Graph documentation [was: Re: Meaning of GF settings]

Jan Schubert Jan.Schubert at
Tue Jan 8 12:50:34 PST 2013

On 01/08/13 21:28, Dirk Hohndel wrote:
> Jan Schubert <Jan.Schubert at> writes:
>> On 01/08/13 21:01, Linus Torvalds wrote:
>>> On Tue, Jan 8, 2013 at 11:49 AM, Pierre-Yves Chibon <pingou at> wrote:
>>>> Attached are the two first dives I did with my computer, the first dives
>>>> of the trip as well.
>>>> We did have a deco stop during the second dive (#9) but we 'only' did
>>>> the safety stop 3min/3m during the first dive (#8)
>>> Ok, that doesn't look horrible, and I think it's just a result of the
>>> default parameters set to "somewhat conservative"
>> I agree saying that it does not look that unhealthy. And there might (!)
>> still be some unwanted features in the code as all of my dives using the
>> dive computer generated deco also ends before the ceiling calculated by
>> subsurface.
> The math looks right. And your correct complaint that we don't take the
> surfacing time into account when drawing the ceiling shouldn't make a
> difference here, correct?

No, should not make a difference, at least not that much. Also sample
times should not be the reason as mine are set to several somples per
minute (one each 10 seconds?) which should not make a difference viewing
a 3hrs profile.

> So I'm not sure what else is left. Of course, there could be bugs...

I just tried to get familar with the code and thought we should check
the correctness of calculating tissues_tolerance..

> One thing to do is to look at the dives with GFlow/high 100/100. If you
> still violate the ceiling, I'd start worrying :-/

Thats really a good point Dirk as I just found some anormalities I could
not understand right now: GFlow _should_ not have any impact to existing
profiles as the "deep" part of the dive will not change anyway. But
playing around with GFlow, it clearly has mentionable impact at my
profiles - as of now I would consider to have a look to this.
Unfortunately I'll not be able to check this the next 72hrs as I have to
leave now, Sorry. But maybe this could be a trigger to someone else to

>>>> I used the default parameters btw (GFlow 30 and GFhigh 75).
>>> If those are more conservative than even a suunto, I suspect we should
>>> make them less so for the default "show" to not make people worry.
>>> Then, for dive *planning* we should probably default to something very
>>> conservative indeed. I dunno.
>> I wouldn't do so (now) but just wait what happens to the deco
>> calculations in the next days when we get more knowledge about the
>> missmatchings.
>> Just a note: Suunto is not using Bühlmann and GF settings at all but
>> RGBM which has a little different approach and might (!) allow divers
>> surface earlier than Bühlmann in case it thinks there are no
>> (mentionable) bubbles in your tissues.
> Linus and I have been diving together for a long time. The only time
> EVER when he has shorter deco than I is if we do deep dives and my Uemis
> is set to "extra careful" mode. But in all realistic scenarios it seems
> that the Suunto at default settings shows shorter NDL for him.

Sorry, I'm not saying that RGBM is by default less conservative than
Bühlmann but that RGBM like VPM uses a different approach which will
come to different results. I know about the implemenation done by
Shearwater which uses VPM backed up with a GFhigh of 95% which gets
hitten by some divers, most likely on longer and deeper dives...


More information about the subsurface mailing list