# M3diff tool help **URL:** \ **Category:** SMOKE\ **Tags:** admin, smoke\ **Created:** [April 9, 2021, 5:10am UTC](https://forum.cmascenter.org/t/m3diff-tool-help/2232 "2021-04-09T05:10:05Z")\ **Posts on this page:** 12\ **Page:** 2
**Author:** ![eyth.alison](https://avatars.discourse-cdn.com/v4/letter/e/7ab992/32.png) [@eyth.alison](https://forum.cmascenter.org/u/eyth.alison)\ **Post date:** [April 13, 2021, 8:04pm UTC](https://forum.cmascenter.org/t/m3diff-tool-help/2232/21 "2021-04-13T20:04:24Z")
This has been a long thread. Do you now have a file for 5/1 that is a similar to the other dates or is this still a problem? If you are still having problems, it would be helpful to attach your smkinven log. The warning message which is tripping up the log analyzer is: 0,False,1,‘WARNING: Duplicate source in inventory. Will store only one value for’,/home/Michael/Documents/SMOKE/2016setupEPA/2028fh1\_16j/intermed/ptegu\_20160501/logs/smkinven\_ptegu\_20160501\_may\_daily\_2028fh1\_16j.log All other messages in the log analyzer, including the “DEFAULT\_CONV\_FAC\_ERROR” and “MAXERROR” messages, are okay. The first column in that level 1 file indicates a message priority. Anything with a 2 or 3 is a known common warning listed in our “known messages” file. Anything with a 0 or 1 is generally worth looking into. 0 usually means the warning is not known, and that is the case with this “duplicate source in inventory” warning. We aren’t familiar with this error, so that is why your log file might be helpful. ---
**Author:** ![Iqbal](https://avatars.discourse-cdn.com/v4/letter/i/bc8723/32.png) [@Iqbal](https://forum.cmascenter.org/u/Iqbal)\ **Post date:** [April 13, 2021, 8:51pm UTC](https://forum.cmascenter.org/t/m3diff-tool-help/2232/22 "2021-04-13T20:51:45Z")
Hello @eyth.alison Yes, I have a file 5/1 which size are same as other dates and EPA. Please see the smkinven log files [smkinven\_ptegu\_20160501\_2028fh1\_16j.log.txt](https://forum.cmascenter.org/uploads/short-url/wd5fMZkakXHLUxg7YimmNRO6eAE.txt) (35.6 KB) [smkinven\_ptegu\_20160501\_may\_daily\_2028fh1\_16j.log.txt](https://forum.cmascenter.org/uploads/short-url/xuDZr6Qn6sjTMVKgMCeeCNx4mZO.txt) (480.0 KB) Thank you ---
**Author:** ![eyth.alison](https://avatars.discourse-cdn.com/v4/letter/e/7ab992/32.png) [@eyth.alison](https://forum.cmascenter.org/u/eyth.alison)\ **Post date:** [April 14, 2021, 12:44pm UTC](https://forum.cmascenter.org/t/m3diff-tool-help/2232/23 "2021-04-14T12:44:12Z")
Your file is probably fine if it’s size is fine. We checked our logs and found that we saw a similar warning in our processing for the 2028fh1 run. The source in question doesn’t actually have any duplicate records in the inventory, so we think this may be a superfluous warning message. This could be added to the known\_messages list so it doesn’t trip up the log analyzer going forward, but for your processing, you should be OK if the file size looks fine. ---
**Author:** ![Iqbal](https://avatars.discourse-cdn.com/v4/letter/i/bc8723/32.png) [@Iqbal](https://forum.cmascenter.org/u/Iqbal)\ **Post date:** [April 14, 2021, 6:47pm UTC](https://forum.cmascenter.org/t/m3diff-tool-help/2232/24 "2021-04-14T18:47:37Z")
Hello @cjcoats @bbaek @eyth.alison Looks like I am able to run it. Here is what I did: 1. I used inlineto2d tool to modify both (EPA & My run) emission files for 20160501 2. Then I ran m3diff infileA infileB DEFAULT, (provides A-B) which gives me [m3diff20160501de.txt](https://forum.cmascenter.org/uploads/short-url/oRGXrYGPMfvEkAJxmoBVLQKBbMu.txt) (586.5 KB) file without an error. I checked MAX A-B value , most of the are zero , but some values are not. Such as top three values are: 3.34E-06 9.54E-07 9.54E-07 **Is this an issue?** 1. Then I tried to run m3diff for ‘non-default’ option (A-B)/RMS , and **I get the error " A:B nowhere defined"** [m3diff20160501mo.txt](https://forum.cmascenter.org/uploads/short-url/3ZXs69yGZKKaalWxSyoDdlPefGx.txt) (9.5 KB) the part of the terminal output is below: ---
