In the early days of the internal combustion engine, compression ratios had to be very low because fuel anti-knock quality was very poor -- in the vicinity of 30-40 octane. Oddly, there was a potential benefit in this low ratio. The three-to-one ratio of engines like Ford's Model T left a lot of room above the piston at TDC, room enough for effective large-scale turbulence to persist during compression.
Of course, until Harry Ricardo discovered the value of turbulence in promoting rapid, knock-free combustion, these low compression engines had relatively little turbulence, and they knocked heavily if conditions were favorable.
Even at much higher ratios -- the 5:1 of the "high compression" Chrysler products of the 1930s -- there was still a lot of room above pistons at TDC. In racing and aircraft engines, with hemi combustion chambers, there was circular symmetry as well, allowing perceptive designers like Harry Weslake to put the entire charge into rotation by offsetting the intake port. The value of this lay not so much in simple rotation, as in that it was a way of storing kinetic energy in the charge during compression. As the piston neared the head, viscosity and friction with combustion chamber features would slow the rotation, converting the large-scale motion into small-scale turbulence that was useful in speeding combustion.
After the war, supercharged racing formulas died out. This, combined with the availability of really good gasoline, resulted in use of much higher compression ratios. In the long-stroke engines that had been the preference pre-war, this was still not a terrible thing. Consider a combustion chamber's "aspect ratio", if you will, comparing its height to its diameter. Why should this be important? Well, the hight-to-diameter ratio of a chamber dictates what sorts of motions are or are not likely to persist at TDC. A low aspect ratio (high in relation to diameter) can easily be filled with a single, large-scale motion -- rotation. This can either be around the cylinder axis, as pioneered by Weslake, or tumbling -- revolving around an axis at right angles to the cylinder axis.
As compression ratio and bore/stroke ratio are increased, the ceiling becomes closer to the floor and the walls recede; the aspect ratio increases. Other geometric problems may intrude, too. If we are considering a hemi chamber, high compression ratios are only possible in an oversquare engine by creating a high dome on the piston, which pokes up into the chamber, making aspect ratio even worse. Also, in the effort to raise compression ratio, the designer must push parts even closer together. This means more or less deep valve cut-outs in the piston crown in addition to the intrusive dome.
This makes it imposible to consider tumbling motion because it will be stopped by the valve cut-outs. Axial rotation is a problem for the same reason. Early users of high compression were effectively stopped by this problem.
Kuzmicki at Norton had been a mechanical engineering lecturer in Poland before the war, had been imprisoned by the Soviets. He escaped, and then walked from Siberia to Pakistan. His first job at Norton was as floor-sweeper. He transformed the Manx engine by making it possible to combine high compression and good combustion. He shaped the piston so that it nearly touched the cylinder head at TDC everywhere except in the valve cut-outs and spark-plug region, effectively making the combustion chamber into a compact spectacles form. Within this small volume, the aspect ratio was much lower than it would have been had the chamber embraced the entire piston diameter; there was room for charge motion and, hence, for rapid light-up. The large squish area injected small-scale turbulence into the compact chamber, accelerating its combustion rate.
Chambers of this same form have served many a two-valve engine builder well ever since. It would remain to Keith Duckworth to discover how to achieve equally rapid combustion with almost no squish at all.
Duckworth saw that the folding of the combustion chamber, inherent in using a large valve included angle with over-square bore/stroke dimensions, interfered with combustion. Whatever room there was above the piston at TDC should be in a simple, compact form -- not folded, and not disturbed by machined or cast features in piston or head. Duckworth did away with the piston dome by swinging the valve stems towards each other radically. This reduced the volume of the chamber, making it unnecessary for the piston to push up into it to reach the desired compression ratio.
What remained was a flat, tent-like chamber above the piston. The combustion volume was deepest at the center, near the central spark plug, and tapered away towards the cylinder wall. It wasn't an ideal combustion space because it was cramped at the edges and it had a rather high aspect ratio, but Duckworth combined it with downdraft intake ports that generated a strong tumbling motion in the cylinder during the intake stroke. This motion was able to store kinetic energy during intake and much of compression, decaying to small-scale turbulence at the point of ignition near TDC.
Suzuki bought rights to a machining method for generating chambers that could make good use of Duckworth's tumbling intake energy -- the TSCC scheme. This produced a smooth chamber interior, tallest in the center of the chamber, more lens-shaped than tent-shaped in cross-section. The idea of this and other related designs is to provide unobstructed room in which the tumbling motion can persist during compression, rather than be dissipated. Yamaha, in its 5-valve Genesis designs, has likewise tried to provide a lens-shaped compression space in which intake motions can persist.
The former god of turbulence was squish, but as engine valve area has increased, the territory remaining in which squish zones can be located has shrunk to 15% of bore area and usually less.
All this has been very well so long as street compression ratios are used. At 10 or 11:1, the tumbling motion can find room to persist during compression, with the flow running across the top of the chamber in one direction, and across the piston crown just below it in the opposite direction. The limiting case is a spherical chamber, in which the charge simply rotates. Since a sphere has minimum surface area to volume, this form presents the whirling charge with the minimum possible surface against which friction tends to slow the motion.
As we raise the compression ratio higher, this notional "sphere" gets squashed more and more. Surface area grows in relation to volume, creating more and more friction to slow the charge motion as the piston comes up on compression. What is worse, this squashing of the sphere causes the rapidly moving charge to oppose its own motion; the current across the cylinder head is opposite to the current just below it, moving across the piston crown. Thus not only does charge motion slow because of friction against the piston and chamber, but it must also "rub itself the wrong way" as well.
As the two parts of the flow pass over each other at high speed, the situation is analogous to high wind blowing across a still pond; this is stable only until a tiny inhomogeneity develops in the interface. Once it does, waves form on the interface. It takes time on the sea, but in the combustion chamber the instability is practically instant. The tumbling flow deteriorates into local cells, small enough to rotate "whole".
Now, when the spark occurs, combustion lights up one of these local cells, then the adjacent ones, and so on. This can be OK if there is plenty of kinetic energy left, but is not so OK if charge kinetic energy has dissipated. In that case, the flame moves from cell to cell lazily.
The result is a Ducati 851 chamber, Yamaha OW-01 chamber, or any of the modern tribe that require spark leads out in the 40-degree range.
Standard works on IC engines indicate that efficiency depends upon combustion time, but only very slightly. However, this work invariably depends upon research performed at under 2000 RPM. In racing/sports engines that reach high RPM, good power is repeatedly associated with rapid combustion and short ignition lead, to the extent that ignition lead can almost be regarded as an inverse figure of merit for combustion chamber designs.