09-15-2018, 12:08 PM
There's been a lot of "How much faster than the Rev. C Prime is the G2?" so I decided to put together a kind of benchmarker that you can use to compare the two.
The benchmark in itself is probably of limited value in absolute terms but gives a good idea of the differences between different machines.
This is a "suite" of 3 tests built into the same app. The "Savage" benchmark gives an idea of the accuracy of the engine, the "Summation" test times how quickly the machine can chew through calculations involving transcendental functions and, finally, the "N Queens" test gives an idea of how well the machine deals with decision making and branching (no floating point is involved).
Download link: https://s3-eu-west-1.amazonaws.com/xferp...appdir.zip
As soon as you start the app you'll get a menu in which you can choose the number of iterations for the first two tests and the size of the board for the N-Queens test:
![[Image: menu1.png]](https://s3-eu-west-1.amazonaws.com/xferperso/forums/primebenchmark/menu1.png)
![[Image: menu2.png]](https://s3-eu-west-1.amazonaws.com/xferperso/forums/primebenchmark/menu2.png)
![[Image: menu3.png]](https://s3-eu-west-1.amazonaws.com/xferperso/forums/primebenchmark/menu3.png)
Hit the [OK] soft key when you're done and let your Prime go away and crunch numbers for a while (you'll see the animated hourglass in the top-right corner of the screen). Quick tip: don't leave your Prime connected to your computer while doing this because the constant chatter over the USB link does slow the Prime down.
When it's finished running the tests, it'll spit out some statistics for you in the Terminal screen. The results I get below are firstly from one of my Rev. C Primes, then from my G2 Prime and finally from the Virtual Prime running on my Core i5 3GHz with 16GB DDR4 RAM.
![[Image: revcresults.png]](https://s3-eu-west-1.amazonaws.com/xferperso/forums/primebenchmark/revcresults.png)
![[Image: g2results.png]](https://s3-eu-west-1.amazonaws.com/xferperso/forums/primebenchmark/g2results.png)
![[Image: vpresults.png]](https://s3-eu-west-1.amazonaws.com/xferperso/forums/primebenchmark/vpresults.png)
The benchmark in itself is probably of limited value in absolute terms but gives a good idea of the differences between different machines.
This is a "suite" of 3 tests built into the same app. The "Savage" benchmark gives an idea of the accuracy of the engine, the "Summation" test times how quickly the machine can chew through calculations involving transcendental functions and, finally, the "N Queens" test gives an idea of how well the machine deals with decision making and branching (no floating point is involved).
Download link: https://s3-eu-west-1.amazonaws.com/xferp...appdir.zip
As soon as you start the app you'll get a menu in which you can choose the number of iterations for the first two tests and the size of the board for the N-Queens test:
Hit the [OK] soft key when you're done and let your Prime go away and crunch numbers for a while (you'll see the animated hourglass in the top-right corner of the screen). Quick tip: don't leave your Prime connected to your computer while doing this because the constant chatter over the USB link does slow the Prime down.
When it's finished running the tests, it'll spit out some statistics for you in the Terminal screen. The results I get below are firstly from one of my Rev. C Primes, then from my G2 Prime and finally from the Virtual Prime running on my Core i5 3GHz with 16GB DDR4 RAM.
