# Code slower than C++ version **URL:** \ **Category:** Help\ **Created:** [October 13, 2021, 8:50pm UTC](https://fortran-lang.discourse.group/t/code-slower-than-c-version/2066 "2021-10-13T20:50:52Z")\ **Posts on this page:** 6\ **Page:** 2
**Author:** ![lmiq](https://yyz2.discourse-cdn.com/free1/user_avatar/fortran-lang.discourse.group/lmiq/32/555_2.png) [@lmiq](https://fortran-lang.discourse.group/u/lmiq)\ **Post date:** [October 14, 2021, 2:56pm UTC](https://fortran-lang.discourse.group/t/code-slower-than-c-version/2066/21 "2021-10-14T14:56:59Z")
> [@certik](#): > > and handle nans. This is probably deliberate to keep a safer option as the default. What I miss there is that `@fastmath` (or the compiler option in Fortran) should skip those checks, shouldn’t it? ~~(it does not)~~ ---
**Author:** ![oscardssmith](https://avatars.discourse-cdn.com/v4/letter/o/b9e5f3/32.png) [@oscardssmith](https://fortran-lang.discourse.group/u/oscardssmith)\ **Post date:** [October 14, 2021, 3:05pm UTC](https://fortran-lang.discourse.group/t/code-slower-than-c-version/2066/22 "2021-10-14T15:05:01Z")
`three_median2` and `three_median1` (and associated assembly) is actually an incorrect implementation if you are following strict ieee rules. `three_median2(0.0,-0.0,0.0)==-0.0` which and wrong, and both `three_median1` and `three_median2` produce `NaN` for `three_median2(0.0,NaN,0.0)==NaN`. If you define `three_median1` as `@fastmath max(min(a,b),min(max(a,b),c))`, Julia is able to optimize to the same assembly as `three_median2` ---
**Author:** ![certik](https://yyz2.discourse-cdn.com/free1/user_avatar/fortran-lang.discourse.group/certik/32/4_2.png) [@certik](https://fortran-lang.discourse.group/u/certik)\ **Post date:** [October 14, 2021, 3:25pm UTC](https://fortran-lang.discourse.group/t/code-slower-than-c-version/2066/23 "2021-10-14T15:25:25Z")
> [@oscardssmith](#): > > If you define `three_median1` as `@fastmath max(min(a,b),min(max(a,b),c))` , Julia is able to optimize to the same assembly as `three_median2` Perfect. Thanks for the feedback @oscardssmith. @lmiq, do you want to try it if you now get the same performance with `three_median1`? ---
**Author:** ![lmiq](https://yyz2.discourse-cdn.com/free1/user_avatar/fortran-lang.discourse.group/lmiq/32/555_2.png) [@lmiq](https://fortran-lang.discourse.group/u/lmiq)\ **Post date:** [October 14, 2021, 4:11pm UTC](https://fortran-lang.discourse.group/t/code-slower-than-c-version/2066/24 "2021-10-14T16:11:10Z")
Yes, it does: ```auto julia> function three_median1(a, b, c) res = @fastmath max(min(a,b),min(max(a,b),c)) return res end three_median1 (generic function with 1 method) julia> @btime test($x,$three_median1) 10.053 μs (0 allocations: 0 bytes) 5017.136463388665 ``` (I think I had put `@fastmath` on the call to the function on the `test` function, and that didn’t work) one curiosity, in this last version I though I would do one comparison less: ```auto julia> function three_median3(a,b,c) a1, a2 = a < b ? (a, b) : (b, a) a3 = a2 < c ? a2 : c res = a3 > a1 ? a3 : a1 return res end three_median3 (generic function with 1 method) ``` but the native code is the same: ```auto julia> @code_native three_median3(1.0,1.0,1.0) .text ; ┌ @ REPL[32] within `three_median3' vminsd %xmm1, %xmm0, %xmm3 vmaxsd %xmm0, %xmm1, %xmm0 ; │ @ REPL[32]:3 within `three_median3' vminsd %xmm2, %xmm0, %xmm0 ; │ @ REPL[32] within `three_median3' vmaxsd %xmm3, %xmm0, %xmm0 ; │ @ REPL[32]:5 within `three_median3' retq nopw %cs:(%rax,%rax) ; └ ``` (the compiler decided that swapping the values is not worth saving one comparison, something like that). But, at the end, Fortran is being able to optimize the `min/max` version to optimal, isn’t it? Does it behave like Julia in this regard (having NaN checks without `fastmath` and skipping those with it)? (I don’t know how to check the assembly codes in these cases). ---
