Why the process time difference on 2 PC's using same parameters?

[Trankuility]

Yesterday morning, on computer 1, I ran walkforward on bid tick data from 1/1/2008 thought 11/30/2008. I had the process optimize on 3 parameters and used Default (not genetic). The entire process took around 35 mins. I did this on several instruments and they all took around 35 mins.

Later in the evening, on computer 2, I ran the same walkforward using the same exact parameters and dates and tick data and the process took several hours. Why is that? The second computer has more memory, better cpu, etc. So if anything it should go faster than computer 1. To take it a step further I then ran the same exact process on computer 3 and it too took several hours. What could cause such a huge difference in process time?

[NinjaTrader_RyanM1]

Hello Tranquility,

Part of this could be the data retrieval process. If the data was previously cached on one computer, then time is gained by skipping the download process.

Most likely it has to do with how these computers run, independent of the hardware available for them. Security software could be interfering or it could mean spyware is installed and affecting your performance. There's a lot of potential differences with the exact software that could be running.

If you are interested in pursuing this further, I would look into benchmarking utilities. This will give you a number to compare among your machines to get an idea on expected performance. An IT professional should be able to help with this or you could try looking for some benchmarking utilities on the web.

[Trankuility]

I'm positive I don't have any malware/spyware on any of the computers. I'm also positive that the 2 computers I mentioned have better benchmarks than the other computer. And Security programs, etc. are not a factor.

As for the cache, the data is not being downloaded. This is offline and using historical data in my database. However, say I run a walk forward (offline) on an instrument for 1/1/10 - 11/30/10. Then I run another walk forward (offline) on the same instrument for 1/1/08 - 11/30/10. Does the second walk forward go faster due to stored information from the 1st walk forward?

[NinjaTrader_RyanM1]

Results are not cached, but the data is. Running the same test back to back could go faster the second time since the data is already in memory.

I understand you have a clear idea of your system capabilities and how they should perform compared to one another but NinjaTrader is not a suitable benchmarking tool. Get some numbers to work with from a dedicated benchmarking utility or contact an IT professional if you're interested in comparing your hardware.