It is an always shifting ceiling... unless you have a tightly controlled environment. Your temp and RH will be important too. other factors are invovled in this equation. DLI is a better way to understand it, then you don't have to account for different hours of use when comparing to someone else's results or previous results etc. So, remember this is a ballpark... and a large ballpark in a less-controlled environment.
800-1000 ppfd is going to be near max over 12 hours (35-42 DLI, if i'm off on the ppfd range, the DLI is more accurate -- google DLI chart.)
over 18 hours, you need 2/3rds of power over 12 hours to acheive same DLI. 500-700ish ppfd for 18 hours.
DLI is 1:1 directly proportional to hours of use... it all makes sense when you understand the math behind it. Google a DLI chart and see how it all meshes as i described.
studies covering just a few hours difference show there is virtually no difference in yield whether you provide 40DLI over 10, 11, 12, 13 hours (all acceptable for bloom). you should only use more or less hours if you are using it to adjust what is given to the plant... I.E. more is not necessarily better in all cases. so, don't assume that. make an informed choice in this regard, not a based on whimsy.
Is there a difference with large variation of how long light is provided? this is more important for an autoflower than aphotoperiod, since there's a limited range of common sense choices for photoperiods in bloom. 10-13 hous is rational and won't cause a hormonal balance issue in bloom. Autos you may choose anything from 24 - 12 hours, give or take... sometimes autos get placed in with photoperiods in bloom. The answer is i don't know. i haven't read anything that compared this to a baseline of some sort and with a suitable sample size to be confident about what was seen/learned. I'd expect at some point you get more stretching with longer dark periods... so at some point a "40 DLI" may have a slightly different outcome if some other important factor is significantly different from a 'norm' or typical expectation.
it's all relative... so "optimal amount of light" shifts... if you udnerstand all the factors, it is easier to translate and apply this understanding of providing photons to the plant.
trial and error... let this math get you in a ballpark as far as what you provide, then you observe and react to what the plant does... internode length will be a good guide. that will tell you if it needs more or less light.
VPD will matter... you may need to feed more, but if vpd is higher than before it may take care of itself.
rates.. change over time... don't think of the plant as a snapshot.. it's an ever evolving thing with constant change. it's not ml/L of X fertilizer... it how many molecules of each needed building block you provide per day... or per second or per hour... it is a rate that we must understand. it gets complicated due to all the relativistic factors invovled with feeding or giving light etc.