**Author:** ![ivanpribec](https://yyz2.discourse-cdn.com/free1/user_avatar/fortran-lang.discourse.group/ivanpribec/32/3290_2.png) [@ivanpribec](https://fortran-lang.discourse.group/u/ivanpribec)\ **Post date:** [October 14, 2021, 4:54pm UTC](https://fortran-lang.discourse.group/t/code-slower-than-c-version/2066/25 "2021-10-14T16:54:18Z")
> [@lmiq](#): > > But, at the end, Fortran is being able to optimize the `min/max` version to optimal, isn’t it? Does it behave like Julia in this regard (having NaN checks without `fastmath` and skipping those with it)? > (I don’t know how to check the assembly codes in these cases). You can check them with [Godbolt](https://godbolt.org/#g:!((g:!((g:!((h:codeEditor,i:(filename:'1',fontScale:14,fontUsePx:'0',j:1,lang:fortran,selection:(endColumn:21,endLineNumber:2,positionColumn:21,positionLineNumber:2,selectionStartColumn:21,selectionStartLineNumber:2,startColumn:21,startLineNumber:2),source:'function+three_median(x,+y,+z)%0A++real(8),+intent(in)+::+x,+y,+z%0A++real(8)+::+three_median%0A%0A++three_median+%3D+max(min(x,+y),min(max(x,+y),+z))%0A++%0Aend+function%0A%0Asubroutine+med_rv(n,+x,+res)%0A++integer(8)+::+n%0A++real(8)+::+x(n),+res++++%0A++real(8)+::+param1,+param2,+param3,+param4,+pi%0A++integer(8)+::+i%0A++real(8)+::+three_median%0A%0A++param1+%3D+2d0%0A++param2+%3D+3d0%0A++param3+%3D+4d0+%0A++param4+%3D+6d0%0A++pi+%3D+acos(-1d0)%0A%0A++!!omp+parallel+do+reduction(%2B:res)%0A++do+i+%3D+1,+n+-+2%0A++++res+%3D+res+%2B+three_median(abs(x(i)),+abs(x(i+%2B+1)),+abs(x(i+%2B+2)))**param1++%0A++end+do%0A++!!omp+end+parallel+do+%0A%0A++res+%3D+res*pi/(param4+-+param3*sqrt(param2)+%2B+pi)*n/(n+-+param1)%0Aend+subroutine'),l:'5',n:'0',o:'Fortran+source+%231',t:'0')),k:50,l:'4',n:'0',o:'',s:0,t:'0'),(g:!((g:!((h:compiler,i:(compiler:ifort202130,filters:(b:'0',binary:'1',commentOnly:'0',demangle:'0',directives:'0',execute:'1',intel:'0',libraryCode:'0',trim:'1'),flagsViewOpen:'1',fontScale:14,fontUsePx:'0',j:1,lang:fortran,libs:!(),options:'-O3++-qopt-report-phase%3Dvec+-qopt-report-file%3Dstderr',selection:(endColumn:16,endLineNumber:9,positionColumn:1,positionLineNumber:1,selectionStartColumn:16,selectionStartLineNumber:9,startColumn:1,startLineNumber:1),source:1,tree:'1'),l:'5',n:'0',o:'x86-64+ifort+2021.3.0+(Fortran,+Editor+%231,+Compiler+%231)',t:'0')),k:50,l:'4',m:82.20973782771536,n:'0',o:'',s:0,t:'0'),(g:!((h:output,i:(compiler:1,editor:1,fontScale:14,fontUsePx:'0',tree:'1',wrap:'1'),l:'5',n:'0',o:'Output+of+x86-64+ifort+2021.3.0+(Compiler+%231)',t:'0')),header:(),l:'4',m:17.790262172284642,n:'0',o:'',s:0,t:'0')),k:50,l:'3',n:'0',o:'',t:'0')),l:'2',n:'0',o:'',t:'0')),version:4): Here is the latest Intel Fortran compiler Classic with `-O3`: ```plaintext three_median_: movsd xmm2, QWORD PTR [rdi] #5.3 movsd xmm1, QWORD PTR [rsi] #5.3 movaps xmm0, xmm2 #7.1 maxsd xmm2, xmm1 #7.1 minsd xmm0, xmm1 #7.1 minsd xmm2, QWORD PTR [rdx] #7.1 maxsd xmm0, xmm2 #7.1 ret ``` And here is the `gfortran` generated assembly from the link in @lkedward’s reply: ```plaintext three_median_: movsd xmm1, QWORD PTR [rdi] movsd xmm2, QWORD PTR [rsi] movapd xmm0, xmm1 minsd xmm1, xmm2 maxsd xmm0, xmm2 minsd xmm0, QWORD PTR [rdx] maxsd xmm0, xmm1 ret ``` ---
**Author:** ![lmiq](https://yyz2.discourse-cdn.com/free1/user_avatar/fortran-lang.discourse.group/lmiq/32/555_2.png) [@lmiq](https://fortran-lang.discourse.group/u/lmiq)\ **Post date:** [October 14, 2021, 5:37pm UTC](https://fortran-lang.discourse.group/t/code-slower-than-c-version/2066/26 "2021-10-14T17:37:38Z")
Nice tool. The version with the conditionals only seems to generate less instructions: [Compiler Explorer](https://godbolt.org/z/TP8ePE7Yx) With conditionals only probably there is a small performance gain, although as far as I understood from the documentation the `max/min` function is dealing with NaNs (I was expecting that `--fast-math` or some other flag made the assemblies converge, but I couldn’t find that flag, if it exists). [Previous page](https://fortran-lang.discourse.group/t/code-slower-than-c-version/2066.md?page=1)