Talk:Super PI

"...assistance of his dog..." I'm thinking this needs to be removed --65.1.103.130 17:48, 1 August 2006 (UTC)

Age of SuperPi

So this program is in fact that old? 11 years? I didn't realize that. How come it's still so popular? There must be other programs that do the same in less time. 85.19.140.9 21:30, 13 September 2006 (UTC)

• The popularity is because performance results are comparable with those achieved in the past. It's a continuous process of achieving world records with Super PI. User:IvanAndreevich

Integer or Floating

I doubt if computations in SuperPi are performed in ALU or FPU. Such high precision couldn't be achieved with floating numbers IMHO. Zuxy 03:45, 16 March 2007 (UTC)

• Agreed. This is an important piece of information for anyone using it as a benchmark. JoaCHIP (talk) 11:23, 30 December 2009 (UTC)

This entry is incorrect

SuperPI uses Gauss-Legendre algorithm, it says so in its help file. Do some research people. —Preceding unsigned comment added by 77.105.61.70 (talk • contribs) 17:32, 31 March 2007

Hyperlink error

OCCT.. busses and students ? i think this need to be corrected —Preceding unsigned comment added by 89.120.220.186 (talk) 00:08, 9 August 2009 (UTC)

External link not working

The externa link was not working when I tried it. Gligoran —Preceding undated comment added 13:48, 4 August 2010 (UTC).

Just tested it and it worked fine for me. --Salix (talk): 16:19, 4 August 2010 (UTC)

"Credibility Concerns"

The section titled "Credibility Concerns", is vague and contains some seriously questionable claims.

In benchmarking hardware, one is measuring the performance for solving one problem to approximate the hardware performance at solving similar problems. The accuracy of a benchmark depends mostly on how similar it is to the target. The notion presented in the article is that there was some competition to produce faster SuperPI computations, and that somehow SuperPI's "flaws" were to blame. These brings up some serious issues.

1. Why was there a competition? Was this just between a handful of people? 2. Was there any actual fraud? Where hardware manufacturers using it to promote product? Were overclockers using it to promote their techniques? The former could actually be an argument for fraud, the latter not. Because in the latter all that is happening is that the SuperPI benchmark wasn't a good measure for other problem-solving. 3. How on earth was this related to SuperPI? Adding a checksum to help detect computation errors is trivial, and error-detection methods are used in many mathematical libraries. The fact that they added one is not proof of any anti-fraud measures. If I were the developer I certainly wouldn't have cared to implement anti-fraud measures since falsifying benchmarks is as easy as typing some numbers. JSory (talk) 05:12, 7 November 2025 (UTC)