11/02/05 These runs were created before today, but I've finally gotten around to catalogueing (sp?) them. For actual comments on the data, go down further. ==== data.raw/XOa ===================================================== NX = 250 XMAX = 2.5 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 0.948683 NSTEP0 = 150 NZ = 250 ZMAX = 2.5 NSTEP1 = 50 ZMAX2 = 0.5 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 ==== data.raw/fine ==================================================== NX = 250 XMAX = 25. NTOTAL = 2100 -> ntim = 211 NY = 1 YMAX = 0.948683 NSTEP0 = 1050 NZ = 250 ZMAX = 2.5 NSTEP1 = 10 ZMAX2 = 0.5 NSTEP2 = 10 NCPU = 5 NSTEP3 = 10 ==== data.raw/finez2 ================================================== NX = 250 XMAX = 2.5 NTOTAL = 2100 -> ntim = 211 NY = 1 YMAX = 0.948683 NSTEP0 = 1050 NZ = 250 ZMAX = 25. NSTEP1 = 10 ZMAX2 = 0.5 NSTEP2 = 10 NCPU = 5 NSTEP3 = 10 ==== data.raw/resx2 =================================================== NX = 50 XMAX = 2.5 NTOTAL = 2100 -> ntim = 211 NY = 1 YMAX = 0.948683 NSTEP0 = 1050 NZ = 250 ZMAX = 25. NSTEP1 = 10 ZMAX2 = 0.5 NSTEP2 = 10 NCPU = 5 NSTEP3 = 10 ==== data.raw/resz2 =================================================== NX = 250 XMAX = 2.5 NTOTAL = 2100 NY = 1 YMAX = 0.948683 NSTEP0 = 1050 NZ = 50 ZMAX = 25. NSTEP1 = 10 ZMAX2 = 0.5 NSTEP2 = 10 NCPU = 5 NSTEP3 = 10 ------------------------------ Comments ------------------------------- 1. There are 3 runs currently going, resx1, resz1, and y1. y1 will test whether or not y=0.948... vs. y=1 has a big difference on these 2-d runs. The hope is that it doesn't, and I can just set y=1 in the future. y1 should be compared with XOa. 2. XOa and fine are basically the same; fine is simply compressed in the x-direction, which makes sense since XMAX have been increased by a factor of 10. What DOESN'T make sense is that the XOa/ run has 3 periods across a z=constant slice while the fine/ run has 25. I'd have expected 30 = 3 x 10. 3. The runs finez2, resx2, and resz2 are qualitatively different from fine/ and XOa/ due to the difference in the ratio of ZMAX and ZMAX2. 4. Nomenclature: resx1/resz1 will have NX/NZ be larger than NZ/NX. This only makes sense - more points = higher resolution. For some reason the runs resx2/resz2 have NZ/NX larger than NX/NZ. I could rename them, but I hesitate to do so for reasons of reproducibility. So, just realize that res?2/ are oddly named. 5. All the data is on the mass store; I'll have on iditarod what I can, but the disk has been filling up due to the work I've been doing on both the triple stuff and the RAMS stuff. ======================================================================= 11/04/05 These comments come from the beta version of the readme file: ****************************************************************************** 1: XOa: The default run. Looks ok. 2: fine: changes the values of NTOTAL, NSTEP#, and XMAX. There is a mapping from fine's timesteps to XOa's. 1 -> 1 6 -> 2 11 -> 3, etc... Also, XMAX is 10x's as big in fine as in XOa, which is made apparent by the dense idl plots. 3: finez: After seeing fine, and with a desire to affect a vertically sensitive initial velocity profile, I decided to make a run w/ ZMAX = 10 * XMAX, with XMAX as it is in XOa ****************************************************************************** ==== data.raw/resx1 =================================================== NX = 250 XMAX = 2.5 NTOTAL = 2100 -> ntim = 211 NY = 1 YMAX = 0.948683 NSTEP0 = 1050 NZ = 50 ZMAX = 2.5 NSTEP1 = 10 ZMAX2 = 0.5 NSTEP2 = 10 NCPU = 5 NSTEP3 = 10 ==== data.raw/resz1 =================================================== NX = 50 XMAX = 2.5 NTOTAL = 2100 -> ntim = 211 NY = 1 YMAX = 0.948683 NSTEP0 = 1050 NZ = 250 ZMAX = 2.5 NSTEP1 = 10 ZMAX2 = 0.5 NSTEP2 = 10 NCPU = 5 NSTEP3 = 10 ==== data.raw/y1 ====================================================== NX = 250 XMAX = 2.5 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.0 NSTEP0 = 1050 NZ = 250 ZMAX = 2.5 NSTEP1 = 50 ZMAX2 = 0.5 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 ------------------------------ Comments ------------------------------- 1. The y1/ data and the XOa/ data are identical; that is to say when I do iditarod>diff XOa/XOa.?xz.#### y1/y1.?xz.#### (for a given ? and ####) I get that there is no difference. Didn't really bother plotting. The upshot of this is that I can set y = 1. 2. There is a noticable difference between resx1 and resz1. The interior ovals are flatter in the resz1 case (flatter on top and bottom). Makes sense I guess? Yeah, if fewer x points, then the points that define the oval top are spread out, so _ _/ \_ becomes _____. Yeah, you got it. Looking at the profiles, but I don't think that'll tell me anything else. ======================================================================= 11/16/05 ==== data.raw/zmax0 =================================================== NX = 250 XMAX = 2.5 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.0 NSTEP0 = 1050 NZ = 250 ZMAX = 0.0 NSTEP1 = 50 ZMAX2 = 2.5 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 ------------------------------ Comments ------------------------------- 1. The init.format file for the zmax0 needs to reflect the fact that only 3 variables, 2, u, and w, were outputted and that no *.txz.tar file exists. ======================================================================= ==== data.raw/zmaxno ================================================== NX = 250 XMAX = 2.5 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.0 NSTEP0 = 1050 NZ = 250 ZMAX = 0.0 NSTEP1 = 50 ZMAX2 = 2.5 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 ------------------------------ Comments ------------------------------- 1. I think this was run after zmax0. Not sure but it doesn't really matter. The re-run was done to see if the failure to obtain temperature output was a fluke. It doesn't look to be and may arise from the fact that the temperature profile needs a non-zero ZMAX. 2. Since this data is the same as zmax0, you can safely delete it. ======================================================================= ==== data.raw/zmax2.5 ================================================= NX = 250 XMAX = 2.5 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.0 NSTEP0 = 1050 NZ = 250 ZMAX = 2.5 NSTEP1 = 50 ZMAX2 = 0.0 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 ------------------------------ Comments ------------------------------- 1. This run is the compliment of zmax0 and was done to see what happens to iprob 12 when zmax span the entire vertical domain and zmax2 = 0 ======================================================================= ==== data.raw/zmaxall ================================================= NX = 250 XMAX = 2.5 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.0 NSTEP0 = 1050 NZ = 250 ZMAX = 2.5 NSTEP1 = 50 ZMAX2 = 0.0 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 ------------------------------ Comments ------------------------------- 1. Hmm. This is identical to zmax2.5. Go ahead and eliminate it. 2. Comparing XOa and zmax2.5, they look fairly similar. Both have ZMAX = 2.5, but XOa has a non-zero value for ZMAX2 ( = 0.5) 3. In looking at the profile along z = const, you can see that XOa has a smaller amplitude than does zmax2.5, though the wavelength and phase are the same. This is only looking at the 1st timestep. 4. By t = 40, both are still in phase and Amp|2.5 > Amp|XOa, but the centerlines for the two have shifted. This is most likely due to different vertical domain sizes. 5. I'm reminded again in looking at the data why I went to more discrete timesteps; it's kinda hard to get a sense of the motion of the wave from this data since the jumps are kinda big. 6. HEHEHE!! Setting ZMAX = 0: a. Results in not *.txz.* files and b. Results in entries of NaN for EVERY LINE of the 2, u, w files. So THAT is an invalid option. ======================================================================= ==== data.raw/even ==================================================== ------------------------------ Comments ------------------------------- 1. This run looks at what happens when ZMAX = ZMAX2. Non-dim. values of x and z were expanded to 5 since 5 divides more nicely than 2.5 2. For some reason this failed to run the first time. ======================================================================= ==== data.raw/all ===================================================== ------------------------------ Comments ------------------------------- 1. This is the same as the even/ run, but w/ ZMAX2 = 0. You should compare it to zmax2.5 for kicks. 