first, the efficacy is not "3.11 μmol/joule" as listed fo this light. So we need to figue out a more accurate guesstimate.
1296 samsung diodes.. unless they are all 5000K (or is it 6500k, check websit spec sheet fo the "1" the list attributes.. all others fall short of this peak by up to 15% give or take) it is definitely not 3.11umol/J, so it is a lower umol/s than listed. they are listing ppfd but not the space covered, so it's easy to manipulate the numbers like this.. it's what i consider a big red flag that the manufactuer is being manipulative, even though this still does look like a quality light.
if top bin diodes and mix of 3000-5000k etc like normal.. could be 2.9-3.0 umol/J. could be 2.7-2.8ish with lower bin diodes.. so 330w x 2.7 /* 2.9 */ = 891 (957)
or 1782 (1914) umol/s when doubled
1782 / 25 * 10.764 = 767 PPFD
So, maybe 25-50% early on (still have to observe and react based on resulting internode length) and ramping up quickly to about 65-70% in mature vege over an 18/6 cycle -- could push 20/4 if you wanted to or just try 7% moe power (1/9th more than 65-70% would be equivalent to adding 2 hours to 18 per day*)
12 hours you want 100%.. i'd keep the same distance -- that is whatever best covers the space... eliminate hotspots and make sure it spreads out without wasting too much to absorption in walls or missing best parts of plant etc.
These suggestion assume you are lighting entire 5x5 area. if you are covering half that, you need to either leave it raised to cover full area and waste watts or proportionally reduce watts, in this case 1/2. all relative.. apples to apples this way.. common sense while covering smaller plants/smaller area requires a change in photons given per day to match.
a cheap lux meter can work for measuring how even your canopy coverage is -- even a free app. it'll be proportional intensity (use a piece of rope o something to measure same distance.. guess you'd need something rigid poking out to sides to help with this type of frame so you can make a proper right angle and measure same distance down). use multiple locations accross canopy, and average it out. or pick some % loss you are willing to accept at corners... maybe 400-500ppfd which should be 50%, or possibly less, than the middle under these lights. so if lux numbers are 50% or better in corners, probably okay... observe and adjust. probably get to same exact point as doing the math.
it's a single bar x2 and you need a wider spread given frame(s). so i'm thinking 20-24" is probably better than 16"-20" they suggest for bloom. basically 4 165w quantum boards spaced out. follow the measurements, though. i can't decide if i'd run them parallel or end to end... maybe 4x6.5 if end to end and if parallel it'd by a tough stretch beyond 4' wide at end of the end of 34.75" length.. 3-4" on each end.. 42" tops, imo. That can be fine too... can leave room for equipment in there, and tighten coverage will increase light per area applied. (increases PPFD->increases DLI, and could take another 10% light per day on 12/12)
It's well-designed for 2 in a 5x5. if you want to boost CO2 and tightly control temps etc etc.. you might want closer to 850watts (erring on high side) of similar quality equipment.