**Author:** ![eyth.alison](https://avatars.discourse-cdn.com/v4/letter/e/7ab992/32.png) [@eyth.alison](https://forum.cmascenter.org/u/eyth.alison)\ **Post date:** [April 14, 2021, 7:07pm UTC](https://forum.cmascenter.org/t/m3diff-tool-help/2232/25 "2021-04-14T19:07:08Z")
I would think that differences that small would be OK. ---
**Author:** ![Iqbal](https://avatars.discourse-cdn.com/v4/letter/i/bc8723/32.png) [@Iqbal](https://forum.cmascenter.org/u/Iqbal)\ **Post date:** [April 14, 2021, 7:16pm UTC](https://forum.cmascenter.org/t/m3diff-tool-help/2232/26 "2021-04-14T19:16:47Z")
@eyth.alison thank you so much for the response. Would you please suggest me why ‘non-default’ option (A-B)/RMS is not working? . ---
**Author:** ![cjcoats](https://avatars.discourse-cdn.com/v4/letter/c/ac8455/32.png) [@cjcoats](https://forum.cmascenter.org/u/cjcoats)\ **Post date:** [April 14, 2021, 8:53pm UTC](https://forum.cmascenter.org/t/m3diff-tool-help/2232/27 "2021-04-14T20:53:02Z")
As I said earlier, you **don’t want an output file** (which you have not set up properly anyway, which is the cause of your errors); what you want is the grid-statistics report to `STDOUT`, using _difference normalized by pointwise mean_ (menu item 11), not _difference normalized by joint root mean square_ (menu item 13). ---
**Author:** ![Iqbal](https://avatars.discourse-cdn.com/v4/letter/i/bc8723/32.png) [@Iqbal](https://forum.cmascenter.org/u/Iqbal)\ **Post date:** [April 14, 2021, 9:33pm UTC](https://forum.cmascenter.org/t/m3diff-tool-help/2232/28 "2021-04-14T21:33:00Z")
@cjcoats Thank you so much. I got it. I just set output file ‘N’ and it gives me output in terminal for option 11. ---
**Author:** ![Iqbal](https://avatars.discourse-cdn.com/v4/letter/i/bc8723/32.png) [@Iqbal](https://forum.cmascenter.org/u/Iqbal)\ **Post date:** [April 16, 2021, 8:50pm UTC](https://forum.cmascenter.org/t/m3diff-tool-help/2232/29 "2021-04-16T20:50:39Z")
Hello @eyth.alison m3diff tool gives difference and top three are 3.34E-06, 9.54E-07, 9.54E-07. You mentioned this tiny differences are okay ( because of different machines?) and I understand that. But I have a question: **What is the limit / threshold of differences to say ‘difference is not significant’ or ’ difference is significant’?** ---
**Author:** ![cjcoats](https://avatars.discourse-cdn.com/v4/letter/c/ac8455/32.png) [@cjcoats](https://forum.cmascenter.org/u/cjcoats)\ **Post date:** [April 16, 2021, 9:01pm UTC](https://forum.cmascenter.org/t/m3diff-tool-help/2232/30 "2021-04-16T21:01:57Z")
This gives the relative difference: 3.34E-06 says that the numbers agree to about one part in 1/3.34E-06 = 299401, i.e., the agree to about 5.5 significant decimal digits. Do a round-off error analysis of the calculations involved,. The inputs are IEEE REAL\*4, accurate to the binary equivalent of almost 7 significant decimal digits. **Mathematics (numerical analysis) problem**: how large do you expect the round-off error to be? You certainly **cannot** count on the 7th significant digit in the results; how much round-off error you can expect depends on the nature of the calculation… agreeing to 5 decimal digits is reasonably to be expected. ---
**Author:** ![Iqbal](https://avatars.discourse-cdn.com/v4/letter/i/bc8723/32.png) [@Iqbal](https://forum.cmascenter.org/u/Iqbal)\ **Post date:** [April 17, 2021, 4:09pm UTC](https://forum.cmascenter.org/t/m3diff-tool-help/2232/31 "2021-04-17T16:09:49Z")
Thank you for the explanation @cjcoats . I greatly appreciate you. ---
**Author:** ![warvid](https://avatars.discourse-cdn.com/v4/letter/w/97f17d/32.png) [@warvid](https://forum.cmascenter.org/u/warvid)\ **Post date:** [October 18, 2021, 1:54pm UTC](https://forum.cmascenter.org/t/m3diff-tool-help/2232/32 "2021-10-18T13:54:51Z")
Thank you for the suggestions. I greatly appreciate that[.](https://get-vidmate.com) [.](https://instasave.onl/) [Previous page](https://forum.cmascenter.org/t/m3diff-tool-help/2232.md?page=1)