Bühlmann vs RGBM: How Dive Computer Algorithms Differ

A decompression algorithm is a mathematical model of how inert gas moves into and out of your body, and it is a model rather than a measurement. No dive computer knows anything about your tissues. It knows depth, time and the gas you told it you are breathing, and it runs arithmetic that decades of diving suggest keeps most people out of trouble most of the time. Bühlmann and RGBM are two different pieces of arithmetic built on two different pictures of what is happening inside a diver, and understanding the difference is worth more than picking a winner, because there is no winner.

What is a decompression algorithm actually doing?

Under pressure, the inert gas in your breathing mix dissolves into your blood and tissues in proportion to its partial pressure. On ascent, the ambient pressure falls and that gas has to come back out. If it comes out slowly it leaves through the lungs uneventfully. If the pressure drops faster than the gas can be carried away, it can form bubbles, and bubbles in the wrong place are decompression sickness.

Every algorithm is an attempt to predict how fast that gas goes in and how fast it can safely come out. It does that by modelling your body as a set of theoretical tissue compartments, each with its own half time, which is how quickly it loads and unloads. Fast compartments handle a short deep dive, slow compartments dominate a long day of repetitive diving. The compartments are not organs. They are mathematical fictions that happen to produce useful answers.

Where the two families separate is what they think matters. A dissolved gas model asks how much gas is dissolved and how much supersaturation a compartment can tolerate. A bubble model says that microscopic bubble nuclei exist in the body all the time, and that the real question is whether an ascent lets them grow. Those two starting points produce different advice about the same dive.

How does Bühlmann ZHL-16C work?

Bühlmann ZHL-16C is a dissolved gas model with 16 tissue compartments, each with a published half time and a limit on how much supersaturation it can tolerate at a given ambient pressure. Those limits are the M-values. The model is fully published, which is why it has been implemented by many manufacturers and studied by many people who are not manufacturers, and that openness is a large part of why the technical diving community settled on it.

Raw ZHL-16C is not what your computer runs. Manufacturers add conservatism on top, and on a Shearwater Research Peregrine Dive Computer or a Shearwater Perdix 2 Ti Dive Computer that conservatism is exposed as gradient factors: two percentages you set yourself. The low number governs how close to the M-value line the model lets you get at depth, which sets your first stop. The high number governs how close you may be when you surface.

That table is descriptive rather than prescriptive. What you should set is a question for your instructor and your agency, because it is a decision about your own acceptable risk and it interacts with your age, fitness, hydration, workload and the kind of diving you do. The important point is that on a gradient factor computer the conservatism is a number you chose and can explain, which is a very different relationship with an instrument than accepting whatever it says.

How does RGBM work, and what does the reduced gradient part mean?

RGBM stands for reduced gradient bubble model, developed by Bruce Wienke. It starts from the position that bubble nuclei are always present in tissue and that decompression should be managed to stop them growing, rather than only managing how much gas is dissolved. The reduced gradient part refers to reducing the permitted supersaturation gradient in situations thought to promote bubble growth.

Those situations are the ones you meet in ordinary recreational diving: repetitive dives across a day, reverse profiles where a deeper dive follows a shallower one, short surface intervals, multi day diving, and rapid ascents. An RGBM implementation penalises those explicitly, which is a design intent rather than a side effect.

Implementations differ and they are proprietary, which is the honest limitation of any discussion of them. A Suunto Zoop Novo Wrist Dive Computer runs Suunto RGBM, which is noticeably conservative against a Bühlmann computer at typical gradient factors. A Suunto D5 Dive Computer runs Fused RGBM 2, which is meaningfully less conservative than the older model. A Mares Genius Air Integrated Dive Computer runs the Mares-Wienke implementation, and it is a different piece of software again. Saying a computer runs RGBM tells you the family, not the numbers.

Bubble models are not exclusive to RGBM. VPM-B is another, available as an option on technical units such as the Shearwater Research Teric Dive Computer , and it is used mainly by divers doing planned decompression with training to match. If that is where you are headed, read the technical diving overview first, because model selection is one of the smaller decisions in that world.

How far apart can two computers be on the same dive?

Far enough to cause an argument, which is why this section exists. Two computers running different models on an identical profile can differ by tens of minutes of no stop time, and the gap widens across a day of repetitive diving. A diver on a conservative bubble model and a diver on a moderately configured dissolved gas model can be looking at genuinely different dives while swimming side by side.

