1 - You may be thinking "how can this be marked"? I have no idea what tool you used to plot your graphs, nor what platform you are using. However, I look at a random sub-sample of submissions. If necessary I will ask people to demonstrate getting the results again. "But, but, but...", a hypothetical student might ask, "what if I don't get the same results next time?". Of course, the results will never be the same the second time, but they should be of similar sizes, with similar trends. "Ah", our friend might say, "my computer has been replaced / I've upgraded my OS / Trump was elected since then". Fine, I don't care. Just do the same process on the new setup and I'll watch you do it.
2 - Note that there is some disagreement on asserts in production code, but when doing peformance optimisation you almost certainly don't want asserts in the inner loop.
3 - If you don't know how to create (and embed) vector format graphs and graphics you need to find out. This is one of the things that often makes strong final year projects look weak. No matter how good the technical content is, if the diagrams are low-resolution embedded bitmaps with DCT-ringing, then the project feels weak. Presentation is important.