(This is the text from boot's 12-step program introducing and explaining the bootMarks. The entire article appeared in boot's June 97 issue, and accompanying bootDisk contained the portion of the bootMarks that may be distributed freely. PC Gamer will also include the same bootMarks on a furute disc as soon as we can arrange it.)
boot 12-Step
This month:
bootMarking your PC
Intro:
boot scoured the known universe in search of benchmarks that would rock your motherboard, flex your MMX, and pump those texture-mapped polygons.
And we found them.
Now you can give your system a taste of what it's like to be sent into the bootLab for evaluation.
The boot-o-meter is out and the bootMarks are in. Now, with precise benchmark scores and frame-rate results, you'll get the 411 you need. This collection of abstract and real-world torture tests will make your computer sweat, and will separate the contenders from the pretenders.
Six of the 10 bootMarks can be found in the Feature section on this month's bootDisc. Each is displayed as a button which launches a dialog box with buttons pointing to the web pages explaining that particular bootMark in detail, and another that installs it.
The bootMarks can play a role in your decision-making when buying a new system, but shouldn't be the deciding factor. No test is 100-percent foolproof, and the art of reviewing a system goes beyond benchmarking-it's looking at how a system is put together, the choice of components, the "out-of-box" experience, and much, much more.
Think your system is Kicks Ass? Step into the shoes of the bootEditors for a day, and see how your system handles the bootMarks.
-Andrew Sanchez
Step One
Tools of the Trade
The only other tool you'll need besides this month's bootDisc is a stopwatch. Try to get a digital one, but don't go overboard.
First, time how long it takes to boot up. From the moment you flip your power switch, your system's under the clock until there's no hourglass mocking you. Record that time. Now, shut down your computer via the Start menu. Begin timing when you click on the OK button, then stop when you see "It's now safe to shut down your computer" or your monitor shuts off.
Step Two
The New Meter
Set your desktop to 1024x768, 16-bit color depth, and set your monitor refresh to 75Hz. Take a deep breath, and prepare to enter the bootMarks.
Step Three
CPU/Motherboard
The main bootMark test, created by Symantec, evaluates subsystem performance-in particular, the ability of your CPU, memory, PCIset, and motherboard to work as a finely tuned number-crunching beast. If you plan to indulge in some overclocking, the bootMarks are a great way to see how much extra speed you've squeezed out of your system.
INSTALLATION: Installation is pretty straightforward. It doesn't create a Programs group so it will need to be launched from the Start Menu under Norton System Information.
NOTE: If you already have Norton System Doctor installed, you already have System Information and don't need to install the bootMarks.
Fire up the program by starting the System Information program found inside the Norton System Information folder.
Click on the Benchmark button at the bottom and step back for a second or two.
Take a gander at "Your System" on the right-hand side, and record. You may want to rerun the benchmark a few times, and get the average score.
The higher the number, the better.
THE RANGE: Scores start at zero, with 75 being the average you should expect. Only the most battle-hearty systems will ever surpass 150.
Step Four
WIN95 APPS
BAPCo's SYSmark32 for Windows 95 is an industry respected, comprehensive suite of script-based real-world application tests which gauge ability to handle the everyday rigors of word processing, spreadsheets, and more. Hey, everyday can't be frag-day.
Once the benchmarks are complete, SYSmark32 computes an Official Score, which is based on the six individual scripted tests.
The higher your Official Score, the better.
This benchmark is the longest one we run, taking anywhere from two hours for a good system, to five hours for "not-so-good" systems. It's long, but it's automated, comprehensive, and it utilizes six real-world applications.
The six categories tested and programs used are:
WORD PROCESSING Microsoft Word 7.0, Lotus WordPro 96
SPREADSHEET Microsoft Excel 7.0
DATABASE Borland Paradox 7.0
DESKTOP GRAPHICS CorelDraw 6.0
DESKTOP PRESENTATION Lotus Freelance Graphics 96, Microsoft PowerPoint 7.0
DESKTOP PUBLISHING Adobe Pagemaker 6.0
THE RANGE: Consider a result of 100 as minimal system scoring, with 200 being the average boot system. If you blast past 300, you're on fire!
Step Five:
Direct3D
Microsoft's much-maligned 3D API is the force behind the PC's explosive growth in texture-mapped polygon presentations. Direct3D benchmarking is the subject of heated controversy, but we've decided on the Terramark v1.0 benchmark-a byproduct of Eidos' soon-to-be-released Descent killer, Terracide. Why?
Because, this benchmark is the lesser of all evils.
Microsoft's own suite of DirectX benchmarks, based on abstract texture-mapped tunnels and cubes, reports less-than-accurate results.
