Add pp_tttt as an example of a process with many P1 subdirectories which need different nwf values

[valassi]

example of process with two matrix-element in the same directory:
define p = j / g
define j = p
generate p p > w+ w+ j j

[valassi]

To make it clearer

• pp to something creates two subdirectories: qq to something and gg to something

• qq to W something has nprocesses>1 within one qq subdirectory (because not all qq pairs are equivalent in MEs, but they are still integrated at the same time)

[valassi]

A first example of a process with two P1 subdirectories is pp to 4 tops. There are issues there, MR #560 is in WIP

[valassi]

As correctly pointed out by @zeniheisser the issue here is that some parts of mgOnGpuConfig are P1-specific, so there should really be a separate mgOnGpuConfig in each P1 subdirectory. Or, some parts of these should be moved back to the CPPProcess class where @oliviermattelaer had essentially added them initially. This is essentially what the cleanup issue #266 should all be about... (also vaguely related to #356)

[valassi]

I renamed this, to mention

• that the issue is not ONLY different P1 subdirectories, it is also/especially having them use different nwf values
• that pp_tttt (in WIP PR add pp_tttt to codegen and fix its builds (two P1 subdirectories which need different nwf values) #560) is an example of this issue

Incidentally, note instead that we now ALREADY have an example with several P1 subdirectoeis: this is gq_ttq (PR #626). There I added tput/tmad test scripts which only test one directoy (I have just created #636 about eventually testing both, low priority)

The followup of this issue is best done in PR #560, where @zeniheisser already added useful comments. I am closing this now to avoid unnecessary confusion.