I´ve done extensive measurements in terms of RAM usage and provide some figures for you and everyone else that wonders how much RAM they´ll need:

Through many tests I´ve found that **1 year of H1 history data takes about 22 MB of RAM in Chrome**. So now you can do the math how much RAM you require to run your strategy building. Let´s take an example with 10 years of H1 data:

**Chrome: 10 x 22 MB = 222 MB per instance.**

Assuming 20 years of H1 data (RAM just simply doubles then as it is double the amount of bars) (OR 10 years of M30 data which equals the same amount of bars as 20 years of H1 data):

**Chrome: 20 x 22 MB = 444 MB per instance.**

In your case you´ve got a 8 core CPU, and to max it out, you´d run at least 8 instances, rather 16 because of SMP / HT. So let´s assume you want to build with 8 instances of 10 years of H1 data, that´s 8 x 222 MB for Chrome, totalling at 1776 MB of RAM that you´ll need in that case for all instances. If you want to completely max out your CPU with 16 instances, then it´s double of that, so 3552 MB of RAM for 16 instances.

If you now double the amount of bars loaded by switching to either 20 years of H1 data or 10 years of M30 data, and run 16 instances again, you are looking at a total of 7104 MB of RAM that will be needed for that - so it´s getting very tight already with your system in that case.

Another thing to think about is multi-market validation, in case you want to use that. In that case it requires even more RAM. **Each additional market you test on, requires about 2,5 MB per each year of H1 history data that it holds.** So let´s assume you want to validate on 10 additional markets, each market having 10 years of H1 history data, you are looking at an additional 250 MB of RAM (2,5 MB x 10 markets x 10 years of H1 data) and that is per Chrome instance once again!

So 1 Chrome instance with 10 years of H1 data and validation on 10 additional markets with 10 years of H1 data each will now require 472 MB in total (222 MB for the main data + 250 MB for the additional markets validation). Now multiply that by 8 (3776 MB) or 16 (7552 MB) and you have your new figures.

If you are wondering about Firefox and how much RAM it takes there compared to Chrome: it requires about 70% of the memory that Chrome needs (yes, more efficient JS engine, but a little slower than Chrome overall - but good if you need to save RAM). **So you can take all the above values and multiply them by 0.7 to get the approximate RAM usage in Firefox.**

Based on all this data, everyone should now be able to easily calculate the RAM required for their needs with EA Studio.