2. For some reason this failed to run the first time. ======================================================================= ==== data.raw/pack0 =================================================== NX = 250 XMAX = 5.0 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.0 NSTEP0 = 150 NZ = 250 ZMAX = 5.0 NSTEP1 = 50 ZMAX2 = 0.0 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 ------------------------------ Comments ------------------------------- 1. This is the first run done with the intent of altering the wave packet, not the background temp. profile. I think I need to control both. Um, right now I'm thinking that if I can specify the packet in the code, this will be my initialization and I can simply watch the model evolve. 2. I haven't actually made a change for this run; just testing to see that I'm starting w/ what I think I am (i.e. the triple code) These results should be the same as the all/ run above. 3. While even, all, and pack1 failed to run, pack0 did ok. Wonder why. Also, trying those other 3 again. 4. I'll put this here rather then create a whole other block for it. I accidentally deleted the contents of pack0/ on kraken. This isn't SO bad since they'd be deleted anyway, but I wanted to check something. For some reason pack1, etc... won't run due to the executable script, e250., not being created. Since pack0 ran, packa should run, where packa's only change from the codein/codeit in pack0 is that pack0 -> packa. If this doesn't run I may have to ask Joe what's up. 5. I compared the pack0/ data to the zmax2.5 data. pack0 zmax2.5 number of wavelenghts: 5 3 amplitude range of t : -.11 to .11 -.02 to .02 XMAX, ZMAX : 5.0 2.5 6. I believe I know what the problem is with my runs. I tried forcing the kz's to zero. For some reason that caused the executable to not be created. I'm not satisfied with that answer though. That doesn't explain why the all/ and even/ runs didn't execute; those still utilize the _zyx code and paths. Could be due to use of 'initbl' instead of initks. Yeah, that '0' explanation really doesn't sound right when you consider only triple_pack2.f made a call to initbl. Gonna run packb exactly as packa, and gonna run packc AFTER I change the k_z's to zero in subtrp_bl.f 7. The reason for the failure of the all/ and even/ runs is most likely due to the fact that the codein referenced the zyx code but the codeit referenced the _bl code. Gonna try copying the original codeit to the even/ and all/ directories and seeing if that executes. 8. all/ was rerun w/ the data on jules untouched. even/ was run after deleting those files. 9. packa and packb both ran successfully. 10. packc used a subtrp_bl.f where all the kz's are set to zero. Yeah, it, um, seemed to run well too. No .tar files yet but it just started running and there are .tgz files. ======================================================================= ==== data.raw/pack1 =================================================== ------------------------------ Comments ------------------------------- 1. pack1 should be like pack0. What I did is: a. in codeit, changed $V{MAIN}="triple_bl"; to "triple_pack1" b. in codein, changed CTDtriple_bl.f to CTDtriple_pack1.f c. created CTDtriple_pack1.f from CTDtriple_bl.f; changed ONLY the first line from triple_bl.f to triple_pack1.f d. created triple_pack1.f from triple_bl.f The idea behind this is that as I change the code, I want to keep my changes, and so I think making a file for each change and compiling that is best. This way too things will be easily reproducible. (sp?) 2. For some reason this failed to run the first time. ======================================================================= ==== data.raw/pack2 =================================================== ------------------------------ Comments ------------------------------- 1. pack2 uses initbl instead of initks. :) I set akz's to zero. We'll see what happens. All I want to start is a horizontally propagating packet. ======================================================================= 11/21/05 ==== data.raw/thick0 ================================================== NX = 250 XMAX = 2.5 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.0 NSTEP0 = 150 NZ = 250 ZMAX = 2.5 NSTEP1 = 50 ZMAX2 = 0.5 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 WAVESTDEV = .1 ------------------------------ Comments ------------------------------- 1. The thick0 run was done with WAVESTDEV set to .1 from the 'standard' value of 1. The idea is that i need to control the input wave packet and the velocity profile and this is a stab at that. 2. You should compare this to the XOa run, since the only differences between the two are the WAVESTDEV parameter and the YMAX parameter, which I don't think matters. ======================================================================= 1/24/06 RRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRR ------------------------------ Comments ------------------------------- Since it's given you a little trouble in the past, I thought I'd state here, clearly, that: nt = 1 + ntotal/nstep2 Be Aware: 1. The +1, while currently true, may not always be so. 2. Currently, nstep1 = nstep2 = nstep3; if those are different, I'm not sure how the above would change. 3. If something isn't working, you should always be able to add up the total amount of data you have then start dividing stuff out, like the spatial dimensions and the word size, to end up w/ the # of timesteps. Cool? ==== data.raw/r4 ====================================================== NX = 250 XMAX = 2.5 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.0 NSTEP0 = 150 NZ = 250 ZMAX = 2.5 NSTEP1 = 50 ZMAX2 = 0.5 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 WAVESTDEV = 1. IBX = IBY = 0; periodic IBZ = 1; bounded ------------------------------ Comments ------------------------------- 1. First of all, having the box bound in the z - direction was a a mistake; should really set that to zero. 2. This run was basically a first go of iprob 12 to make sure I'd installed the code correctly. 3. Actually, this was run w/ _bl not _zyx and it worked, so yay. ======================================================================= ==== data.raw/r5 ====================================================== NX = 250 XMAX = 2.5 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.0 NSTEP0 = 150 NZ = 250 ZMAX = 2.5 NSTEP1 = 50 ZMAX2 = 0.5 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 WAVESTDEV = 1. IBX = IBY = 0; periodic IBZ = 1; bounded ------------------------------ Comments ------------------------------- 1. the r5/ run ran using dtriple_bl2/ instead of dtriple_bl, though no changes were made. If you do a diff between the 2 sets of un- tarred files, you'll see they're the same. ======================================================================= ==== data.raw/r6 ====================================================== NX = 250 XMAX = 2.5 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.0 NSTEP0 = 150 NZ = 250 ZMAX = 2.5 NSTEP1 = 50 ZMAX2 = 0.5 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 WAVESTDEV = 1. IBX = IBY = 0; periodic IBZ = 1; bounded ------------------------------ Comments ------------------------------- 1. again, need to change that damned IBZ. 2. This is the same as r4, r5 except that I deleted the call to the 'wavepacket' subroutine in dtriple_bl3/. Whee. 3. Yeah, every output file contains 0.00E+00 for every data entry. This means (I'm pretty sure) that the output is only for the wave/ perturbation data and says nothing about the back ground, which confirms what I thought I saw with the other runs. 4. -> You may want to look at a KH problem; I thought THAT output isotherms of the background. Dunno, could be something else. 5. Yeah, this got partially overwritten by the stupid r7/ run. But, due to the timestamp on the tar files and the *.mean* files and the data in the *.tar (NOT *.tgz) files, I think things are alright. Just, be aware again that when r6/ was run the first time the executable had the wavepacket routine removed. ======================================================================= 1/25/05 ==== data.raw/r7 ====================================================== NX = 250 XMAX = 2.5 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.0 NSTEP0 = 150 NZ = 250 ZMAX = 2.5 NSTEP1 = 50 ZMAX2 = 0.5 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 WAVESTDEV = 1. IBX = IBY = IBZ = 0; periodic ------------------------------ Comments ------------------------------- 1. A kh run, just to see what the temp looks like again; I really would like to be able to look at the background instead of just the wave packet. 