Yup, need a better light... Not even sure if the one in that diary could give you 1g/w, for frame of reference, so less than 50g is probably your ceiling.
It's a fairly complicated thing, actually. Samsung diodes (lm301 B or H) are the best 'white' LED diodes out there as far as efficacy -- you pay for efficacy as any modern LED should have a very narrow range of potential spectrums (spectral flux density? maybe proper term).
With any samsung lm301-based light, simply divided total number of diodes by watts... if near 0.25watts per diode, it's top of the line in efficacy, assuming it also has a decent heat sink and otherwise good build quiality. 0.40-ish and below is good. 0.40-0.50 you get into midrange stuff and expect fewer than 50,000 hours of use and less efficacy.
If you are looking for maybe a half pound yield per grow, a 240w LED of midrange quality around 0 will work well and you don't miss out on too much longevity.
If you want a higher end light, compare to the FC-3000 by mars. Has great features and high efficacy. At 0, i'd spend the extra here, for sure.
Even with the midrange stuff you can push 1-1.33g/watt without sweating too much about the process. You'll go another 10% beyond that with the higher efficacy lights, but it'll also last longer. You will get the full 50,000 hours before it reaches 90% of original intensity. Many lights advertise 50k but if you drive them harder than 0.25w/diode, don't expect to match the diagram at samsun.com related to lifespan curve. it will degrade faster, a mathematical certainty that is very predictable.
But, if oyu want a certain sized garden, figure 1g/w from a midrange light, then figure out how many watts you want working backward from the yield you want. This will be conservatively safe. You don't have to do any extraordinary gardening methods to hit 1g/w with an average light... in fact if you count larf, it'll be higher.