249 lux @ 8.17m.
Candlepower = 249 x 8.17 x 8.17 = 16620 candlepower.
Slightly lower than UF-980L of nearly 18k plus and U80 of 17k plus.
For us mentally challenged members whats that it lumens? And not DX lumen's, Real ones! lol
For lumens we do it another way....either in an Integrating Sphere or a white walled/white ceiling corner. I do it along the corridor with 2 walls that are 1m apart and 2.5m ceiling. Position your head real close to the floor coz that'd affect the readings. (the light and lux meter are on the floor). Or use a stool and put the meter and light on it thus raised. Do not move too near to the sensor. Then you get a known lumens light like XM-L driven at 2.5A and the emitter lumens for that is about 800 lumens.
Good or bad heatsinking affects the readings of the different lights, but basically most lights are ok if you let it stabilise after the initial 10s and do the reading from 10-15 seconds. The result would be good as a ballpark figure, any number is better than guessing.
Eg, 40mm C8 style, 2.5A is 800 lumens. Say it gives 40 lux. If your measurement light is 80 lux, then that is 1600 lumens. The type of spill and hotspot size makes just a very small difference if your ceiling and walls are white. Use the same position and make sure your hotspot is shining at the middle spot of the ceiling, you will cut down on the variables this way. Note that reflector quality also plays a part, actually this way you are measuring something quite close to OTF (Out the Front).
SR3800 is about 2000 lumens from my readings. Langcjl said about 2300 lumens for his case. I actually estimated 1600-1700 based on calculations, I am wrong.
Thats alot of work just to get a number, But a very important number to people tho. Even tho it doesn't tell the whole story.
I don't like to point the meter at any area the light directly illuminates which most people do (point both at ceiling). I mean, you wouldn't do that in an integrating sphere.
I am waiting for the skyray,both the m3x clone and this 3x xml ,i am waiting too for tk70 and some more lights, i will be able to give pretty accurate numbers of lux and lumen readings.
The most important thing is to keep everything constant. You can even use the corner of a white wall and floor, ie 1/8th pi. Actually you somewhat get the same mechanism in Audio, HiFi etc.... SPL/Sound Pressure Level. Another thing is to not do it too close to a wall/ceiling, that causes different spot sizes. That's all I can think of at the moment.
The thing is, I can use the same flashlight and I can point at the wall 20 deg slanted to the left, 20 deg slanted to the right, 0 deg straight up, and I get very close readings. The sensor takes in light from about a 120 deg arc.
You don't need very close readings anyway, coz even things like batteries and to a much bigger degree the heating of the emitter, would at the different measurement times return different results. 2-5% is not uncommon for the good lights. But we usually let it stabilise just after the emitter and immediate driver board/heatsink heats, up but before the host heat saturates. And then we have things like batch variations etc.... So we'll contend with what we can do....
Now with P60 lights that is very difficult to do, because the thing just dives and drives....till the thing heat saturates. I have a feeling that if you leave it in high in my country, the thing will get killed after 10 minutes without it ever stabilising, should you forget to wrap the drop-in. LOL! 
In short, we are talking about getting a relative figure.
It is good to have more people to chime in and build up a DB. For my 3xT6 it is 2000 lumens, Langcjl is 2300L. We can get a pretty good range and paint a good picture.
One thing I thought about doing but kind of lazy to is a simple design w/ standard milk cartons that everyone can make and get normalized readings. Basically just away the spouts and bottoms of two of these, tape them together and shine light at one spout and measure at the other:
That way everyone can be more or less talking about the same thing when speaking of lumens.