TuneUp Blog, read this blog post.) Here are the performance results—both before and after disabling “Dr. Watson”.

PC Mark Vantage

As you can see, the productivity benchmark, PCMark Vantage, did not show any noticeable differences, regardless of whether “Dr. Watson” was enabled or disabled. “Dr. Watson” doesn’t affect performance in Windows XP, and it’s not even a factor if you’re using Vista or 7.

Are there any tweaks that you think don’t have a purpose or that you’ve always doubted? Leave us a comment; we’d love to prove if it’s just a myth—or if there’s any truth behind it.

Installing the latest ForceWare driver

The following options and recommendations are based on the latest available ForceWare drivers (266.66 for the XP 32-bit version, and 266.66 for Vista and 7 32-bit and 64-bit versions).

Download the files, extract them, and install the driver by double-clicking on “setup.exe”

Notebook users, make sure you click on “INF Modified”, download the nv_disp.inf, and overwrite the original nv_disp.inf in the extracted driver folder.

Open NVIDIA Control Panel

To open up the NVIDIA control panel, right-click on your desktop, and select “NVIDIA Control Panel”. Click on “Adjust image settings with preview”. You can easily change the visual quality of games just by moving the sliders. Click on “Use my preference emphasizing”, and move the sliders to the middle to balance the visual quality and performance of games.

Adjust Image Setting with Preview

If you move the slider all the way to the left, the preview will appear jagged as the smoothness of curves (also known as anti-aliasing) will be disabled. Anti-aliasing is one of the most performance-hogging features—especially for lower-end graphics cards—but it allows for smoother visuals.

Manual control over all 3D settings

Changing the anti-aliasing level of games is only one way to improve the performance or quality of your games. For settings that will have a positive impact on frames per second (FPS) or visual quality, click on “Manage 3D settings” in the NVIDIA control panel.

Manage 3D Settings

Ambient Occlusion

Possible Options: On, Off
Description: This setting can enhance realism in some games by improving lighting and shadows. However, by enabling Ambient Occlusion, the FPS will drop significantly in most cases. We advise you to try it out and see if the visual improvement is worth the drop in performance.

Anisotropic filtering

Possible Options: Application-controlled, Off, 2x, 4x, 8x, 16x
Description: This filter makes all kinds of textures in games appear more crisp and clean, especially when looking at objects that are far away. But, the higher you set Anisotropic filtering, the less FPS you will get out of your game. If you set it to the standard “Application-controlled” option, games can choose between the levels of filtering. If you choose another option, the games’ settings will be overwritten.

Antialiasing – Gamma correction

Possible Options: On, Off
Description: This setting will allow for improved color quality in games, and its effect on performance is minimal. That’s why we suggest you make sure it’s set to “On”.

Antialiasing – Setting

Possible Options: Application-controlled, 2x, 4x, 8x, 8xQ, 16x, 16xQ
Description: Like the previous setting described, this is one of the most important options for games. The higher the Antialiasing is set, the smoother all images in a game will appear. However, especially if you use Antialiasing 8x or above, you will notice a performance hit. The “Q” modes, only available on GeForce 8000 series or newer, will further improve the visual quality of the anti-aliasing, but this is hardly noticeable – if at all. Again, you should probably test each of these settings to find the right balance of quality and performance.

Antialiasing – Transparency

Possible Options: Off, Multisampling, Supersampling
Description: The transparent anti-aliasing mode adds more realism to curves in games. Try the “Multisampling” option, since it won’t noticeably affect the performance yet will increase the visual quality.

Maximum pre-rendered frames

Possible Options: 0–8
Description: This controls the number of frames that the processor prepares before transferring them to the graphics card. Increasing this value results in smoother game play, but you may notice a lag when using the mouse and keyboard. To eliminate the lag, try the “0″ setting.

Multi-display/mixed-GPU acceleration

Possible Options: “Multiple display performance mode”, “Single display performance mode”, and “Compatibility display performance mode”
Description: If you have only one display, you should select “Single display performance mode”. If you have two or more displays, select “Multiple display performance mode”. If you notice glitches or strange textures on a multi-display setup, go for the “Compatibility display performance mode”.

Power management mode

Possible Options: Adaptive, Prefer maximum performance
Description: Only available on the GeForce 9000 series, the Power management mode is designed to adapt the performance of the graphics card to the needs of the game. If you are on a notebook and only play a lower-end game, you should stick to the “Adaptive” setting. However, if you want full performance at all times, go with the “Prefer maximum performance setting”.

Texture filtering – Anisotropic sample optimization

Possible Options: Off, On
Description: Set this setting to “On” to improve your performance. Keep in mind that, in some games, you may notice a slight decrease in the visual quality of textures.

Texture filtering – quality

Possible Options: Quality, High quality, Performance, High performance
Description: This function specifies how all textures (e.g. wall textures and character textures) will appear in a game. On lower-end hardware, you might want to try the “Performance” or even the “High Performance” settings to get more FPS.

Threaded optimization

Possible Options: On, Off
Description: This option should always be on “On”, as it allows the support of multi-threaded optimization for modern multi-core processors.

Vertical sync

Possible Options: Use the 3D application setting, Force On, Force Off
Description: Vertical Sync (or Vsync) synchronizes the frames that your graphic card renders with the refresh rate of your monitor. If you disable Vsync, you might find that games run more smoothly, however, in most cases, you will notice that some parts of the screen might not be rendered correctly and appear to lag. More explanation is provided over here.
Disable Vsync only if your monitor has a higher framerate (for example 60hz) than your gameplay (for example 40 FPS).

That’s it! We have covered the most important 3D settings in the NVIDIA control panel. Hopefully, this post has helped you find the right balance of gaming performance and visual quality. Is there another option that you didn’t find here? Post a comment, and we’ll explain the setting and give you the right recommendation!