skip to content
( note: this is a translation from this german article: )
With this article we want you to show the small but pioneering benchmark MaxxPI² - PreView - Multi something closer.
This benchmark is based on calculation PI and uses to which, in contrast to the previous MaxxPI ² - PreView - Single,
use all CPU cores for computation. So, this benchmark will use complete existing computing power more efficiently
than most comparable programs on the basis of PI calculations.
We hope you enjoy reading this and look forward to your opinions.
MaxxPI² - PreView - Multi
This benchmark uses as formula the Chudnovsky algorithm, unlike the MaxxPI ² - PreView - Single
that use the Gauss-Legendre algorithm. The advantage of the Chudnovsky algorithm is, that principally, multi-core capability is possible.
MaxxPI ² - PreView - Multi use this. That means: That all available CPU cores work together on a single calculation.
It was made sure that all CPU's equally served. There is no optimization, so that no CPU manufacturer is preferred.
Intensive useage of MMX/SSE2 (if possible SSE3) and of course an extensive use of multithreading.
MaxxPI² - PreView - Multi needs at least a dual-core processor and supports in the current version 1.03
CPU's with 2,3,4 and even 8 Core's. So the benchmark will only run on all CPU's, with more than one core.
Maximum depth of calculation:
268.435.456 decimal places
Designed for following operating systems: 
Microsoft Windows 2000, 32/64bit (experimental)
Microsoft Windows XP, 32/64bit
Microsoft Windows Server 2003, 32/64bit
Microsoft Windows Vista, 32/64bit
Microsoft Windows Server 2008, 32/64bit
Microsoft Windows 7, 32/64bit
Memory usage: 
Digits of PI  Memory usage (MByte)
128K 2
256K 4
512K 6
1M 15
2M 25
4M 45
8M 80
16M 170
32M 300
64M 580
128M 1150
(*)256M 2280
 (*) only available if x64 OS is present 

CPU load:

Unlike previous programs with a similar principle, the CPU load generated with MaxxPI ² - Multi is very high.
Achieved PMU (Performance Monitoring Unit) load with Prime95's "Small FFT's" will give about 61 percent, 
MaxxPI² - PreView - Multi this will go upto almost 80 percent. The values given here have been reported
with an Core i7 without active HT.


The following eight processors were used for comparison. All running with a standard clocks
and on different operating systems. The used Intel Xeon UP W3540, we additionally benched
with three different operating systems, to show how the benchmark will perform.
  • Intel Celeron Dual-Core E1200     1.6GHz,0.5MB Cache,     2 Cores, 2GB RAM, XP 32bit
  • Intel Pentium Dual-Core E2180     2.0GHz,  1MB Cache,     2 Cores, 2GB RAM, XP 32bit
  • Intel Core 2 Duo E8200           2.66GHz,  6MB Cache,     2 Cores, 2GB RAM, Vista 64bit
  • Intel Core 2 Quad Q6600           2.4GHz,  8MB Cache,     4 Cores, 6GB RAM, Vista 64bit
  • Intel Xeon UP W3540              2.93GHz,  8MB Cache,4+4(HT)Cores, 6GB RAM, XP/Vista/Win7 32/64/64bit
  • 2x Intel Xeon DP X5460          3.16GHz,2x12MB Cache,     8 Cores, 8GB RAM, XP 64bit (dual CPU board) 
  • AMD Phenom II X2 550 Black Edt.   3.1GHz,  6MB Cache,     2 Cores, 4GB RAM, Vista 64bit
  • AMD Phenom II X4 940 Black Edt.   3.0GHz,  6MB cache,     4 Cores, 4GB RAM, Vista 64bit


If you liked MaxxPI² - PreView - Multi and this awakened your interest,
the download can be found exclusively in our download area.
We would like that you also look at our forums. Your are welcome to exchange experiences with us,
to benchmark for comparing against other users and for answering upcomming questions. On the homepage of MaxxPI²
you will find additional benchmarks MAXMEM² - Preview and MaxxPI² - Preview - Single,
which you also can download for testing.
At this point, our special thanks goes to M.Bicak for the provision of multi MaxxPI² - Preview - Multi.
If you have any questions or comments regarding this benchmark, we will answer this in our topic.
Published: Wednesday, 10 June 2009  Author: Frank Hempel
(translated by M.Bicak)