gem__lo_decode_locator: locator line sequence number out of rangecompilation aborted for

[dfumento]

I get the following message with the most recent 8.0 Fortran.

gem__lo_decode_locator: locator line sequence number out of rangecompilation aborted for ababcut2.f

When I compile with the -axW option.

Do you know what is causing it and how to fix it. It doesn't tell the line number so I don't know what I should fix in the code.

Thanks,
David

[dfumento]

Hi,
I meant to add that this is legacy code that works with most recent 7.1 when compiled with the same options.

Thanks,
David

[Steven_L_Intel1]

I see other reports of this in our support database. Try compiling without -ip or -g as a workaround.

[dfumento]

These are the only flags that I was using besides the -axW

-w -c -r8 -O3 -tpp7

It compiles about 200,000 lines of code ok (over an hour), then I get the error, so is there anything particular in the code that I can change?

Thanks,
David

[Steven_L_Intel1]

What specific compiler version are you using?

All of the cases we have seen of this were using -g, and that triggered the problem.

[dfumento]

Intel Fortran Compiler for 32-bit applications, Version 8.0 Build 20040412Z Package ID: l_fc_pc_8.0.046

I've attached the file. There are lots of "include" files but perhaps there is something in the syntax. I wish the compiler would tell you which line numbers it was having trouble with.

Thanks,
David

[dfumento]

Here are the rest of the messages from the compile.


gem__lo_decode_locator: locator line sequence number out of range/home/friesner/fumento/dinger/r2004-1/jaguar-src/main/atei2ndb/ababcut2.f(107) : (col. 11) remark: LOOP WAS VECTORIZED.
/home/friesner/fumento/dinger/r2004-1/jaguar-src/main/atei2ndb/ababcut2.f(146) : (col. 9) remark: LOOP WAS VECTORIZED.
compilation aborted for /home/friesner/fumento/dinger/r2004-1/jaguar-src/main/atei2ndb/ababcut2.f (code 3)
Command failed with exit code 3 (0)

[Steven_L_Intel1]

I would suggest you report this through Premier Support.

[dfumento]

Done. Thanks,David

[dfumento]

This is the answer I got from Premier Support


Actually the problem occurs when you use both -O3 -axW together. The compiler breaks during the cache optimization phase. If you disable cache optimization, the code compiles OK. I'm not sure how much performance penaly you may pay by disabling cache optimization. I filed a report on this issue and will let you know when I get an update.

Here's how you can compile:

$ ifort -w -c -r8 -O3 -axW -mP2OPT_hlo_cache_opt=FALSE ababcut3.f
ababcut3.f(1) : (col. 17) remark: ababcut2_ has been targeted for automatic cpu dispatch.