I'm going to develop an application for android that uses the whetstone algorithm to measure CPU performance. I've chosen the Whetstone algorithm because my research tells me it's an appropriate way of measure performance of less powerful cpu's.
I've got the source code and pseudo code(from the sweet 60's) for the whetstone algorithm, and so far so good. But the whole whetstone algorithm seem's a bit secret, and it's hard to find useful information about it. So my questions are:
- Why are the Whetstone algorithm suitable to measure performance of less powerful cpu's?
- In brief, can anyone tell me what exactly makes an cpu performance algorithm being a whetstone algorithm?
- Can anyone explain in short, the pseudo code of the Whetstone algorithm?
Answer on any of this questions is really appreciated?
Back when Whetstone was developed processor performance was measured in KWIPS where the K stands for "kilo" which is either 1000 or 1024 depending on who you ask.
A recent Core i7 benchmark put most modern CPU performance in the GWIPS where G means "giga" and is roughly a billion, one million times more than kilo.
Considering that phone processors are considerably more powerful than the processors in the 60s and 70s and less powerful than a Core i7 I think it's perfectly suitable.
As far as an explanation of what the Whetstone actually does: it's made up. It's purpose is to execute many instructions to see how powerful a processor is. There's no input data, no computed output, nothing. It doesn't actually do anything useful while it's running. The usefulness of Whetstone is that you can measure how long it takes to execute and thus determine how fast the processor is.
http://www.roylongbottom.org.uk/whetstone.htm
http://www.guru3d.com/articles_pages/core_i7_3770k_and_3750_review_with_z77,15.html