Dear Referee,
thank you for your positive review and kind words.

I have added the four references you recommended (and a few related ones). As this was in any case going to lead to (a fraction of) one more page, I also did add a very short paragraph in the introduction explaining how our work is related to and differs from these.

In Fig 1, I chose to explicitly indicate what "gt00256" referred to (this was already mentioned in the caption, but the link was not obvious).

In Fig 2, I have modified the caption to introduce the word Hyper-Threading and indicate the number of physical and logical cores in the node.

Thanks again, kind regards
Andrea Valassi

---

Dear Andrea,

Please find the referee's comments below:

This is a relevant and clearly written proceedings contribution, highlighting the progress of the Madgraph5_aMC@NLO vectorisation/GPU offloading project since the earlier ICHEP2022 proceedings.

I recommend accepting this contribution after some minor improvements:

There need not be much discussion (given the page limit), but the introduction should mention earlier work on porting matrix element calculations to GPU (such as 0909.5257, 1002.3446), as well as more recent/ongoing efforts (2106.06507, 2106.10279).
In Fig. 1, the framed label "gt00256" does not seem to be explained or to add any information. Perhaps it could be removed for clarity.
In Fig. 2, the last part of the plot title could be modified as follows: "2x HT" -> "2x hyperthreading (HT)". This would introduce the HT abbreviation, which is also used within the figure, and clarifies to the reader what is meant here. I think this is sufficient and it is not necessary to mention this in the main text further, as it is not essential for the rest of the discussion.