Hi Daniel, I'm not sure I follow what you mean by "plot using the dustfrac array instead of trying to plot particles." Maybe this is a feature of Splash I don't know about? How do you display N dust densities for the disc together without using plot particles? You can't very well do this with rendering, so what else is there? If you can turn on the particles with subtypes, then I think this is an easy solution for plotting multiple dust species on the same plot.
However, I often find that I want to compare a single-phase one-fluid simulation with a multigrain simulation. If I run Splash with a single-phase one-fluid simulation, I only have one dust fraction column. If I run Splash with a multigrain simulation, I have N dust fraction columns (plus a dustfracsum column depending on what version of Phantom was used). The same goes for dust densities and deltav. As far as I can tell, there is no way to plot, e.g., the dust density of these two files on the same plot since Splash generates different numbers of columns for each file so the data won't line up. I was just thinking that perhaps there is a way to align the relevant columns, perhaps by comparing grain sizes or with a command-line flag. The reason I need to compare single-species dumps with multigrain dumps is so that I can isolate the effects of back reaction. As for the Splash issues for the new Phantom dump files, I just checked out the latest version of Phantom and compiled and ran a disc setup with MAXDUSTSMALL greater than 1. I checked out the newest version of Splash and tried to read the file. It seg faults because Phantom dumps the tag "ndustsmall" while Splash looks for the old tag "ndusttypes". If I edit the read_data_sphNG.f90 file to additionally look for ndustsmall, then I can get Splash to get to the menu, but you'll see immediately there are issues. I think it is still trying to look for dustfracsum and deltavsum, which were dumped in earlier version of Phantom. When these columns/tags are not in the dump file, Splash is calculating or assigning the dust columns incorrectly. I don't know whether dustfracsum and deltavsum are being calculated somewhere else in Splash or not, but they are needed to make the fake dust particles. If I try to plot something from the menu despite the menu problems, I again get a seg fault. Cheers, Mark On Fri, Nov 9, 2018 at 12:46 AM Daniel Price <[email protected]> wrote: > Hi Mark (cc to users list), > > > OK, the types thing makes more sense now. That will allow you to display > one or more dust grains on the same figure. That will be really helpful and > way better than what I currently do (dump to ascii and separate the dust > with awk commands). This may be a piper's dream, but it would also be > really nice if we could plot single-grain simulations and multiple-grain > simulations together. I don't know if you guys have run into this problem > yet in your two-fluid stuff, but it either requires combining separate pdfs > or else going through the ascii route again. I wonder if we could have > Splash either generate placeholder arrays filled with zeros for the missing > dust phase arrays in the single-grain case… > > I think this already exists, the method is to always plot using the > dustfrac array instead of trying to plot particles. The dustfrac array > contains both small and large grain information carried by, or interpolated > to, the SPH particles. This is also how we pass the dust information to > MCFOST, i.e. we interpolate all information to a single set of particles > which are used as the generating points for the Voronoi mesh. > > I gather what you’re after is to do the same thing the other way around, > i.e. pretending that the dust species were all handled as large grains? > With subtypes I think this should be possible also, we just need to > generate the fake dust particles as extra sets of dust particles (of > different subtypes). > > > I agree the density problem is definitely a bug, I'm just not sure which > one is correct...both versions are acting funny. I am pretty sure that I > had the densities working correctly at some point, because the densities > matched the independent calculations I did in the Multigrain paper. But > I've tried checking out versions of Splash from a few months ago and the > densities are doing something weird...this is kind of worrying to me. > > Let’s track this via the issue tracker: > https://github.com/danieljprice/splash/issues/12 > > > > > Also, I can't even plot one-fluid/multigrain simulations with the newest > version of Phantom and Splash...it always seg faults. I think this is > partly due to the tags problem I discussed with you earlier. Have you come > up with a fix for this yet? If I go and manually change the tag in Splash, > I can get it to make it to the menu, but some of the columns are missing > (i.e. are blank or are labeled by integer values with no indication of what > they are numbering). > > If you could send me a dump file that causes a seg fault, I will try to > reproduce this from my end… please also file an issue on GitHub > > Cheers, > > Daniel > > > > > > > On Wed, Oct 31, 2018 at 9:47 PM Daniel Price <[email protected]> > wrote: > > Hi Mark, > > > > 1) We made some changes to phantom output to allow for multiple dust > species, which means dust cannot simply be type 2 any more. The types thing > is complicated, what we intend to implement is a “subtypes” feature in > splash, so dust particles are still "type 2", but have subtypes. We haven’t > yet done this, so currently dust particles are types 7->n. > > > > 2) Plotting of dust densities not much should have changed… so this may > be a bug. Is this for one fluid or two fluid? > > > > 3) Would be great to have a Skype chat to discuss where we’re up > to/plans for the hybrid scheme (we’d also like to involve you in the > development). Daniel M has been unwell, but some time in the next week or > so would be good from my perspective… any day that suits you? > > > > Best wishes, > > > > Daniel > > -- You received this message because you are subscribed to the Google Groups "SPLASH users forum" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at https://groups.google.com/group/splash-users. For more options, visit https://groups.google.com/d/optout.
