Firstly, thank you for the copies of Millard’s copies of A3 and KAV 21-24. Secondly, great illustrations you have placed on the forum superimposing your arrangement of limus onto copies of these documents.
Maybe you could point to the exact reference you are using because I can not find it. On the contrary, in my internet search, I find the complete opposite.
The first construction I find is in George Smith’s 1785 book, “The Assyrian Eponym Canon”, at
with relevant pages being 40-41 and 69-71. His arrangement was
659 Silim-Assur 658 ? 657 ? 656 Sa-nabu-su and for him Assur-gimil-turri was to be placed after 644
In 1902 the Rev. C.H.W. Johns in “The Chronology of Ašurbânipal’s Reign. I.”, PSBA 24 (1903) 235-241 published the first in a series of papers on the chronology arrangement of the early to mid limus of Aššur-bani-apli. His summary table is here
Subsequent papers in 1903-1907 explained his reasons for revising Smith’s assignments. His arrangement forms the basis of the modern conventional scheme and is
659 Silim-Ašur 658 Ša-Nabű-šű with Aššur-gimillu-tere being placed sometime after 644.
When I turn to Schroeder’s 1920 explanation of Canon A3, here at the top of page VIII
we see that he considered Aššur-gimillu-tere (which name he reads as “Ášur-bęl-la-mur?”) to be a post canonical and therefore several years after 659.
Finally Ungnad. In his 1938 article “Eponymen” in RlA, on pages 428 and 429 his summary has the following arrangement
659 Silim-Aššur 658 Ša-Nabű-šű with Aššur-gimilli-terra sometime after 648
So I say again, for over 100 years, no one but yourself has placed Aššur-gimillu-tere (and others) between Silim-Aššur at 659 and Ša-Nabu-šu at 658. All place him (and the others) in the post canonical period. Clearly therefore no one considers him to be the last limu of column IX of KAV 21-24.