- 신규로 표시
- 북마크
- 구독
- 소거
- RSS 피드 구독
- 강조
- 인쇄
- 부적절한 컨텐트 신고
I have tried this in Intel 12, 13, 14 - the same result. I have a computation that crashes in release but it doesn't crash in debug and thus I cant find out why it crashes. Any ideas on what to do? Compiler options that can be tried? Thanks a lot in advance!!!!!
링크가 복사됨
- 신규로 표시
- 북마크
- 구독
- 소거
- RSS 피드 구독
- 강조
- 인쇄
- 부적절한 컨텐트 신고
Not a lot to go on here. What is the exact error you get? Typically this situation means you have a programming error that is masked by the lack of optimization. In release mode you should get a traceback indicating about where in the code the error occurred. You can study that to see if you can figure out what went wrong. You could try, in the debug configuration, changing the optimization level to 1 and then 2 to see if the error appears. You may also need to turn off run-time checks that are disabled in a Release configuration.
- 신규로 표시
- 북마크
- 구독
- 소거
- RSS 피드 구독
- 강조
- 인쇄
- 부적절한 컨텐트 신고
Well, what exactly did you do? /O2 is the default for a release configuration.
- 신규로 표시
- 북마크
- 구독
- 소거
- RSS 피드 구독
- 강조
- 인쇄
- 부적절한 컨텐트 신고
Used /O2 with Debug. The problem was that an array subroutine argument was sometimes a pointer, sometimes an allocatable array and sometimes an array from the stack.
BTW, can you guys comment on my performance benchmark computation using two processors, parallel computation
Intel 12: 136 seconds
Intel 13: 156 seconds
Intel 14: 160 seconds
Has anyone else found the same thing?
- 신규로 표시
- 북마크
- 구독
- 소거
- RSS 피드 구독
- 강조
- 인쇄
- 부적절한 컨텐트 신고
We regularly test performance of new versions and compare to previous versions. We don't find it acceptable for there to be a general slowdown, and for most applications, performance improves. We can't test every possible application though and there have been sporadic cases of specific applications slowing down. If you're willing to provide a test case, we'll be glad to investigate.