Hellbender and Monster Truck Madness have no way of running a scripted benchmark to provide exacting results each and every time.
WizMark 2.0? We use it to test Direct3D functionality, but because the benchmark is made by 3Dfx (and is biased toward the Voodoo chipset), it's hard to stand behind it.
Not only do most chip developers dig the Terramark, it's the only Direct3D benchmark that uses an actual gaming engine to derive its score. As each subsequent scripted test runs, Terramark enables more 3D acceleration features, such as bilinear filtering, alpha blending, fogging, MIP mapping, and more. If a card fails a particular test, it's noted in the final result. There are three levels of Direct3D card functionality, with each completed level of suites producing a numerical score.
The higher the score, the better.
We take all three scores and add them up. For example, a 3Dfx-laden system may produce a Level 1 score of 267, Level 2 score of 270, and Level 3 score of 269. Add all these up and you get a Total Terramark score of 806. This is the score we present. If a level fails, we add the remaining scores. Some cards, such as Matrox's Mystique, fail the Level 3 suite, so the Total Terramark score represents the sum of the Level 1 and Level 2 scores. Others, such as the S3 ViRGE, refuse to run the benchmark at all. This doesn't mean that the ViRGE can't run Direct3D apps, it's simply a case of the ViRGE and Terramark not seeing eye-to-eye.
As Direct3D drivers mature and Terramark becomes more robust, we'll utilize new revisions and keep you up to date.
THE RANGE: If all the tests fail, you'll come up with a big goose egg-500 would be the median score. Score more than 1000, and we'll be asking you where you got that prototype AGP card.
Step Six.
SUB: Hard Drive
Adaptec's ThreadMark v1.0 tortures your hard drive subsystem, recreating intense I/O activity under WinNT and Win95 with a mixture of single and multithreaded read and write requests across a range of requested block sizes. Once the benchmark is complete, it will indicate your hard drive subsystem's data transfer rate (in MB/sec) and CPU utilization (as a percentage).
INSTALLATION: Install the ThreadMark application via the bootDisc interface and launch it through the Start Menu. (It can be found under ThreadMark in the Programs group.)
NOTE: You'll need 126MB free on the drive you wish to benchmark.
Once you launch the benchmark, simply press the ThreadMark Start button on the right-hand side and sit back.
This test will take about 30 to 45 minutes. Upon completion, you'll get a detailed report of your hard drive subsystem.
THE RANGE: A 1MB/sec result is considered bare-bones minimum scoring, with average EIDE systems scoring about 2.5MB/sec. Score anything past 5MB/sec and you're ablaze.
Step Seven:
CD-ROM
CD Tach/Pro v1.65 is the Win95-native benchmark from TestaCD Labs. Previously responsible for the equally exhaustive DOS-based CD Stone, this benchmark pokes at your CD-ROM drive and reports transfer rates at different points on the disc (to illustrate the difference between CLV and CAV drives), CPU utilization at maximum obtainable speed, full-stroke and random access times, and burst tests.
On the bootDisc is a lighter version of the test we use for benchmarking, but it'll give you an idea of how your system's CD-ROM drive stacks up against the suggested specs.
INSTALLATION: Launch CD Tach's setup program via the bootDisc interface. Once the icons are created, launch the program. When prompted, click on the Next button to proceed. It will give the throughput of your CD-ROM drive, the CPU utilization up through 6x, and a taste of the video test we run in-house (VidTach). The video included on the disc is different from the one tested and is strictly for illustrative purposes. If you're interested in getting the full CD Tach suite of tests, check out www.azstarnet.com/~gcs/.
As you go through each step, record your score. When you've finished, you'll have the stats you need.
THE RANGE: 1500K/sec is boot's minimum acceptable CD-ROM score, with 1800K/sec being the average. Flying past 2100K/sec ensures your CD-ROM subsystem of Kick-Ass status.
Step Eight
WIN95 FMV
TestaCD Labs' VidTach v1.52 streams a 640x480 video off the CD-ROM and plays it scaled full-screen at 30fps onto a 1024x768 screen. This tests your CD-ROM's sustained throughput, your CPU's raw horsepower to calculate the resized image, and your graphics card's ability to display it all.
Or maybe none of that. Most modern graphics chips have a hardware video scaler that can do all this work without any load on your CPU and without dropping a frame, but surprisingly, few graphics drivers accelerate Video for Windows, which VidTach uses.
THE RANGE: Scores on VidTach range from 0 to 100 percent frames played.
