- 新着としてマーク
- ブックマーク
- 購読
- ミュート
- RSS フィードを購読する
- ハイライト
- 印刷
- 不適切なコンテンツを報告
Hi,
a colleague of mine is experiencing bizarrely long compilation times for one particular
file - more than a minute whereas if I compile it with the same version of the Intel compiler
it takes less than a second.
The file is several 100s of lines of code, but nothing specifically complex, and it includes
a few other source files. I estimate the total number of lines to be 3000.
If the compilation time had been consistently long, I could have imagined that there
was something awkward in the source code, but that is not the case. It is just on this
one laptop where it is taking so long (I tried two other machines myself and another
colleague does not experience it either).
The compilation time drops to 1 second or thereabout if we remove the included files.
Any suggestions as to what might be slowing down the compilation process? Should
we reinstall the compiler?
Regards,
Arjen
a colleague of mine is experiencing bizarrely long compilation times for one particular
file - more than a minute whereas if I compile it with the same version of the Intel compiler
it takes less than a second.
The file is several 100s of lines of code, but nothing specifically complex, and it includes
a few other source files. I estimate the total number of lines to be 3000.
If the compilation time had been consistently long, I could have imagined that there
was something awkward in the source code, but that is not the case. It is just on this
one laptop where it is taking so long (I tried two other machines myself and another
colleague does not experience it either).
The compilation time drops to 1 second or thereabout if we remove the included files.
Any suggestions as to what might be slowing down the compilation process? Should
we reinstall the compiler?
Regards,
Arjen
コピーされたリンク
10 返答(返信)
- 新着としてマーク
- ブックマーク
- 購読
- ミュート
- RSS フィードを購読する
- ハイライト
- 印刷
- 不適切なコンテンツを報告
Look at his INCLUDE path and see if there are network shares on it. Are the include files in the same folder as the source? Are these on a local disk or network share?
- 新着としてマーク
- ブックマーク
- 購読
- ミュート
- RSS フィードを購読する
- ハイライト
- 印刷
- 不適切なコンテンツを報告
That should not matter: my own experiment was on a network drive. Besides the project only uses files
and include directories and the like from the directory tree that contains it.
Oh, the compiler version is Intel Fortran 11.0.072
Regards,
Arjen
and include directories and the like from the directory tree that contains it.
Oh, the compiler version is Intel Fortran 11.0.072
Regards,
Arjen
- 新着としてマーク
- ブックマーク
- 購読
- ミュート
- RSS フィードを購読する
- ハイライト
- 印刷
- 不適切なコンテンツを報告
It could matter. Please try again with files local and no network shares on the INCLUDE path and see if the behavior changes.
- 新着としてマーク
- ブックマーク
- 購読
- ミュート
- RSS フィードを購読する
- ハイライト
- 印刷
- 不適切なコンテンツを報告
Okay, Steve,
I will ask him to do that experiment.
Regards,
Arjen
I will ask him to do that experiment.
Regards,
Arjen
- 新着としてマーク
- ブックマーク
- 購読
- ミュート
- RSS フィードを購読する
- ハイライト
- 印刷
- 不適切なコンテンツを報告
Hi Steve,
it turns out that the disk on which he worked was encrypted (laptops get stolen :() - this encryption software
normally does not have any impact on the performance, but curiously enough in this case it had a great
impact.
Thanks for pointing us to the solution.
Regards,
Arjen
it turns out that the disk on which he worked was encrypted (laptops get stolen :() - this encryption software
normally does not have any impact on the performance, but curiously enough in this case it had a great
impact.
Thanks for pointing us to the solution.
Regards,
Arjen
- 新着としてマーク
- ブックマーク
- 購読
- ミュート
- RSS フィードを購読する
- ハイライト
- 印刷
- 不適切なコンテンツを報告
That's great news. For what it's worth, all of our laptops use PGP Whole Disk Encryption and I frequently do Fortran builds on mine without noticing any slowdown.
- 新着としてマーク
- ブックマーク
- 購読
- ミュート
- RSS フィードを購読する
- ハイライト
- 印刷
- 不適切なコンテンツを報告
Well, my colleague uses Trucrypt, but only this particular file seems to be affected.
Anyway, the riddle is brought back to manageable proportions ;)
Regards,
Arjen
Anyway, the riddle is brought back to manageable proportions ;)
Regards,
Arjen
- 新着としてマーク
- ブックマーク
- 購読
- ミュート
- RSS フィードを購読する
- ハイライト
- 印刷
- 不適切なコンテンツを報告
Some further information: the long compile time happens if optimisation is in effect.
The inclusion of the files makesfour subroutines internal. Can that have such a large
effect?
Regards,
Arjen
The inclusion of the files makesfour subroutines internal. Can that have such a large
effect?
Regards,
Arjen
- 新着としてマーク
- ブックマーク
- 購読
- ミュート
- RSS フィードを購読する
- ハイライト
- 印刷
- 不適切なコンテンツを報告
I have no idea why that should matter, unless the disk encryption really slows down disk writes.
- 新着としてマーク
- ブックマーク
- 購読
- ミュート
- RSS フィードを購読する
- ハイライト
- 印刷
- 不適切なコンテンツを報告
If it does, then it is still odd that it should happen with that particular file and that set of compiler options.
At least we now know what was causing the delay.
Regards,
Arjen
At least we now know what was causing the delay.
Regards,
Arjen
