Without benefit of a lux meter, just whitewall viewing the unmodded lights' output side-by-side (l'll take photos and post this weekend) I'd declare the HD2010 (single emitter version) the winner by matter of preference. Here's what led me to that conclusion:
Hotspot intensity: advantage no one
The intensity of (the brightest part of) the hotspots of each beam appear to be equal to each other
Hotspot uniformity: advantage no one*
The HD2010 has the most uniform XML in SMO hotspot I've ever seen, with only the very faintest appearance of a brighter edge than center - it's barely noticeable even on high, on a white wall. *The D1 hotspot, now that I've replaced the LED with the separated dome, also produces a hotspot that is uniformly intense throughout.
Hotspot size: advantage HD2010
Despite having a wider mouth than on the D1 (wider mouths often produce smaller hotspots) the reflector of the HD2010 produces a hotspot that is noticeably wider in diameter (and equally as intense, and more uniform). The unconventional variation in the relationship between reflector diameter and hotspot diameter among the two being compared here can be accounted for in the difference in the reflector depth to width ratio (the D1 reflector is deeper).
Spot to spill ratio: advantage HD2010*
If throw is your primary concern, then a higher spot to spill ratio is usually desirable for isolating subjects in the spot and not providing distracting illumination to subjects in the (nearby) spill. If, on the other hand, your desire is for brighter spill and a lower ratio of spot to spill, then the advantage would go to the D1.
Corona to hotspot size ratio: advantage HD2010
This is arguably unimportant to most people. If your preferences are like mine, and you think of the corona as light that didn't make it into the hotspot but should have, then you would prefer the HD2010. The corona of the D1 is both larger and brighter than that of the HD2010. The impression it leaves me with is that a larger LED die would flesh-out the hotspot to completely fill the corona. The first time this host and reflector were offered, it was with an SST-50. I'd be curious to prove my theory by comparing the beam profile of the D1 with its predecessor.
Spill width/angle: HD2010=wider, D1=narrower
(advantage by matter of preference)
Spill brightness: Advantage D1
The impact of the extreme depth of the D1's reflector is that of a narrower, more concentrated spill.
Sustained current to LED on high: Advantage D1
With it's two cell configuration and 3A max buck driver, the D1 will maintain its target current to the LED for more of the discharge cycle than the HD2010. Note that the HD2010 does not regulate current at all. Both lights' drivers use PWM dimming.
Estimated runtime on high: Advantage unknown
With the OEM drivers unmodded, the HD2010 drive current is initially considerably higher than that of the D1 (5A vs. 3A). As the HD2010's single cell depletes and Vbatt drops, drive current will decrease, nearing ~2A, perhaps lower, at the end of the discharge. With it's lower initial drive current, the D1's two cell configuration might appear to initially have an advantage, but as Vbatt decreases, current drawn from the cells will increase. Total cell capacity when comparing 1*26650 to 2*18650 of the same chemistry is comparable. In support of that statement I offer a comparison of BestInOne's 3500mAH IMR26650 vs. their 1500mAH IMR18650, or as in the case of my own cells, eOne Moli 3.3AH 26650 vs AW 1600mAH IMR18650.
Visible PWM modes: advantage HD2010
While still visible to those sensitive and looking for it, the HD2010 PWM frequency is higher than that of the D1, on which the use of PWM dimming is more obvious.
Mode memory (unmodified): HUGE advantage HD2010
The HD2010 will go to the next mode in the sequence at next power-on only if it has only been within a few seconds since last power-off, otherwise it will always start on high. The D1 on the other hand will always start up in the NEXT mode in the sequence unless it has been off for ~20 minutes (in which case it will start in high). (A 22k-33k ohm resistor placed in the location shown in post #70 will mod the D1 OEM driver to operate as the HD2010 driver does).
Ergonomics: HD2010
Balance is leaps and bounds better on the HD2010 than the head-heavy, switch-way-down-at-the-far-end D1.
Uniqueness: HD2010
While there are several logos/"brands" of this host, it is truly unique with it's 66mm bezel and single 26650 cell configuration among a sea of 2*18650 lights with bezel diameter very similar to the D1 (the D1 being a scant few mm wider than conventional WF500 style heads).
Fit and finish: Advantage HD2010
Admittedly a small advantage, considering that both lights are excellent in this category. You would be right to accuse me of being swayed by the UltraFire logo vs.f**king "Prairie Fire" (I suppose I still prefer a black PF to a silver UF for this light)
I can recommend both lights as being unique and superior LED throwers compared to most others, but if you only have the budget or desire to own one (and battery size is of no consequence to you), then my advice would be to go for the HD2010 with a 26650 IMR cell.