Step Nine
DOS GAMING
For DOS performance, we use id software's Quake-everyone's favorite CPU pusher and VESA 2.0 inspector. While
INSTALLATION: You'll need to have either the full or shareware version of Quake installed. The shareware version can be found on bootDiscs 02 and 04 as well as at id's web site (www.idsoftware.com). The Quake button in the bootDisc interface launches a setup application that will copy over the bootMark.dem file to the directory you specify.
NOTE: Make sure you choose your Quake\id1 subdirectory as the target for this file.
Now, drop to DOS and fire up Quake. Make sure you have Quake version 1.06 or greater. If your version is not up to snuff, we've included the v1.01-to-v1.06 patch, as well as the latest VΘritΘ Quake and GLQuake patches in the Patches folder on this month's bootDisc.
Now set your video resolution to 640x480.
What? You don't see that resolution as an option? Congratulations! Your video card is not VESA 2.0 compliant! Run out and download the latest version of SciTech's Display Doctor, or check your video card manufacturer's web site for VESA drivers. Once you've set your resolution, ensure that Quake is set to full screen-no status bars on the bottom.
Start a new game. Pull the console down by pressing the tilde key (~). Type in timedemo bootmark and hit enter. Immediately press the tilde key once and relax. Quake will now run through our little demo and at the end will spew an average frame-per-second count.
Higher frame rates are always better.
If you own an accelerated version, run the bootmark.dem with that and observe your frame rates. You should be getting at least twice the rate of unaccelerated Quake.
THE RANGE: 10fps is the minimum a DOS-based video card should do in 640x480 Quake, with 15fps being the average score. Hit 20fps unaccelerated, and you may not need Voodoo.
Step Ten
DirectX Gaming
Shiny's MDK Performance Test v1.4 prods your floating-point processing power and your video card's DirectX compliance and DirectDraw performance.
A higher score is better.
INSTALLATION: The installation will create several icons on your desktop. Run the MDK Performance Test for Win95. You'll need DirectX drivers. The 3.a drivers can be found in the bootDisc Feature interface. You'll need to restart your computer after installing DirectX. The desktop icon that launches the bootMark is the MDK Performance Test for Win95. Launch it. After five seconds, observe the bottom status bar and number labeled This Machine. This is your DirectX score.
THE RANGE: Burn 60 in your brain. This is the bare minimum a system should perform, with 90 being the optimal score. Cruise by a score of 120 and your system is the DirectX bomb!
Step Eleven
MMX Performance
Armed with the latest v4.0.0.109 patch, Equilibrium's palette-processing powerhouse DeBabelizer Pro turns into a mighty MMX-armed application, capable of drastically increasing your system's processing performance... provided you have those 57 new instructions.
This one requires your stopwatch.
INSTALLATION: Setup will direct you to install the bootMark.avi movie as well as the super palette used into a bootmark\debab directory on your hard drive. Install these anywhere you want, you'll need to open them through DeBabelizer Pro. Next, Setup will install the DeBabelizer Pro demo.
NOTE: This demo does not offer MMX compatability, which this test is supposed to test. It is here for illustrative purposes only, although it can be upgraded to the full version. If you already own it, make sure you're running the newest version. The patch files to bring it to date can be found in the \patches\debab directory off the root of the bootDisc. Both updates will need to be applied.
Now, launch the program from the Start menu. Maximize the window, go to the File menu and then Open the boottro.avi file. A window containing the movie will appear. Go to File and select Open Super Palette. Find the boottro.dbp file and open it.
Take a close look at the Super Palette window on the right and pay attention to the little Action Arrow in the upper right-hand corner.
Move your cursor over to the Super Palette's window. Click and drag that Action Arrow from the boottro.dbp window into the center of the boottro.avi window. When prompted at the Apply Movie Operation, choose Entire Movie, get your stopwatch ready, and start timing the second you hit that OK button.
When the operation's complete, stop the clock and record your time.
THE RANGE: 410secs is the bottom end of the scale, with 370secs being the comfort zone. If your system can blast past 330secs, it rocks.
Step Twelve
CPU/Disk
Microsoft's Visual C++ v4.1 delivers a swift kick to your CPU, memory, and hard drive subsystem by timing how long it takes to recompile the Microsoft Foundation Classes.
This one also needs your stopwatch.
If you own Visual C++, fire it up, go to the File menu and choose Open Workshop. Find the mfcdll.mak file that comes with Visual C++, and open it. Accept all the default values. Now, add the "libname=mfc40" command by going to the Build menu, choosing Settings and manually edit the command line.
Once you've done that, choose the Rebuild All from the Build menu, start your timer, and watch the fur fly.
When the build is complete, stop timing and record your results.
As usual, the lower the time, the better.
THE RANGE: Accept 250secs as your bottom of the barrel, with 175secs being the average time. Compile below 100secs and you'll be creating your own apps in no time.