2. BUT, I messed up, and while I edited the codein.re250 file, I did not then copy it to codein, so I don't really know what ran; something stupid no doubt. I wish I knew how to target and quit my job, but I've started running r8 as well and would hate to get the wrong one. Still, I should learn how to abort a job. 3. Huh. r7 didn't even seem to execute. Probably b/c I used the codein for r6; I wonder if it overwrote my data. Dammit, I bet it did. It did, though it left the data alone. ======================================================================= ==== data.raw/r8 ====================================================== Basically, r4/ with IBZ = 0 NX = 250 XMAX = 2.5 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.0 NSTEP0 = 150 NZ = 250 ZMAX = 2.5 NSTEP1 = 50 ZMAX2 = 0.5 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 WAVESTDEV = 1. IBX = IBY = IBZ = 0; periodic ------------------------------ Comments ------------------------------- 1. Yeah, just to see if I recover XOa 2. I did. ======================================================================= ==== data.raw/r9 ====================================================== NX = 32 XMAX = 12.5 NTOTAL = 180 -> ntim = 19 NY = 20 YMAX = 4.2 NSTEP0 = 180 NZ = 64 ZMAX = 25.0 NSTEP1 = 10 ZMAX2 = 0.0 NSTEP2 = 10 NCPU = 4 NSTEP3 = 10 WAVESTDEV = 1. IBX = IBY = 0; periodic IBZ = 1; bounded ------------------------------ Comments ------------------------------- 1. A kh run, just to see what the temp looks like again; I really would like to be able to look at the background instead of just the wave packet. 2. Also, this is an excellent opportunity to see that 19 = 1 + 180/10 3. Then ntim calc is fine. The data looks stupid in idl though; I think I'm gonna try running again but with NY = 1 ======================================================================= ==== data.raw/r10 ===================================================== NX = 250 XMAX = 2.5 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.0 NSTEP0 = 150 NZ = 250 ZMAX = 2.5 NSTEP1 = 50 ZMAX2 = 0.5 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 WAVESTDEV = 1. IBX = IBY = IBZ = 0; periodic ------------------------------ Comments ------------------------------- 1. A run to start probing the background temperature profile. 2. Oh, look for the phrase "Brians Addition" somewhere in the output, maybe the dat0 file? Nope, dat0 file has other stuff (which you can read about). "Brians Addition" appears 5 times in a row in the r5.out file. 3. Did find this out; Nstep0.ge.Nstep1, and Nstep2.ge.Nstep3, by necessity. Not sure 'bout the relationship between 1 and 2. 4. SO, what r10 did is take the _bl2.f code and, in the wavepacket subroutine, change the line T(l) = T(l) - tbar(k,1) -- to -- T(l) = tbar(k,1) It would be nice if this let me just look at the background profile. 5. It failed to execute; if it fails again I'll try w/ the _bl code. It shouldn't matter though since the r5/ run ran and IT used the _bl2 code. 6. Part of the problem may be that for the comment character I used ';' instead of '!'. Stupid IDL and fortran similarities and differences. ======================================================================= ==== data.raw/r11 ===================================================== NX = 250 XMAX = 2.5 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.0 NSTEP0 = 150 NZ = 250 ZMAX = 2.5 NSTEP1 = 50 ZMAX2 = 0.5 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 WAVESTDEV = 1. IBX = IBY = IBZ = 0; periodic ------------------------------ Comments ------------------------------- 1. This is like run10/, but with the bl2 code further modified; I don't like the small values of tbar. ======================================================================= ==== data.raw/r12 ===================================================== NX = 32 XMAX = 12.5 NTOTAL = 180 -> ntim = 19 NY = 1 YMAX = 4.2 NSTEP0 = 180 NZ = 64 ZMAX = 25.0 NSTEP1 = 10 ZMAX2 = 0.0 NSTEP2 = 10 NCPU = 4 NSTEP3 = 10 WAVESTDEV = 1. IBX = IBY = 0; periodic IBZ = 1; bounded ------------------------------ Comments ------------------------------- 1. An attempt to nicely reproduce the KH data. ======================================================================= 1/30/06 Both r9 and r12 are the same, I think. The reason it looked like I'd failed to reproduce the original KH data is that I changed the way I wrote the formatted data; before I had one column of data and now I have 3 (nx, nz, data), and the idl code I was using failed to take this into account. r11 is a success in the sense that I was able to multiply tbar. I looked at a profile of the data though, and it looks a little sinuous for my taste. Time to muck around in the xopause routine. ==== data.raw/r13 ===================================================== NX = 250 XMAX = 2.5 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.0 NSTEP0 = 150 NZ = 250 ZMAX = 2.5 NSTEP1 = 50 ZMAX2 = 0.5 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 WAVESTDEV = 1. IBX = IBY = IBZ = 0; periodic ------------------------------ Comments ------------------------------- 1. This run will make use of the _bl code. I'm also changing the _bl2 code back to the _zyx code since the changes I had made were minimal and I'd prefer to save it for a more meaningful change. 2. In _bl, in xopause routine, I just set tbar(k,1) = 5.0. I didn't touch tbar(k,2) which may cause trouble. 3. I'm pretty sure that tbar(k,1) = Temp. Profile. and tbar(k,2) = dtdz of tbar(k,1) because of the code notes and the fact that it just kinda' appears that way ======================================================================= ==== data.raw/r14 ===================================================== NX = 250 XMAX = 2.5 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.0 NSTEP0 = 150 NZ = 250 ZMAX = 2.5 NSTEP1 = 50 ZMAX2 = 0.5 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 WAVESTDEV = 1. IBX = IBY = IBZ = 0; periodic ------------------------------ Comments ------------------------------- 1. editing the _bl2 code: in xopause, tbar(k,1) = 5.0 * z tbar(k,2) = z ======================================================================= ==== data.raw/r15 ===================================================== NX = 250 XMAX = 2.5 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.0 NSTEP0 = 150 NZ = 250 ZMAX = 2.5 NSTEP1 = 50 ZMAX2 = 0.5 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 WAVESTDEV = 1. IBX = IBY = IBZ = 0; periodic ------------------------------ Comments ------------------------------- 1. editing the _bl3 code: in xopause, tbar(k,1) = 5.0 tbar(k,2) = 0. ======================================================================= 2/2/06 I wish I had more to say about r13-r15, but I don't. ==== data.raw/r16 ===================================================== NX = 250 XMAX = 2.5 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.0 NSTEP0 = 150 NZ = 250 ZMAX = 2.5 NSTEP1 = 50 ZMAX2 = 0.5 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 WAVESTDEV = 1. IBX = IBY = IBZ = 0; periodic IPROB = 30 ------------------------------ Comments ------------------------------- 1. Adding an IPROB number to the code for my experiment. Modified the _bl code, first by restoring it to _zyx. ======================================================================= ==== data.raw/r17 ====================== This should be default: XOa == NX = 250 XMAX = 2.5 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.0 NSTEP0 = 150 NZ = 250 ZMAX = 2.5 NSTEP1 = 50 ZMAX2 = 0.5 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 WAVESTDEV = 1. IBX = IBY = IBZ = 0; periodic IPROB = 30 ------------------------------ Comments ------------------------------- 1. r16/ had a problem w/ extra ')'s in the code. r17 fixed it. 2. SWEET; r17 matches XOa, so I have a good starting point to making my own IPROB. I'm gonna go ahead and copy the _bl code I have to _bl1, _bl2, and _bl3. 3. The ONLY difference between _bl and _zyx is that _bl has iprob 30 everywhere there is an iprob 12 ======================================================================= 2/3/06 So, I'd like to set the value of freq = 1 since I want just a horizontally propagating wave and: T(l) = amp*cos(wkxxi+wkz*z(k))*emask (k)*(exp(-diff*ww2i) +aa*z(k)**2 +bb*z(k) +cc ) where wkz=2.