AstroSkipper

You are somehow talking your head off here. You and only you were trying, statistically completely absurdly, to make a statement about how much faster my CPU is than yours from artificial, completely arbitrarily defined values. And by that I mean the CPU score, since you didn't offer detailed values. That is statistically and mathematically pure nonsense. If you really want to know more about our CPUs in a more meaningful way, you have to compare the values in the section CPU Test Suite to get information how many times faster my CPU is than yours. So please be more serious when making statements!

Unfortunately, nonsense. You didn't provide your values. Only the values from the System Test Suite via screenshot. Where is the list of all values? You only get them via exporting from PCMark05. Look at my post: https://msfn.org/board/topic/186451-overall-performance-comparison-between-different-computers-under-windows-xp/?do=findComment&comment=1272362 I mean the detailed values I posted in the code area. And the values in the section CPU Test Suite have to be compared to get information how many times faster my CPU is than yours.

That seems exaggerated to me. I have disclosed all my values in detail, even from the CPU Test Suite. Where are yours? Without these values, your statement is just conjecture and not statistically proven. You just wanted to set a value for calculating a geometric mean to zero, and now you want to be able to read from an artificially, completely arbitrarily obtained value how many times faster my CPU is than yours. That's very funny.

Test completed. Unfortunately, as reported, no PCMark Score was calculated due to two tests in the System Test Suite that failed: Here are my results: My self-calculated PCMark Score gained by reducing the discrete characteristic values is 1290. Of course without any guarantee, as it is not quite clear which units PCMark05 uses in its algorithm. But I think I did it right. Here are my detailed values: <<< System Information >>> ------------------------------------------------------------------------------------------ System Model MICRO-STAR INTERNATIONAL CO., LTD MS-6391 Processor Intel(R) Pentium(R) 4 CPU 2.80GHz Physical Memory 1.5 GB Graphic NVIDIA GeForce 6200 Video Memory 256 MB Operating System Microsoft Windows XP (5.1.2600) 32-bit Application PCMark05 <<< Result >>> ------------------------------------------------------------------------------------------ < Main Test Results > PCMark 0,0 PCMarks CPU 3.049,0 Memory 2.034,0 Graphics 1.335,0 HDD 5.107,0 < System Test Suite > HDD - XP Startup 11.035.388,0 B/s Physics and 3D 76,2 FPS Transparent Windows 391,1 windows/s 3D - Pixel Shader 33,5 FPS Web Page Rendering 1,6 pages/s File Decryption 54.567.500,0 B/s Graphics Memory - 64 lines 354,6 FPS HDD - General Usage 7.412.013,0 B/s Audio Compression 0,0 Video Encoding 0,0 Text edit 42,1 pages/s Image Decompression 8.869.950,3 pixels/s File Compression 2.236.637,5 B/s File Encryption 14.321.510,0 B/s HDD - Virus Scan 17.429.826,0 B/s Memory Latency - Random 16 MB 4.948.194,5 accesses/s < CPU Test Suite > File Compression 8.475.572,0 B/s File Decompression 92.539.640,0 B/s File Encryption 61.333.580,0 B/s File Decryption 51.141.084,0 B/s Image Decompression 19.719.909,7 pixels/s Audio Compression 2.027.561,7 B/s File Compression 4.393.688,0 B/s File Encryption 29.755.038,0 B/s File Decompression 23.497.330,0 B/s File Decryption 13.191.807,0 B/s Audio Decompression 532.577,2 B/s Image Decompression 5.039.531,7 pixels/s < Memory Test Suite > Memory Read - 16 MB 926.653.120,0 B/s Memory Read - 8 MB 928.947.840,0 B/s Memory Read - 192 KB 21.672.546.304,0 B/s Memory Read - 4 KB 39.685.492.736,0 B/s Memory Write - 16 MB 901.376.320,0 B/s Memory Write - 8 MB 900.723.200,0 B/s Memory Write - 192 KB 12.219.063.296,0 B/s Memory Write - 4 KB 12.250.072.064,0 B/s Memory Copy - 16 MB 867.653.760,0 B/s Memory Copy - 8 MB 862.753.856,0 B/s Memory Copy - 192 KB 7.173.510.656,0 B/s Memory Copy - 4 KB 12.217.476.096,0 B/s Memory Latency - Random 16 MB 4.948.194,5 accesses/s Memory Latency - Random 8 MB 5.298.853,9 accesses/s Memory Latency - Random 192 KB 143.647.140,5 accesses/s Memory Latency - Random 4 KB 1.404.406.372,1 accesses/s < Graphics Test Suite > Transparent Windows 392,7 windows/s Graphics Memory - 64 lines 354,1 FPS Graphics Memory - 128 lines 248,9 FPS WMV Video Playback 19,7 FPS 3D - Fill Rate Multi-Texturing 1.334.127.441,4 texels/s 3D - Polygon Throughput Multiple 8.536.705,0 triangles/s 3D - Pixel Shader 33,6 FPS 3D - Vertex Shader 5.565.621,9 vertices/s < HDD Test Suite > HDD - XP Startup 10.599.348,0 B/s HDD - Application Loading 8.399.760,0 B/s HDD - General Usage 7.344.671,5 B/s HDD - Virus Scan 63.417.920,0 B/s HDD - File Write 43.694.524,0 B/s For my PCMark Score calculation I took MB/s instead of B/s as shown in the details here: Comparing my values with @NotHereToPlayGames's ones, his Intel Atom CPU seems indeed to be weaker than mine.