Neither computer is broken. Neither diver is being reckless or timid. The spread is the real state of decompression science, where several defensible models fit the available data and none of them can be validated against the inside of your body in real time.

The rule that follows is simple and non negotiable: a buddy pair runs the dive on the more conservative of the two computers. If one shows 18 minutes and the other 31, the dive has 18 minutes in it. Agree that on the surface, where it takes thirty seconds, rather than at depth where the conversation is impossible. The failure mode this prevents is the one that matters: a diver quietly following their own longer numbers past their buddy's limit, and the pair separating during the ascent.

The same applies to surface intervals. If one computer wants longer before the next dive, that is the interval. Nobody wins an argument with the more cautious model, and the cost of accepting it is a coffee. The Suunto and Shearwater comparison covers what this looks like across two specific brands.

Look at the curve rather than the individual numbers. Time collapses fast with depth, from 205 minutes at 35 feet to 20 minutes at 100 feet and 8 minutes at 140. Every model produces a curve of roughly that shape, and they disagree about where exactly it sits. The no decompression limits table carries the same warning and the same caveats.

What matters more than the algorithm?

Almost everything you personally control. Ascent rate matters enormously, and the standard maximum is 30 feet per minute with slower being better in the last 30 feet. A safety stop of three minutes at 15 feet on any dive to 30 feet or deeper is cheap insurance that every model approves of. Hydration, fitness, thermal stress, workload and how well you slept all affect gas elimination in ways no algorithm can see.

So does not diving to the limit. A no stop limit is the edge of a model's comfort, not a target, and divers who habitually surface with time and gas in hand are managing risk more effectively than divers optimising their choice of algorithm. The single most conservative thing you can do is come up earlier.

On the hardware side, other properties affect your diving on every dive rather than in theory. Whether you can read the screen in poor visibility, whether the buttons work with cold hands, whether the battery is user replaceable, and whether the computer is one you will actually wear all matter more day to day than the model inside. A Aqua Lung i300C 2-Gauge Console Dive Computer and a Scubapro Aladin H Matrix Dive Computer will feel more different to dive because of their formats than because of their models.

Why do the models disagree at all?

Because decompression cannot be measured on a living diver in real time, so every model is fitted to outcomes rather than derived from first principles. Researchers know how much gas went in, roughly how it came out, and whether the diver got hurt. What happens in between is inferred. Several sets of arithmetic fit that evidence, and they diverge most exactly where the evidence is thinnest: repetitive diving, long multi day exposures, and profiles that real divers swim but studies rarely reproduce.

That is why the disagreement is respectable rather than embarrassing. A bubble model and a dissolved gas model are two reasonable readings of an incomplete picture, and the difference between them is smaller than the difference between a well hydrated, warm, unhurried diver and a cold, tired, fast ascending one on the same profile.

It is also why chasing the model with the longest numbers is the wrong instinct. Every extra minute a model grants you is a minute closer to the edge of its own assumptions, and models are least reliable at their edges. The safest diver on any boat is usually not the one with the best algorithm.

Which algorithm should you choose?

Choose a Bühlmann computer with gradient factors if you want to set and explain your own conservatism, you are heading toward technical training where gradient factors are part of the vocabulary, or you simply want to understand what the instrument is doing rather than trusting it. Being able to state your settings as two numbers is a real advantage when you are planning with other divers.

Do not choose a gradient factor computer if you would be tempted to loosen the settings to get longer numbers. A configurable model in the hands of a diver who has not been taught what the configuration means is worse than a fixed one, and this is the honest argument against them.

Choose an RGBM computer if you want the computer to handle conservatism for you, you do a lot of repetitive and multi day diving and want a model designed to penalise exactly that, or you dive with a buddy or a club that already runs them and you want your numbers to broadly agree.

Do not choose an RGBM computer if a conservative model would frustrate you into working around it, or if you specifically want to reason about your own decompression rather than accept a setting chosen from a list. Frustration is a safety issue: a diver who resents their computer starts ignoring it.

For most divers this should be one factor among several rather than the deciding one. The dive computer guide works through the rest, and the computer index lists the model each unit runs where the maker publishes it, and omits it where they do not. Whatever you buy, a computer displays a model rather than a plan, and no instrument extends the limits of the certification you hold.

Deciding on a specific model? We review the Shearwater Research Peregrine dive computer review and Suunto Zoop Novo review in full.