*pi*sqrt(1-freq*freq) so if freq = 1, wkz = 0 and there is no z dependence in the cosine term. aa, bb, and cc are all just constants. I need to see where freq is set, since it's not OBVIOUS that its in the codein :( Okay, I checked into tsetup.f: _______________________________ Codein |tsetup/triple ----------------+-------------- FREQOVERN |freq WAVESTDEV |wide WAVESTDEV2 |wide2 ----------------+-------------- Turns out I've been running with freq = 1 as is. Hm. ==== data.raw/r18 ===================================================== NX = 250 XMAX = 2.5 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.0 NSTEP0 = 150 NZ = 250 ZMAX = 2.5 NSTEP1 = 50 ZMAX2 = 0.5 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 WAVESTDEV = .1 IBX = IBY = IBZ = 0; periodic IPROB = 30 ------------------------------ Comments ------------------------------- 1. I'm running this w/ WAVESTDEV = .1 to see what happens. I know I did this w/ thick0, but shutup. Also, using _bl since there were no code changes, only codein changes. ======================================================================= ==== data.raw/r19 ===================================================== NX = 250 XMAX = 2.5 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.0 NSTEP0 = 150 NZ = 250 ZMAX = 2.5 NSTEP1 = 50 ZMAX2 = 0.5 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 WAVESTDEV = .1 IBX = IBY = IBZ = 0; periodic WAVESTDEV2= .1 IPROB = 30 ------------------------------ Comments ------------------------------- 1. I edited the _bl1 code to try and constrain the wave in the horizontal. Also, since I'm hacking, I need to hardwire wide2 in the code since currently, wide2=0 for iprob.ne.13, and 30.ne.13 2. wide1 is the sig. for both x and z, as I have modified it. ======================================================================= ==== data.raw/r20 ===================================================== NX = 250 XMAX = 2.5 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.0 NSTEP0 = 150 NZ = 250 ZMAX = 2.5 NSTEP1 = 50 ZMAX2 = 0.5 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 WAVESTDEV = 1 IBX = IBY = IBZ = 0; periodic IPROB = 30 ------------------------------ Comments ------------------------------- 1. I edited the _bl2 code to elminate emask(k). We'll see what happens. ======================================================================= ==== data.raw/r21 ===================================================== NX = 250 XMAX = 2.5 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.0 NSTEP0 = 150 NZ = 250 ZMAX = 2.5 NSTEP1 = 50 ZMAX2 = 0.5 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 WAVESTDEV = 1 IBX = IBY = IBZ = 0; periodic IPROB = 30 TZRATIO = 1; usually = 2.828. ------------------------------ Comments ------------------------------- 1. Running _bl. The idea is that tzr = 1, and so there is no kink in the T-profile. Since FREQOVERN = 1, I'm assuming horizontal motion only. But in some of the other runs, I see what looks like vertical motion. So, I got rid of the vertical variation with the hopes that only horizontal motions result. We'll see ======================================================================= ==== data.raw/r22 ===================================================== NX = 250 XMAX = 2.5 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.0 NSTEP0 = 150 NZ = 250 ZMAX = 2.5 NSTEP1 = 50 ZMAX2 = 0.5 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 WAVESTDEV = .1 IBX = IBY = IBZ = 0; periodic IPROB = 30 TZRATIO = 1; usually = 2.828. ------------------------------ Comments ------------------------------- 1. I like the clarity of the WAVESTDEV = .1 plot, and so I thought I'd look at TZRATIO = 1 with that. ======================================================================= ==== data.raw/r23 ===================================================== NX = 250 XMAX = 2.5 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.0 NSTEP0 = 150 NZ = 250 ZMAX = 2.5 NSTEP1 = 50 ZMAX2 = 0.5 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 WAVESTDEV = 1 IBX = IBY = IBZ = 0; periodic IPROB = 30 TZRATIO = 2.828. ------------------------------ Comments ------------------------------- 1. Edited _bl3 to have aa = bb = cc = dd = 0. The idea is to eliminate the polynomial and see how badly behaved the solution is at the bndy ======================================================================= ==== data.raw/r24 ===================================================== NX = 250 XMAX = 2.5 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.0 NSTEP0 = 150 NZ = 250 ZMAX = 2.5 NSTEP1 = 50 ZMAX2 = 0.5 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 WAVESTDEV = 1. IBX = IBY = IBZ = 0; periodic WAVESTDEV2= 1. IPROB = 30 ------------------------------ Comments ------------------------------- 1. Companion run to r19. I increased the value of wide in the codein to give me a closer comparison to XOa ======================================================================= 2/6/06 Pitt beat the Seahawks - Go Bus! RESULTS: r17 - This is the same as XOa and is the base/default run. It appears as though the waves propagate down and to the left (-z, -x), though if the timestep is funky you realize this could be misleading. r18 - This run is cool; used WAVESTDEV = .1 instead of 1. There seems to arise a symmetry around the middle of the domain (vertical middle) r19 - funky. The data doesn't make a lot of sense. It's all very jagged and discontinuous. Look at 24 before making any decisions. r20 - eliminated emask. the first time step is different from r17, but not by much, and the run seems to evolve in the same way r21 - tzr = 1; attempt to eliminate horizontally travelling waves. The first time step looks the same as that of r20, which is odd since I don't think I changed the emask there. - The results definitely seem out of step with the r17 data, though it still looks like there is vertical motion. r22 - definitely contains propagation down and to the left. r23 - eliminating the polynomial REALLY messes with the data/model. Don't go there. r24 - Yeah... identical to r19. i.e. bad. SUMMARY So what have I learned? 1. The waves do indeed appear to propagate (-z,-x) 2. DON'T SET aa = bb = cc = dd = 0; the continutity of the derivative at the boundaries IS important 3. It's hard to wrap this thing in X like Z How about an iprob=13 run or two? ==== data.raw/r25 ===================================================== NX = 250 XMAX = 2.5 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.0 NSTEP0 = 150 NZ = 250 ZMAX = 2.5 NSTEP1 = 50 ZMAX2 = 0.5 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 WAVESTDEV = .1 IBX = IBY = IBZ = 0; periodic WAVESTDEV2= 1. IPROB = 13 ------------------------------ Comments ------------------------------- 1. Run using the _bl code. Wanted to see how this compares to r19 ======================================================================= ==== data.raw/r26 ===================================================== NX = 250 XMAX = 2.5 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.0 NSTEP0 = 150 NZ = 250 ZMAX = 2.5 NSTEP1 = 50 ZMAX2 = 0.5 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 WAVESTDEV = .1 IBX = IBY = IBZ = 0; periodic WAVESTDEV2= 1. IPROB = 13 ------------------------------ Comments ------------------------------- 1. r25 w/ a wider WAVESTDEV ======================================================================= 2/7/06 ==== data.raw/r27 ===================================================== NX = 250 XMAX = 2.5 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.0 NSTEP0 = 150 NZ = 250 ZMAX = 3.0 NSTEP1 = 50 ZMAX2 = 0.0 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 WAVESTDEV = 1 IBX = IBY = IBZ = 0; periodic IPROB = 30 TZRATIO = 1.0000 ------------------------------ Comments ------------------------------- 1. running w/ zmax2 = 0 in another attempt at homogeneity ======================================================================= ==== data.raw/r28 ===================================================== NX = 250 XMAX = 2.5 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.0 NSTEP0 = 150 NZ = 250 ZMAX = 3.0 NSTEP1 = 50 ZMAX2 = 0.0 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 WAVESTDEV = 1 IBX = IBY = IBZ = 0; periodic IPROB = 30 TZRATIO = 2.828 ------------------------------ Comments ------------------------------- 1. This run changed the _bl2 code. Emask was restored to _bl and z0 was changed from z0 = zmx1-3.*wide to z0 = zmx1-1.*wide ======================================================================= ==== data.raw/r29 ===================================================== NX = 250 XMAX = 2.5 NTOTAL = 51 -> ntim = 52 NY = 1 YMAX = 1.0 NSTEP0 = 51 NZ = 250 ZMAX = 2.5 NSTEP1 = 1 ZMAX2 = 0.5 NSTEP2 = 1 NCPU = 5 NSTEP3 = 1 WAVESTDEV = 1. IBX = IBY = IBZ = 0; periodic IPROB = 30 ------------------------------ Comments ------------------------------- 1. _bl code, but with a high temporal resolution so as to see the nature of the propagation. 