So are you going to fill in a "zero" for the failed tests? Or just geomean a smaller subset? A geometric mean is the nth root of a product with n factors. If you set one value to zero, the whole product will be zero and the nth root of zero is zero. My answer to your first question is therefore no. You can't set a value from your data to zero. And yes, you have to reduce the set of values, i.e. you have to build a smaller subset by removing characteristics.

Small interim info. All scores are displayed except for the PCMark score. It seems that one test from the systems test suite fails, and therefore no score can be calculated as a multiple of the geometric mean of all individual test results in the systems test suite. I will calculate the PCMark Score from the successfully performed tests in the systems test suite myself.

@NotHereToPlayGames BTW, you unfortunately have to select the option "Adjust for best appearance" for the test only. You can switch back to your settings after all is done.

Ok. I thought I had to submit first to get the PCMark Score. After performing all tests, I got all scores except the PCMark Score which had the value n/a. I will install Windows Media Encoder 9 and see if it helps. BTW, I thought it was already there as I installed all Windows Media compinents in the past. A bit strange.

Correct. You have to click the Submit button to upload your results, and then you get a PCMark Score.

This is the only official source by the developers. Completed all tests. All eye candy in Windows enabled by default. I was missing WM Encoder, so I installed it, wasn't sure if 64-bit would be fine, so installed 32-bit just in case, but otherwise, I never needed it before, I rarely manipulate media files. Disk tests were performed on 1 TB HDD WD WD10EZEX, memory is DDR2, 2x 2 GB Mushkin 996671 in first 2 slots and another 2 GB no-name Kingston in 3rd slot, all supposedly 800 MHz, but the mismatching Kingston slows things down to 667 MHz (add some MHz due to slight overclock), guess due to timings. Ok. You was able to submit your results to get a PCMark Score. Great! Under my Windows XP 32-bit, I wasn't, unfortunately. After finishing all tests, I got an error when trying to submit my results. Either I have to install Windows Media Encoder 9 or I have to update FutureMark SystemInfo as it is perhaps outdated. I will try again later.

We will compare both computers as they are. And that means I won't pull a 512 MB module out of my computer. They are very old and are only removed if really necessary.

https://www.softpedia.com/get/System/Benchmarks/PCMark05-Basic.shtml

No. We do not violate the MSFN Forum Rules. This version is provided as a full version for everyone. It's a legal and well-known website CHIP.de. The tool PCMark05 is abandoned software and offered to be used on old systems for free.

