Well I’ll certainly do what I can when a new interesting emitter comes out. I agree testing like this is important; so much of my mod planning is based on djozz’s testing. And also all the modders who post their lux measurements helps to inform on what to expect in the real world. 
I guess Easyb's measurements can be anchored some by comparing or normalizing them to "luminous exitance" (total flux in all solid angles divided by emitter area) based on cree data sheets at relatively low power. I think lumens per emitter area is actually more the correct thing than illuminance, which is only measured at single angles, for throw, and is essentially what an integrating sphere is measuring.
To me, it's also pretty important to compare to total power though. A 12V light at 2.5A is using 30W of power. An XP-L can draw what, 24W max? pushing it darn hard? So even if max throw is about the same, the xp-l uses less power and wins. So it still sounds like the xhp35 is "pretty close" to being still not as good as the xp-l in throw per watts. Of course a big part of what throw measures in a generic sense is ability to get light down field with a small reflector, so if you need 30W worth of good throw, well I guess in xp-l you'll need two emitters, which means a bigger light head, so effectively worse throw. If you need 50W worth though :)... It all depends on specific design goals.
As far as the power consumption, it depends on what your priority is; if you just want throw, a dedomed XPE2 torch has a high luminance and lower power consumption, but not a lot of output. Lately I have been interested in how to get a whole lot of lumens down range, and for this purpose I think the XHP35 is the best choice. It has more output than the top binned XPL and apparently has close to the same throw. The XHP50 and XHP70 have more output, but they don’t throw nearly as well. The only problem with the XHP35 is the difficulty with driving it well, especially in single cell lights.
I edited the end of my post. Yeah, for 30W of good throw xhp35 is better, and since nothing else can get the same throw (in the sense of downfield light per lighthead size) in that class, you can't even compare on a per watt basis. It's just the only thing. But for 50W of throw, you need two emitters anyway, and xp-l might be better again. I'm not always impressed with "about the same throw but more output" That means about the same throw, but more night vision blinding from light up close. Again, it depends on the application. Actual ability to see downfield isn't just in the geometrical math. You have to calculate for how the observer works too. This is at least as true, probably much more, for cameras as it is for people (if you're not using a good zoom lens to match the throw), so another application.
I realized I can get a good estimate of the total output (lumens) from my measurements with a bit of math trickery. Because the radiation from the LED is distributed in angle in a known pattern (it is very close to a Lambertian surface), I can derive the total output if I know the luminous intensity directly above the LED. I did this calculation when I was trying to get an absolute value of the output from a ceiling bounce test.
I will outline it here:
Just multiply the luminous intensity (in cd) above the LED by pi to get the total output in lumens.
As a sanity check I will compare the output of the XPL V6 1A using this method with the XPL V6 2C that djozz measured, at 4.0A. LEDs with dome are also very close to Lambertian surfaces so this should be a good estimate. From here, the luminous intensity above the XPL V6 1A with dome at 4A was 522.7cd. Multiplied by pi we get 1642 lumens, which is close to the ~1575 lumens that djozz measured at 4A.
I added the plot with total output for the XHP35 to the OP.
2600 lumens from an XHP35 HI? That’s more than I was expecting. Very nice results. SO I could theoretically get 7000 lumens plus 300+kcd if I mod my SP03.
Nice algebra!
Right PD, the Thrunite TN42 makes 2643 lumens out the front in my lightbox. I have a modified L2 with XHP-35 making 2986 lumens at 3.17A. So, as long as the 3 in your SP-03 are still able to pull a current in the 2.6A range, you should see nice results. And amazing throw as well, probably in the 320-330Kcd range unless I miss my quess.
Nice work with that figuring EasyB, seems to correlate pretty well.
When you said to multiply by pi I figured I was in trouble, ate the last of the pumpkin pie just last night… 
Didn’t know you had one hitting 2900 Dale. I thought 2300 was pushing it.
I’ve got one doing 2.8A in a MaxToch Shooter 2X, another in a MaxToch SN50 doing about 2.83A and the one in the L2 doing 3.13A. These are all using the LD-2 modified to work on 4 cells. The big MaxToch is using 4 18500’s, the Shooter 2X is using 4 18650’s. The L2 of course is using 4 of the 26350’s.
The TN42 is running on 2S2P Samsung 30Q’s. Been meaning to test it and see if it makes the same numbers on a lesser cell…
Edit: the one doing 2900 started out domed and I sliced it. The Shooter 2X is an HI variant and the SN50 is an HI as well.
Did the output fall after you sliced it?
I think i remember reading Dale had the XHP35 HD hitting over 2900 lumen. I think he also sliced the dome to test and lost about 600-700 lumen?
I just tested my L2 in the box, 2349 with the little 26350’s at 4.17V well rested. So yeah, about a 600 lumen’s loss in the end. But over 300Kcd so it’s still very viable.
Thanks for measuring again. I think you misremembered the throw number though. You said it was around 208kcd.
The die luminance is sensitive to the thickness of silicone left after slicing, so maybe with a HI emitter or a closer slice it could get up pretty close to 300kcd in the L2.
He may be referring to my comment of getting 300kcd in my SP03…… or he might be just being Dale and forgetting like you said 
300kcd should be possible in an SP03 with the 3 emitters.
Yeah well, doing good to remember my own name after nights like last night. With so many lights its tough to keep all the particulars separated. I know my Shooter 2X got about 322Kcd, very close to what it was as a de-domed XM-L2. Why on earth would the larger L2 be less throwy? Now I gotta go get a new reading on that too. 
Yep, 206.75Kcd as of 2 minutes ago. The Shooter 2X has a larger diameter and deeper reflector, as well as being powered by 4 Samsung 30Q cells instead of the short 26350’s.
Sorry, my bad.
Are those 26350’s sagging under 2.6amps? I think it must be more about the reflector and focus of the Shooter than it is the cells. Or of course the emitter itself. Are they both HI’s or is one sliced?
L2 with dedomed XPL V6 1A can reach 280Kcd. Based on my measurements, the XHP35 HI could reach ~93% of that. If the dome slice is not very close, that could take away an additional 20% to make ~208Kcd.
I built an L6 with de-domed XP-G2 S2 0D and a Buck driver that makes 439Kcd. The driver is Richards 5.5A 27.5mm. Pretty nice in the L6.
I didn’t try to get the slice super close, like to protect the bond wires a bit ya know.