I've been riding GDIT development for a while now and LSPI is quite the thing.
Gasoline in autoignition combusts at a much quicker rate than diesel, so when it detonates at a low RPM (regardless of the reason) it'll do very bad things, always. Gasoline detonation (pre-inition or even at ignition, aka the evil version of 'spark controlled compresion ignition') has always been a potential issue for high-output FI engines but not much summoned or seen by the older designs. It's because of our new engine operating regimes ie. diesel-like torque from gassers, that has brought it to light. In the history of Forced induction gasoline engines, never ever has the boost pressure risen or the torque come on so early in the range than today's implementations, not helped in the least by engine-bogging transmission programming.
LSPI wasn't much of a thing back then because of the low geometric compression ratios, the turbo lag and higher RPM torque curves vs today's GDIT. It was just 'PI' back then. The new GDIT engines have very high cylinder pressures at very low RPM, enabled by small/multi-stage turbos and the charge-cooling of high-pressure direct injection compensating for the elevated compression temperatures by high static compression ratios (ie 12+:1 vs 8:1 of the past). Nevertheless, if that Gasoline AF mixture in the meatiest, low RPM area of the powerband is somehow made to self-ignite, either by the presence of incandescent particles, low-octane species and/or the pressure spike induced by the spark-controlled flame kernel (the very basis for Spark Controlled Compression Ignition) then it's bombs away, we don't have the larger chamber size to buffer the brutal spike of detonating gasoline as we've had in the past. According to the future SCCI combustion regimes, the mode of combustion is similar to some instances of LSPI, but the lean fuel content in the cylinder means much less energy for gasoline detonation, and thus a normally destructive force (fully stoich AF) is scaled way down to be useful and very fuel efficient (controlled, lean "limited energy potential" charge mixture). Cracked ringlands is nothing new, but they used to happen at higher RPMs in the old days, when boost finally peaked, now it's a huge problem because boost peaks early and at low RPMs where the piston dwells in position for longer periods of time, combined with pushing other design limits for efficiency. Whether its pre- or properly timed detonation of gasoline matters little IMO because a fully stoich charge load of gasoline will always do very bad things when detonated/compression ignited especially at low engine speeds.
Also to touch more upon the potential energy, charges are becoming more potent with the new engines as well. With the use of Miller-like valve timing (delayed IVC intended to control charge mass and pumping losses), more of the charge "compression" is happening at the turbine wheel, and not at the compression stroke. This means that since more of the compression work is occuring before the intercooler, it can lose heat at the intercooler, before even reaching the combustion chamber saving the temperature-rise during the compression stroke and controlling compression temperatures. Of course, this means that the charge entering the cylinder is much denser, and will have a compression-stroke temperature peak lower than if standard (non-Miller) IVC timing was used. The phenomenon is then exploited by retaining ignition advance (allowing a larger burn window resulting in a more complete burn) and high-static compression ratio (smaller chamber enhancing burn completion). But should things ever go wrong, like when that nice dense charge stuffed in that small chamber decides to autoignite- the likeliness of major damage is high.
Since naturally aspirated DI engines are atmospherically hard-limited on the maximum density of charge they can intake by the atmosphere, they're not very likely suffer LSPI, and for the same reason are less likely to catastrophically fail by it.