'n stuff. ======================================================================= 2/8/06 Observations: r25 - Same as r26. I made a mistake and in BOTH run WAVESTDEV = .1, WAVESTDEV2 = 1. I'm gonna do 2 more runs to rectify this. :( r26 - IPROB = 13; good stuff. Maybe not what I want, but it certainly makes sense to me. r27 - Uninteresting; motion not constrained to the horizontal. r28 - Changing zo resulted in a shifting of the positive and negative values of the peaks. Also, the magnitudes of the peaks are greater (.02 vs .006) - The evolution is also different; retains that cell like structure r29 - The wave doesn't seem to move in the horizontal. Odd. r19 - I made a mistake and didn't square the x(i)-x0 term. Probably why the data sucked. SO, 3 runs. r30 - Try running r19, Iprob 30, but with the correct code change to _bl1 r31 - r30 but w/ different WAVESTDEV r32 - r25, but with WAVESTDEV = 1., WAVESTDEV2 = 1. r33 - r26, but with WAVESTDEV = .1, WAVESTDEV2 = .1 ==== data.raw/r30 ===================================================== NX = 250 XMAX = 2.5 NTOTAL = 2100 -> ntim = 52 NY = 1 YMAX = 1.0 NSTEP0 = 150 NZ = 250 ZMAX = 2.5 NSTEP1 = 50 ZMAX2 = 0.5 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 WAVESTDEV = .1 IBX = IBY = IBZ = 0; periodic WAVESTDEV2= .1 IPROB = 30 ------------------------------ Comments ------------------------------- 1. Modified the _bl1 code with the correct (x-xo)**2 ======================================================================= ==== data.raw/r31 ===================================================== NX = 250 XMAX = 2.5 NTOTAL = 2100 -> ntim = 52 NY = 1 YMAX = 1.0 NSTEP0 = 150 NZ = 250 ZMAX = 2.5 NSTEP1 = 50 ZMAX2 = 0.5 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 WAVESTDEV = 1. IBX = IBY = IBZ = 0; periodic WAVESTDEV2= 1. IPROB = 30 ------------------------------ Comments ------------------------------- 1. Like r30, but not quite. Duh. ======================================================================= ==== data.raw/r32 ===================================================== NX = 250 XMAX = 2.5 NTOTAL = 2100 -> ntim = 52 NY = 1 YMAX = 1.0 NSTEP0 = 150 NZ = 250 ZMAX = 2.5 NSTEP1 = 50 ZMAX2 = 0.5 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 WAVESTDEV = 1. IBX = IBY = IBZ = 0; periodic WAVESTDEV2= 1. IPROB = 30 ------------------------------ Comments ------------------------------- 1. Like r30, but not quite. Duh. ======================================================================= ==== data.raw/r32 ===================================================== NX = 250 XMAX = 2.5 NTOTAL = 2100 -> ntim = 52 NY = 1 YMAX = 1.0 NSTEP0 = 150 NZ = 250 ZMAX = 2.5 NSTEP1 = 50 ZMAX2 = 0.5 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 WAVESTDEV = .1 IBX = IBY = IBZ = 0; periodic WAVESTDEV2= .1 IPROB = 13 ------------------------------ Comments ------------------------------- 1. The corrected version of r25 or r26. ======================================================================= ==== data.raw/r33 ===================================================== NX = 250 XMAX = 2.5 NTOTAL = 2100 -> ntim = 52 NY = 1 YMAX = 1.0 NSTEP0 = 150 NZ = 250 ZMAX = 2.5 NSTEP1 = 50 ZMAX2 = 0.5 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 WAVESTDEV = 1. IBX = IBY = IBZ = 0; periodic WAVESTDEV2= 1. IPROB = 13 ------------------------------ Comments ------------------------------- 1. Like r32, but different WAVESTDEV. ======================================================================= ==== data.raw/r34 ===================================================== NX = 250 XMAX = 2.5 NTOTAL = 51 -> ntim = 52 NY = 1 YMAX = 1.0 NSTEP0 = 51 NZ = 250 ZMAX = 2.5 NSTEP1 = 1 ZMAX2 = 0.5 NSTEP2 = 1 NCPU = 5 NSTEP3 = 1 WAVESTDEV = .1 IBX = IBY = IBZ = 0; periodic WAVESTDEV2= .1 IPROB = 30 ------------------------------ Comments ------------------------------- 1. r30 but with shorter time steps. ======================================================================= 2/10/06 kraken goes offline tomorrow for the weekend, so I need to get some runs going. Talked with Joe yesterday, very informative. Unfortunately he won't be here next week, so any questions will have to wait till after. That gives me time to work on Comps III though, I suppose. Anyway, the 2 main things I need to play with are EDGELENGTH, which is a smoothing parameter and was set to 5 when the vertical domain size was 3 (resulting in a massive over smoothing), and to determine if the normalization is wrt to the wavelenght or the vertical wavelength. As I'm seeking a kick with a horizontal wave only, it would be BAD if my normalization used the non-existent vertical wavelength. Finally, I've been playing with IDL and will continue to do so since I've written some nice and flexible code. I don't like the need for a .pl and an o.pl but I don't see a way around it. Finally, to reflect the fact that change has occured, the runs will now be labeled with an 's', not an 'r'. SSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSS ==== data.raw/s01 ===================================================== EXECUTABLE: triple_bl NX = 1000 XMAX = 10.0 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.00 NSTEP0 = 150 NZ = 1000 ZMAX = 10.0 NSTEP1 = 50 ZMAX2 = 00.0 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 IPROB = 30 WAVESTDEV = 1. IBX = IBY = IBZ = 0; periodic TZRATIO = 1; usually = 2.828. EDGELENGTH= 1. ------------------------------ Comments ------------------------------- 1. The first of corrected EDGELENGTH runs. We'll do a couple. ======================================================================= ==== data.raw/s02 ===================================================== EXECUTABLE: triple_bl NX = 1000 XMAX = 10.0 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.00 NSTEP0 = 150 NZ = 1000 ZMAX = 10.0 NSTEP1 = 50 ZMAX2 = 00.0 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 IPROB = 30 WAVESTDEV = .1 IBX = IBY = IBZ = 0; periodic TZRATIO = 1; usually = 2.828. EDGELENGTH= 1. ------------------------------ Comments ------------------------------- 1. s01 with a smaller wavestdev. ======================================================================= ==== data.raw/s03 ===================================================== EXECUTABLE: triple_bl NX = 1000 XMAX = 10.0 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.00 NSTEP0 = 150 NZ = 1000 ZMAX = 10.0 NSTEP1 = 50 ZMAX2 = 02.0 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 IPROB = 30 WAVESTDEV = 1. IBX = IBY = IBZ = 0; periodic TZRATIO = 1; usually = 2.828. EDGELENGTH= 1. ------------------------------ Comments ------------------------------- 1. This run should be identical to s01 since tzratio = 1 (even though zmax2 =/= 0.0 2. I realized that zmax + zmax2 =/= constant, so in the other runs where I set zmax2 = 0.0 and zmax = 3 to compensate, that was stupid. ======================================================================= ==== data.raw/s04 ===================================================== EXECUTABLE: triple_bl NX = 1000 XMAX = 10.0 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.00 NSTEP0 = 150 NZ = 1000 ZMAX = 10.0 NSTEP1 = 50 ZMAX2 = 02.0 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 IPROB = 30 WAVESTDEV = 1. IBX = IBY = IBZ = 0; periodic TZRATIO = 2.828. EDGELENGTH= 1. ------------------------------ Comments ------------------------------- 1. The first real run. edgelength = 1, tzr =/= 1, zmax2 = 1/5 zmax ======================================================================= ==== data.raw/s05 ===================================================== EXECUTABLE: triple_bl NX = 1000 XMAX = 10.0 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.00 NSTEP0 = 150 NZ = 1000 ZMAX = 10.0 NSTEP1 = 50 ZMAX2 = 02.0 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 IPROB = 30 WAVESTDEV = .1 IBX = IBY = IBZ = 0; periodic TZRATIO = 2.828. EDGELENGTH= 1. ------------------------------ Comments ------------------------------- 1. Shrunk wavestdev. ======================================================================= ==== data.raw/s06 ===================================================== EXECUTABLE: triple_bl NX = 1000 XMAX = 10.0 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.00 NSTEP0 = 150 NZ = 1000 ZMAX = 10.0 NSTEP1 = 50 ZMAX2 = 02.0 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 IPROB = 30 WAVESTDEV = 2. IBX = IBY = IBZ = 0; periodic TZRATIO = 2.828. EDGELENGTH= 1. ------------------------------ Comments ------------------------------- 1. Increased wavestdev. Just wanna make sure '1' isn't scaled. i.e., is '1' on the same scale as zmax = 10 or does '1' mean 'all' ======================================================================= ==== data.raw/s07 ===================================================== EXECUTABLE: triple_bl1 - T(l) -> T(l) * e^[-(x-xo)**2/wide**2] recall: wavestdev = wide, wavestdev2 = wide2 NX = 1000 XMAX = 10.0 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.00 NSTEP0 = 150 NZ = 1000 ZMAX = 10.0 NSTEP1 = 50 ZMAX2 = 02.0 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 IPROB = 30 WAVESTDEV = 1. IBX = IBY = IBZ = 0; periodic TZRATIO = 2.828. EDGELENGTH= 1. ------------------------------ Comments ------------------------------- 1. I really want to constrain the wave in x and z, so I did. This is being done w/ FREQOVERN = 1, which I ASSUME give no vertical wave, but I'm not sure. ======================================================================= ==== data.raw/s08 ===================================================== EXECUTABLE: triple_bl1 - T(l) -> T(l) * e^[-(x-xo)**2/wide**2] recall: wavestdev = wide, wavestdev2 = wide2 NX = 1000 XMAX = 10.0 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.00 NSTEP0 = 150 NZ = 1000 ZMAX = 10.0 NSTEP1 = 50 ZMAX2 = 02.0 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 IPROB = 30 WAVESTDEV = .1 IBX = IBY = IBZ = 0; periodic TZRATIO = 2.828. EDGELENGTH= 1. ------------------------------ Comments ------------------------------- 1. s07, but with a smaller WAVESTDEV. ======================================================================= ==== data.raw/s09 ===================================================== EXECUTABLE: triple_bl NX = 1000 XMAX = 10.0 