No. Under no circumstances. I have enabled only two settings under Performance Options: "Smooth edges of screen fonts" and "Use visual styles on windows and buttons".

Your code is wrong. Download one of the linked versions and use its provided key!

@NotHereToPlayGames Internet testing has to be skipped, as this is not relevant for our comparison, and IE6 is obsolete anyways. So, are you in or out?

Probably every Windows XP user has it installed. Look here: https://archive.org/details/WMEncoder9Series The same applies to DirectX 9.0c. It's essential and needed for testing graphics. Since you don't really use your computer anymore anyway, it's probably no problem to install these things briefly. https://benchmarks.ul.com/en/legacy-benchmarks https://www.chip.de/downloads/PCMark05-Vollversion_13003873.html

@NotHereToPlayGames Here we are: Define the parameters of such a test and I will engage in this battle. Quantifiable measurements where neither one of us can "cheat". Then we'll let the rest of MSFN members "decide for themselves". Since the word offtopic is apparently not in your vocabulary, I hereby move our conversation to this thread created specifically for the purpose of computer performance comparisons in order to finally be ontopic again. Your proposed test procedure is far too fiddly and time-consuming. I am more in favour of using a well-known, locally installed test tool (i.e. offline tool) that tests all components and generates an overall performance index for the purpose of comparison. My suggestion: PCMark05 which still supports Windows XP.

This topic is designed to compare the overall performance of different computers with each other under Windows XP.

I am not always right. But a pure comparison of CPUs is not sufficient. Read here, for example, about the difference between SD-RAM and DDR2-RAM: https://www.transcend-info.com/support/faq-296#:~:text=DDR2 SDRAM(Double Data Rate,(double of DDR SDRAM). Maybe now you'll realise what I'm talking about. And apart from that, an Intel Pentium 4 is not the same as an Intel Pentium 4. There were different series like Willamette, Northwood, and Prescott. And within these series different FSB clock rates. So, forget about these comparison sites! The only way to compare our computers is doing the same test regarding all hardware components. P.S.: Due to my motherboard layout, there are several bottlenecks. Firstly, the extremely slow SD-RAM memory, then the very low bus clock rate and finally the AGP 4x interface, although my graphics card actually is an AGP 8x one.

Please, don't come up with your "gut feelings"! I have presented all the essential facts that you have not commented on. My presentation has nothing to do with subjectivity and hypotheses. But as I said, it's offtopic here anyway and actually totally irrelevant. Your cucumber is old and mine is many years older.

Thanks for linking! Unfortunately, such comparisons are not particularly useful. It also depends considerably on other components such as the RAM memory, north and south bridge, the bus clock, the graphics adapter, processor features, hyper-threading and so on.

Not quite true. I use both my old Windows XP computer and my Android tablet for what you call "real work". And in some special cases, a notebook with Windows 7 and Windows 10. But that was not the point here. Anyway! This thread is actually about Mypal 68, and I am happy to confirm that this browser works well on my old hardware. And much better when optimised.

Ok. As far as I can see, my Intel Pentium 4 Northwood 2.8GHz 32-bit has a higher operating frequency than your Intel Atom 1.6GHz. In all other categories, your CPU is better than mine. L1/L2 cache, thermal values, instruction sets and so on. And you use DDR2-RAM which is much faster than the old, slow SD-RAM I use. Furthermore, I assume your Intel Atom CPU supports Hyper-Threading technology, my CPU does not. Even your Intel GMA 950 graphics is more efficient than my NVIDIA GeForce 6200 AGP 4x (8x is not supported on my motherboard). And as we all know, the operation frequenzy is difficult to compare regarding completely different processors, i.e., a higher frequenzy value does not necessarily mean a higher performance. BTW, which Intel Atom CPU is it? N270? All in all, I think your old Acer is faster than my old computer.

That's a general problem here on MSFN. If members/users have problems, they are very communicative and willing to provide information. But when it comes to selflessly contributing something to the cause, a certain listlessness prevails. Very regrettable. In principle, the same people always make a contribution here to move the cause forward.