[llvm-dev] Aliasing rules difference between GCC and Clang (original) (raw)

Jonas Paulsson via llvm-dev llvm-dev at lists.llvm.org
Fri Sep 21 00:32:00 PDT 2018

Hi Matthieu,

Oh indeed, sorry, you are doing an outer product on a and b, storing the result in c, and these cannot alias (or they should not, if you do make they alias, it's your responsibility and I think you can get UB). So clang should do better indeed. It would be very nice if you could motivate this in detail with references to the standard. Is it enough that the two struct types have different tags? It might possibly be argued that the matrix contains the vector struct type, and therefore they may alias.  Why do you think this is not the case in this example?

/Jonas

Regards,

Le ven. 21 sept. 2018 à 08:21, Jonas Paulsson <paulsson at linux.vnet.ibm.com <mailto:paulsson at linux.vnet.ibm.com>> a écrit :

I would say that GCC is wrong and should also have a version where a could be equal to b. There is no restrict keyword, so they could be equal. This was between a/b and c, not between a and b. Could you explain your opinion a bit more in detail, please? /Jonas Cheers, Matthieu Le jeu. 20 sept. 2018 à 16:56, Jonas Paulsson via llvm-dev <llvm-dev at lists.llvm.org <mailto:llvm-dev at lists.llvm.org>> a écrit : Hi, I found a difference between Clang and GCC in alias handling. This was with a benchmark where Clang was considerably slower, and in a hot function which does many more loads from the same address due to stores between the uses. In other words, a value is loaded and used, another value is stored, and then the first value is loaded once again before its second use. This happens many times, with three loads instead of one for each value. GCC only emits one load. The values are the arguments to this function: void su3projector( su3vector *a, su3vector *b, su3matrix *c ){ register int i,j; register double tmp,tmp2; for(i=0;i<3;i++)for(j=0;j<3;j++){_ _tmp2 = a->c[i].real * b->c[j].real; tmp = a->c[i].imag * b->c[j].imag; c->e[i][j].real = tmp + tmp2; tmp2 = a->c[i].real * b->c[j].imag; tmp = a->c[i].imag * b->c[j].real; c->e[i][j].imag = tmp - tmp2; } } The types are: typedef struct { complex e[3][3]; } su3matrix; typedef struct { complex c[3]; } su3vector; So the question here is if the su3vector and su3matrix pointers may alias? If they may alias, then clang is right in reloading after each store. If the standard says they cannot alias, then gcc is right in only loading the values once each. It seems to me that either GCC is too aggressive or LLVM is too conservative, but I don't know which one it is... As far as I understand, there is the fact of the different struct types of the arguments (which means they cannot alias), but also the question if su3vector is included in su3matrix (which would mean they may alias). I made a reduced test case, where the same difference seems to be present. It has just one struct type which contains a matrix of double:s. A store to an element of the struct via a pointer is surrounded with two loads of a global double variable. Only Clang emits two loads. typedef struct { double c[3][3]; } STRUCTTY; double e = 0.0; STRUCTTY *f; int g = 0; void h() { int i = e; f->c[0][i] = g; g = e; } clang -O3-march=z13 : h:                                      # @h # %bb.0:                                # %entry larl    %r1, e ld      %f0, 0(%r1)        // LOAD E lrl     %r2, g cfdbr   %r0, 5, %f0        // CONVERT E lgfr    %r0, %r0           // EXTEND E cdfbr   %f0, %r2 lgrl    %r2, f sllg    %r3, %r0, 3 std     %f0, 0(%r3,%r2)    // STORE F ELEMENT ld      %f0, 0(%r1)        // 2nd LOAD E <<<<<<< cfdbr   %r0, 5, %f0        // CONVERT strl    %r0, g             // 2nd USE br      %r14 gcc -O3-march=z13 : h: .LFB0: .cfistartproc larl    %r1,e ld      %f0,0(%r1)// LOAD E lrl     %r2,g lgrl    %r3,f cfdbr   %r1,5,%f0// CONVERT E cdfbr   %f0,%r2 lgfr    %r2,%r1           // EXTEND E sllg    %r2,%r2,3 std     %f0,0(%r2,%r3)    // STORE F ELEMENT strl    %r1,g             // 2nd USE br      %r14 I hope somebody with enough experience and knowledge can guide the way here as this seems to be quite important. /Jonas

LLVM Developers mailing list llvm-dev at lists.llvm.org <mailto:llvm-dev at lists.llvm.org> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev -- Quantitative analyst, Ph.D. Blog: http://blog.audio-tk.com/ LinkedIn: http://www.linkedin.com/in/matthieubrucher -- Quantitative analyst, Ph.D. Blog: http://blog.audio-tk.com/ LinkedIn: http://www.linkedin.com/in/matthieubrucher

-------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20180921/d73e2d2a/attachment-0001.html>

More information about the llvm-dev mailing list