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.00 NSTEP0 = 150 NZ = 1000 ZMAX = 10.0 NSTEP1 = 50 ZMAX2 = 02.0 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 IPROB = 13 WAVESTDEV = 1. IBX = IBY = IBZ = 0; periodic WAVESTDEV2= 1. TZRATIO = 2.828. EDGELENGTH= 1. ------------------------------ Comments ------------------------------- 1. Same setup as s07, but iprob = 13. Iprob 13 wraps the envelope around the wavepacket in along the phase and group speeds, not x and z. But, w/ FREQOVERN = 1, I would assume these are the same, and so this little test. ======================================================================= ==== data.raw/s10 ===================================================== EXECUTABLE: triple_bl NX = 1000 XMAX = 10.0 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.00 NSTEP0 = 150 NZ = 1000 ZMAX = 10.0 NSTEP1 = 50 ZMAX2 = 02.0 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 IPROB = 13 WAVESTDEV = 1. IBX = IBY = IBZ = 0; periodic WAVESTDEV2= 1. TZRATIO = 2.828. EDGELENGTH= 1. FREQOVERN = .5 ------------------------------ Comments ------------------------------- 1. In the spirit of s09 v. s07, s10 and s11 have FREQOVERN = .5, with the idea being the results should be different since the group and phase velocity directions now no longer align with x and z. ======================================================================= ==== data.raw/s11 ===================================================== EXECUTABLE: triple_bl1 - T(l) -> T(l) * e^[-(x-xo)**2/wide**2] recall: wavestdev = wide, wavestdev2 = wide2 NX = 1000 XMAX = 10.0 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.00 NSTEP0 = 150 NZ = 1000 ZMAX = 10.0 NSTEP1 = 50 ZMAX2 = 02.0 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 IPROB = 30 WAVESTDEV = 1. IBX = IBY = IBZ = 0; periodic TZRATIO = 2.828. EDGELENGTH= 1. FREQOVERN = .5 ------------------------------ Comments ------------------------------- 1. Compare this to s10; they should be different. ======================================================================= ==== data.raw/s12 ===================================================== EXECUTABLE: triple_bl NX = 1000 XMAX = 10.0 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.00 NSTEP0 = 150 NZ = 1000 ZMAX = 10.0 NSTEP1 = 50 ZMAX2 = 02.0 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 IPROB = 13 WAVESTDEV = .1 IBX = IBY = IBZ = 0; periodic WAVESTDEV2= .1 TZRATIO = 2.828. EDGELENGTH= 1. FREQOVERN = .5 ------------------------------ Comments ------------------------------- ======================================================================= ==== data.raw/s13 ===================================================== EXECUTABLE: triple_bl1 - T(l) -> T(l) * e^[-(x-xo)**2/wide**2] recall: wavestdev = wide, wavestdev2 = wide2 NX = 1000 XMAX = 10.0 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.00 NSTEP0 = 150 NZ = 1000 ZMAX = 10.0 NSTEP1 = 50 ZMAX2 = 02.0 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 IPROB = 30 WAVESTDEV = .1 IBX = IBY = IBZ = 0; periodic WAVESTDEV2= .1 TZRATIO = 2.828. EDGELENGTH= 1. FREQOVERN = .5 ------------------------------ Comments ------------------------------- ======================================================================= ==== data.raw/s14 ===================================================== EXECUTABLE: triple_bl NX = 1000 XMAX = 10.0 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.00 NSTEP0 = 150 NZ = 1000 ZMAX = 10.0 NSTEP1 = 50 ZMAX2 = 02.0 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 IPROB = 13 WAVESTDEV = .1 IBX = IBY = IBZ = 0; periodic WAVESTDEV2= .1 TZRATIO = 2.828. EDGELENGTH= 1. FREQOVERN = 1 ------------------------------ Comments ------------------------------- ======================================================================= ==== data.raw/s15 ===================================================== EXECUTABLE: triple_bl1 - T(l) -> T(l) * e^[-(x-xo)**2/wide**2] recall: wavestdev = wide, wavestdev2 = wide2 NX = 1000 XMAX = 10.0 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.00 NSTEP0 = 150 NZ = 1000 ZMAX = 10.0 NSTEP1 = 50 ZMAX2 = 02.0 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 IPROB = 30 WAVESTDEV = .1 IBX = IBY = IBZ = 0; periodic WAVESTDEV2= .1 TZRATIO = 2.828. EDGELENGTH= 1. FREQOVERN = 1 ------------------------------ Comments ------------------------------- ======================================================================= Right, so compare IPROB , D , F 7 - 9: 30 to 13, 1., 1. 10-11: 13 to 30, 1., .5 12-13: 13 to 30, .1, .5 14-15: 13 to 30, .1, 1. 2/13/06 It took all weekend to download the txz and uxz files from these runs. Mostly b/c I kept screwing with the damned kftp utility which would hang after ever large .tar file transfer. Also, the initial run where I submitted 15 concurrent jobs was... not the best of ideas. Ultimately I used the scp ability FROM jules TO earth. I've been attempting to look at the output I got and I'm finding that with a 1000 by 1000 data block IDL is having difficulty. Difficulty = 1. a long time to read in and display the data = 2. the maximum window size is ~1010 x 680, and so I can't look at the whole domain at once, which I would like. So. New set of runs. Redo the above 15 runs but this time with nnxp = 500 nnzp = 500 This should be quicker and more managable. I'll change the prefix from 's' to 'f', 'f' for 'five' As those are running I'll play around with IDL in an attempt to look at the data I already have since a lot of time was spent creating it and getting it to Iditarod. Also, in speaking with Joe I learned that the txz files do NOT have the background/mean temperature profile included in them; again, they are contained in the .mean and .mean2 files (.mean2 I believe). You might want to add the mean to the .txz.* file just so you have a better physical grasp of what you're looking at, though I highly recommend waiting until the nnxp = nnzp = 500 files are done before trying this. FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF ==== data.raw/f01 ===================================================== EXECUTABLE: triple_bl NX = 500 XMAX = 10.0 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.00 NSTEP0 = 150 NZ = 500 ZMAX = 10.0 NSTEP1 = 50 ZMAX2 = 00.0 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 IPROB = 30 WAVESTDEV = 1. IBX = IBY = IBZ = 0; periodic TZRATIO = 1; usually = 2.828. EDGELENGTH= 1. ------------------------------ Comments ------------------------------- 1. The first of corrected EDGELENGTH runs. We'll do a couple. 2. ALSO, first of the 'f' runs, with smaller NX and NY ======================================================================= ==== data.raw/f02 ===================================================== EXECUTABLE: triple_bl NX = 500 XMAX = 10.0 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.00 NSTEP0 = 150 NZ = 500 ZMAX = 10.0 NSTEP1 = 50 ZMAX2 = 00.0 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 IPROB = 30 WAVESTDEV = .1 IBX = IBY = IBZ = 0; periodic TZRATIO = 1; usually = 2.828. EDGELENGTH= 1. ------------------------------ Comments ------------------------------- 1. s01 with a smaller wavestdev. ======================================================================= ==== data.raw/f03 ===================================================== EXECUTABLE: triple_bl NX = 500 XMAX = 10.0 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.00 NSTEP0 = 150 NZ = 500 ZMAX = 10.0 NSTEP1 = 50 ZMAX2 = 02.0 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 IPROB = 30 WAVESTDEV = 1. IBX = IBY = IBZ = 0; periodic TZRATIO = 1; usually = 2.828. EDGELENGTH= 1. ------------------------------ Comments ------------------------------- 1. This run should be identical to s01 since tzratio = 1 (even though zmax2 =/= 0.0 2. I realized that zmax + zmax2 =/= constant, so in the other runs where I set zmax2 = 0.0 and zmax = 3 to compensate, that was stupid. ======================================================================= ==== data.raw/f04 ===================================================== EXECUTABLE: triple_bl NX = 500 XMAX = 10.0 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.00 NSTEP0 = 150 NZ = 500 ZMAX = 10.0 NSTEP1 = 50 ZMAX2 = 02.0 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 IPROB = 30 WAVESTDEV = 1. IBX = IBY = IBZ = 0; periodic TZRATIO = 2.828. EDGELENGTH= 1. ------------------------------ Comments ------------------------------- 1. The first real run. edgelength = 1, tzr =/= 1, zmax2 = 1/5 zmax ======================================================================= ==== data.raw/f05 ===================================================== EXECUTABLE: triple_bl NX = 500 XMAX = 10.0 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.00 NSTEP0 = 150 NZ = 500 ZMAX = 10.0 NSTEP1 = 50 ZMAX2 = 02.0 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 IPROB = 30 WAVESTDEV = .1 IBX = IBY = IBZ = 0; periodic TZRATIO = 2.828. EDGELENGTH= 1. ------------------------------ Comments ------------------------------- 1. Shrunk wavestdev. ======================================================================= ==== data.raw/f06 ===================================================== EXECUTABLE: triple_bl NX = 500 XMAX = 10.0 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.00 NSTEP0 = 150 NZ = 500 ZMAX = 10.0 NSTEP1 = 50 ZMAX2 = 02.0 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 IPROB = 30 WAVESTDEV = 2. IBX = IBY = IBZ = 0; periodic TZRATIO = 2.828. EDGELENGTH= 1. ------------------------------ Comments ------------------------------- 1. Increased wavestdev. Just wanna make sure '1' isn't scaled. i.e., is '1' on the same scale as zmax = 10 or does '1' mean 'all' ======================================================================= ==== data.raw/f07 ===================================================== EXECUTABLE: triple_bl1 - T(l) -> T(l) * e^[-(x-xo)**2/wide**2] recall: wavestdev = wide, wavestdev2 = wide2 NX = 500 XMAX = 10.0 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.00 NSTEP0 = 150 NZ = 500 ZMAX = 10.0 NSTEP1 = 50 ZMAX2 = 02.0 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 IPROB = 30 WAVESTDEV = 1. IBX = IBY = IBZ = 0; periodic TZRATIO = 2.828. EDGELENGTH= 1. ------------------------------ Comments ------------------------------- 1. I really want to constrain the wave in x and z, so I did. This is being done w/ FREQOVERN = 1, which I ASSUME give no vertical wave, but I'm not sure. ======================================================================= ==== data.raw/f08 ===================================================== EXECUTABLE: triple_bl1 - T(l) -> T(l) * e^[-(x-xo)**2/wide**2] recall: wavestdev = wide, wavestdev2 = wide2 NX = 500 XMAX = 10.0 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.00 NSTEP0 = 150 NZ = 500 ZMAX = 10.0 NSTEP1 = 50 ZMAX2 = 02.0 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 IPROB = 30 WAVESTDEV = .1 IBX = IBY = IBZ = 0; periodic TZRATIO = 2.828. EDGELENGTH= 1. ------------------------------ Comments ------------------------------- 1. s07, but with a smaller WAVESTDEV. ======================================================================= ==== data.raw/f09 ===================================================== EXECUTABLE: triple_bl NX = 500 XMAX = 10.0 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.00 NSTEP0 = 150 NZ = 500 ZMAX = 10.0 NSTEP1 = 50 ZMAX2 = 02.0 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 IPROB = 13 WAVESTDEV = 1. IBX = IBY = IBZ = 0; periodic WAVESTDEV2= 1. TZRATIO = 2.828. EDGELENGTH= 1. ------------------------------ Comments ------------------------------- 1. Same setup as s07, but iprob = 13. Iprob 13 wraps the envelope around the wavepacket in along the phase and group speeds, not x and z. But, w/ FREQOVERN = 1, I would assume these are the same, and so this little test. ======================================================================= ==== data.raw/f10 ===================================================== EXECUTABLE: triple_bl NX = 500 XMAX = 10.0 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.00 NSTEP0 = 150 NZ = 500 ZMAX = 10.0 NSTEP1 = 50 ZMAX2 = 02.0 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 IPROB = 13 WAVESTDEV = 1. IBX = IBY = IBZ = 0; periodic WAVESTDEV2= 1. TZRATIO = 2.828. EDGELENGTH= 1. FREQOVERN = .5 ------------------------------ Comments ------------------------------- 1. In the spirit of s09 v. s07, s10 and s11 have FREQOVERN = .5, with the idea being the results should be different since the group and phase velocity directions now no longer align with x and z. ======================================================================= ==== data.raw/f11 ===================================================== EXECUTABLE: triple_bl1 - T(l) -> T(l) * e^[-(x-xo)**2/wide**2] recall: wavestdev = wide, wavestdev2 = wide2 NX = 500 XMAX = 10.0 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.00 NSTEP0 = 150 NZ = 500 ZMAX = 10.0 NSTEP1 = 50 ZMAX2 = 02.0 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 IPROB = 30 WAVESTDEV = 1. IBX = IBY = IBZ = 0; periodic TZRATIO = 2.828. EDGELENGTH= 1. FREQOVERN = .5 ------------------------------ Comments ------------------------------- 1. Compare this to s10; they should be different. ======================================================================= ==== data.raw/f12 ===================================================== EXECUTABLE: triple_bl NX = 500 XMAX = 10.0 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.00 NSTEP0 = 150 NZ = 500 ZMAX = 10.0 NSTEP1 = 50 ZMAX2 = 02.0 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 IPROB = 13 WAVESTDEV = .1 IBX = IBY = IBZ = 0; periodic WAVESTDEV2= .1 TZRATIO = 2.828. EDGELENGTH= 1. FREQOVERN = .5 ------------------------------ Comments ------------------------------- ======================================================================= ==== data.raw/f13 ===================================================== EXECUTABLE: triple_bl1 - T(l) -> T(l) * e^[-(x-xo)**2/wide**2] recall: wavestdev = wide, wavestdev2 = wide2 NX = 500 XMAX = 10.0 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.00 NSTEP0 = 150 NZ = 500 ZMAX = 10.0 NSTEP1 = 50 ZMAX2 = 02.0 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 IPROB = 30 WAVESTDEV = .1 IBX = IBY = IBZ = 0; periodic WAVESTDEV2= .1 TZRATIO = 2.828. EDGELENGTH= 1. FREQOVERN = .5 ------------------------------ Comments ------------------------------- ======================================================================= ==== data.raw/f14 ===================================================== EXECUTABLE: triple_bl NX = 500 XMAX = 10.0 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.00 NSTEP0 = 150 NZ = 500 ZMAX = 10.0 NSTEP1 = 50 ZMAX2 = 02.0 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 IPROB = 13 WAVESTDEV = .1 IBX = IBY = IBZ = 0; periodic WAVESTDEV2= .1 TZRATIO = 2.828. EDGELENGTH= 1. FREQOVERN = 1 ------------------------------ Comments ------------------------------- ======================================================================= ==== data.raw/f15 ===================================================== EXECUTABLE: triple_bl1 - T(l) -> T(l) * e^[-(x-xo)**2/wide**2] recall: wavestdev = wide, wavestdev2 = wide2 NX = 500 XMAX = 10.0 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.00 NSTEP0 = 150 NZ = 500 ZMAX = 10.0 NSTEP1 = 50 ZMAX2 = 02.0 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 IPROB = 13 WAVESTDEV = .1 IBX = IBY = IBZ = 0; periodic WAVESTDEV2= .1 TZRATIO = 2.828. EDGELENGTH= 1. FREQOVERN = 1 ------------------------------ Comments ------------------------------- 1. Dammit. f14 and f15 seem to both have the same damned codein. Like identical. Both try and write to the f14 directory. So 15 doesn't exactly tell you anything new. ======================================================================= 2/14/06 Cool; examining the f?? output. f07: This is the first really neat one. You can clearly see the x and z envelope. f01/f02: In looking at these two, WAVESTDEV seems to have a huge role which I don't understand very well yet. f14/f15: Seem to be identical in the last time step. f10/f11: Seem to be identical in the last time step. f12/f13: Seem to be identical in the last time step. f09/f07: Seem to be identical in the last time step. I't kinda' distressing that the 13 and 30 runs are so similar when one should do an envelope in the phase/perp dir and the other in the x/z dir. Also, I seem to be getting a LOT of edge effects. Looking at f15, it doesn't seem like the envelope in x is very effective. Maybe its a scale or resolution thing causing me to miss it? f05/f08: The only diff between these 2 is the executable; f08 tries to constrain both x and z. WAVESTDEV = .1 2/15/06 ==== data.raw/f16 ===================================================== EXECUTABLE: triple_bl NX = 500 XMAX = 10.0 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.00 NSTEP0 = 150 NZ = 500 ZMAX = 10.0 NSTEP1 = 50 ZMAX2 = 00.0 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 IPROB = 30 WAVESTDEV = 2. IBX = IBY = IBZ = 0; periodic WAVESTDEV2= 1. TZRATIO = 1.000 EDGELENGTH= 1. FREQOVERN = 1 ------------------------------ Comments ------------------------------- ======================================================================= ==== data.raw/f17 ===================================================== EXECUTABLE: triple_bl NX = 500 XMAX = 10.0 NTOTAL = 2100 -> ntim = 43 NY = 1 YMAX = 1.00 NSTEP0 = 150 NZ = 500 ZMAX = 10.0 NSTEP1 = 50 ZMAX2 = 00.0 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 IPROB = 30 WAVESTDEV = 10. IBX = IBY = IBZ = 0; periodic WAVESTDEV2= 1. TZRATIO = 1.000 EDGELENGTH= 1. FREQOVERN = 1 ------------------------------ Comments ------------------------------- 1. f16 and f17 play with WAVESTDEV; compare these to runs f01 and f02. Also, realize from f03 that zmax2 =/= 0 may affect more than you know. ======================================================================= 