I don’t use math in performance testing, do I ?
Submitted by Ainars Galvans on Thu, 08/05/2008 - 14:26.
performance testing
I’ve seen testers recommending The Art of War or Weinberg books which are not about testing at all. I’ve also seen performance testers recommending knowledge of probability theory, statistics and modeling principles. I don’t apply this knowledge in performance testing myself – well at least not directly. I never think about things like distribution function, mean deviation, etc. Do you? Don’t I ?!
I don’t use math that match, do you?
Yesterday I gently (without even slowing down) parked my car in an empty spot where my car hardly fit in. I realized that my brain did an extremely hard job to calculate the right path, figure out vehicle’s speed and deciding at what moment to move the wheel, when to press on breaks and synchronize everything… Brain don’t use mathematics (algebra and geometry), did it? My brain is used to this task that that’s why it goes so smoothly: controlling pedals and wheel, validating and adjusting as I drive
How do I analyze performance.
I’m used to think about application performance and it happens smoothly in my brain when I see how fast client side responds, look are CPU and memory utilization, graphs created by my load tools, etc. Quite fast I come up with “it works fast enough”, “too slow”, “I need to investigate/test this one deeper”, “my tests are somewhat wrong”, etc.
I believe this is because human brain is very powerful in recognizing patters, especially as analyzing graphs. There is only an example of what I’m looking for. But my brain is actually looking for patterns in cognitive experience of a whole lifetime. I name this phenomena an informed guess and use it whenever I have enough experience.
beware of those who thinks they know…
I’m pretty good at math: modeling, statistics, probability theory… I would be able to write a great academic work about software performance and performance modeling. But I don’t do that. Moreover I used to know some people who have done such paperwork… they pretend to be performance gurus and are eager to teach me how to do proper performance testing. Just as professor using high math would teach me how to better park my car…
I don’t use math that match, do you?
Yesterday I gently (without even slowing down) parked my car in an empty spot where my car hardly fit in. I realized that my brain did an extremely hard job to calculate the right path, figure out vehicle’s speed and deciding at what moment to move the wheel, when to press on breaks and synchronize everything… Brain don’t use mathematics (algebra and geometry), did it? My brain is used to this task that that’s why it goes so smoothly: controlling pedals and wheel, validating and adjusting as I drive
How do I analyze performance.
I’m used to think about application performance and it happens smoothly in my brain when I see how fast client side responds, look are CPU and memory utilization, graphs created by my load tools, etc. Quite fast I come up with “it works fast enough”, “too slow”, “I need to investigate/test this one deeper”, “my tests are somewhat wrong”, etc.
I believe this is because human brain is very powerful in recognizing patters, especially as analyzing graphs. There is only an example of what I’m looking for. But my brain is actually looking for patterns in cognitive experience of a whole lifetime. I name this phenomena an informed guess and use it whenever I have enough experience.
beware of those who thinks they know…
I’m pretty good at math: modeling, statistics, probability theory… I would be able to write a great academic work about software performance and performance modeling. But I don’t do that. Moreover I used to know some people who have done such paperwork… they pretend to be performance gurus and are eager to teach me how to do proper performance testing. Just as professor using high math would teach me how to better park my car…