2/17/06 I took off yesterday. I'll work tomorrow. Cosmic Balance is thus ensured. Had I said "thus insured" that'd have a whole different meaning. RIGHT. So, instead of dicking around with 43 timesteps creating giant files that time evolve something I don't really care about (since I don't have the initialization correct yet) I've set up some 1 time step runs. The idea is to see if these are useful in setting up my initial conditions (wave amplitude, the envelope, the emask, the location of the packet, etc..) These are labled with a 't', for 'time'. Neat, huh? Also, I'm slight worried about using the phrase 'dicking around' in the above paragraph as this is a work related document which could potentially be viewed by others. But it really is the most accurate and descriptive term out there and for the purposes of this investigation (???) will stand. ==== data.raw/t01 ===================================================== EXECUTABLE: triple_bl NX = 500 XMAX = 10.0 NTOTAL = 50 -> ntim = 2 NY = 1 YMAX = 1.00 NSTEP0 = 50 NZ = 500 ZMAX = 10.0 NSTEP1 = 50 ZMAX2 = 00.0 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 IPROB = 30 WAVESTDEV = 1. IBX = IBY = IBZ = 0; periodic WAVESTDEV2= 1. TZRATIO = 1.000 EDGELENGTH= 1. FREQOVERN = 1 ------------------------------ Comments ------------------------------- 1. A zero timestep run; compare to f01 ======================================================================= ==== data.raw/t02 ===================================================== EXECUTABLE: triple_bl NX = 500 XMAX = 10.0 NTOTAL = 2100 -> ntim = 2 NY = 1 YMAX = 1.00 NSTEP0 = 2100 NZ = 500 ZMAX = 10.0 NSTEP1 = 2100 ZMAX2 = 00.0 NSTEP2 = 2100 NCPU = 5 NSTEP3 = 2100 IPROB = 30 WAVESTDEV = 1. IBX = IBY = IBZ = 0; periodic WAVESTDEV2= 1. TZRATIO = 1.000 EDGELENGTH= 1. FREQOVERN = 1 ------------------------------ Comments ------------------------------- 1. Also a zero timestep run; compare to the end of f01. ======================================================================= ==== data.raw/t03 ===================================================== EXECUTABLE: triple_bl NX = 500 XMAX = 10.0 NTOTAL = 50 -> ntim = 2 NY = 1 YMAX = 1.00 NSTEP0 = 50 NZ = 500 ZMAX = 10.0 NSTEP1 = 50 ZMAX2 = 00.0 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 IPROB = 30 AMP = 2.0 WAVESTDEV = 1. IBX = IBY = IBZ = 0; periodic WAVESTDEV2= 1. TZRATIO = 1.000 EDGELENGTH= 1. FREQOVERN = 1 ------------------------------ Comments ------------------------------- 1. This is run t01, but with AMP set to 2 instead of 0.175070. I want to see if I can get my disturbance to the order of the background temperature profile. ======================================================================= ==== data.raw/t04 ===================================================== EXECUTABLE: triple_bl NX = 500 XMAX = 10.0 NTOTAL = 50 -> ntim = 2 NY = 1 YMAX = 1.00 NSTEP0 = 50 NZ = 500 ZMAX = 10.0 NSTEP1 = 50 ZMAX2 = 00.0 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 IPROB = 30 WAVESTDEV = 1. IBX = IBY = IBZ = 0; periodic WAVESTDEV2= 1. TZRATIO = 10.00 EDGELENGTH= 1. FREQOVERN = 1 ------------------------------ Comments ------------------------------- 1. Increased TZRATIO to see its effect. BUT, I need a non-zero zmax2. runs t05 and t06 will take care of that. ======================================================================= ==== data.raw/t05 ===================================================== EXECUTABLE: triple_bl NX = 500 XMAX = 10.0 NTOTAL = 50 -> ntim = 2 NY = 1 YMAX = 1.00 NSTEP0 = 50 NZ = 500 ZMAX = 10.0 NSTEP1 = 50 ZMAX2 = 02.0 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 IPROB = 30 WAVESTDEV = 1. IBX = IBY = IBZ = 0; periodic WAVESTDEV2= 1. TZRATIO = 10.00 EDGELENGTH= 1. FREQOVERN = 1 ------------------------------ Comments ------------------------------- ======================================================================= ==== data.raw/t06 ===================================================== EXECUTABLE: triple_bl NX = 500 XMAX = 10.0 NTOTAL = 50 -> ntim = 2 NY = 1 YMAX = 1.00 NSTEP0 = 50 NZ = 500 ZMAX = 10.0 NSTEP1 = 50 ZMAX2 = 01.0 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 IPROB = 30 WAVESTDEV = 1. IBX = IBY = IBZ = 0; periodic WAVESTDEV2= 1. TZRATIO = 10.00 EDGELENGTH= 1. FREQOVERN = 1 ------------------------------ Comments ------------------------------- ======================================================================= ==== data.raw/t07 ===================================================== EXECUTABLE: triple_bl NX = 500 XMAX = 100. NTOTAL = 50 -> ntim = 2 NY = 1 YMAX = 1.00 NSTEP0 = 50 NZ = 500 ZMAX = 10.0 NSTEP1 = 50 ZMAX2 = 01.0 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 IPROB = 30 WAVESTDEV = 1. IBX = IBY = IBZ = 0; periodic WAVESTDEV2= 1. TZRATIO = 10.00 EDGELENGTH= 1. FREQOVERN = 1 ------------------------------ Comments ------------------------------- 1. t06, but with a greatly expanded xmax. This means the horizontal resolution is not as fine as the vertical. ======================================================================= ==== data.raw/t08 ===================================================== EXECUTABLE: triple_bl NX = 500 XMAX = 10.0 NTOTAL = 50 -> ntim = 2 NY = 1 YMAX = 1.00 NSTEP0 = 50 NZ = 500 ZMAX = 10.0 NSTEP1 = 50 ZMAX2 = 00.0 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 IPROB = 30 AMP = 10.0 WAVESTDEV = 1. IBX = IBY = IBZ = 0; periodic WAVESTDEV2= 1. TZRATIO = 1.000 EDGELENGTH= 1. FREQOVERN = 1 ------------------------------ Comments ------------------------------- 1. t03 is neat since you can see the variation in the temp field due to the wave over the background/mean. t08 is to see what happens if I bump it up further. ======================================================================= COMENTS: comparing f01 to t03: For the 'u' plot, the tv plots look the same, but comparing profile plots along z = 150 shows that the amplitude of t03 is ~.51 and for f01 is ~.05 It would be interesting to run t03 out to 2100 and see how it compares, and since I can do that as a 1 timestep run, I will. ==== data.raw/t09 ===================================================== EXECUTABLE: triple_bl NX = 500 XMAX = 10.0 NTOTAL = 2100 -> ntim = 2 NY = 1 YMAX = 1.00 NSTEP0 = 2100 NZ = 500 ZMAX = 10.0 NSTEP1 = 2100 ZMAX2 = 00.0 NSTEP2 = 2100 NCPU = 5 NSTEP3 = 2100 IPROB = 30 AMP = 2.0 WAVESTDEV = 1. IBX = IBY = IBZ = 0; periodic WAVESTDEV2= 1. TZRATIO = 1.000 EDGELENGTH= 1. FREQOVERN = 1 ------------------------------ Comments ------------------------------- 1. t03, but at 2100. Compare with f01 at tstep 43 to see if the 2 runs evolve similarly. ======================================================================= ==== data.raw/t10 ===================================================== EXECUTABLE: triple_bl NX = 500 XMAX = 10.0 NTOTAL = 50 -> ntim = 2 NY = 1 YMAX = 1.00 NSTEP0 = 50 NZ = 500 ZMAX = 10.0 NSTEP1 = 50 ZMAX2 = 02.0 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 IPROB = 30 WAVESTDEV = 1. IBX = IBY = IBZ = 0; periodic WAVESTDEV2= 1. TZRATIO = 100.0 EDGELENGTH= 1. FREQOVERN = 1 ------------------------------ Comments ------------------------------- 1. Comparing 4,5,and 6 makes me wonder about the t profile. t10 has an ungodly large tzratio to explore this. ======================================================================= ==== data.raw/t11 ===================================================== EXECUTABLE: triple_bl NX = 500 XMAX = 10.0 NTOTAL = 50 -> ntim = 2 NY = 1 YMAX = 1.00 NSTEP0 = 50 NZ = 500 ZMAX = 10.0 NSTEP1 = 50 ZMAX2 = 0.10 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 IPROB = 30 WAVESTDEV = 1. IBX = IBY = IBZ = 0; periodic WAVESTDEV2= 1. TZRATIO = 10.0 EDGELENGTH= 1. FREQOVERN = 1 ------------------------------ Comments ------------------------------- 1. exploring the tprofile further; zmax2 is set to a very small, though non-zero number. ======================================================================= ==== data.raw/t12 ===================================================== EXECUTABLE: triple_bl NX = 500 XMAX = 10.0 NTOTAL = 50 -> ntim = 2 NY = 1 YMAX = 1.00 NSTEP0 = 50 NZ = 500 ZMAX = 10.0 NSTEP1 = 50 ZMAX2 = 0.50 NSTEP2 = 50 NCPU = 5 NSTEP3 = 50 IPROB = 30 WAVESTDEV = 1. IBX = IBY = IBZ = 0; periodic WAVESTDEV2= 1. TZRATIO = 10.0 EDGELENGTH= 1. FREQOVERN = 1 ------------------------------ Comments ------------------------------- ======================================================================= 2/19/06 d-runs. These are to edit the _bl3 code to accept user supplied values of x0 and z0 in the wavepacket routine. Since my runs are all 2D and have rx=ry=0, I'm going to use THEM to be my x0, y0. I'm doing this since I'm not entirely sure how to edit the code to accept new values from codein. Currently, ry=0 since, for a 2D run, ry=/=0 causes the code to not even execute. Also, just affirmed that editing the tsetup file doesn't help, as the triple code it probably what needs editing.