The AST looked very bad: see https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67754#c2 There were a lot of testcases failing with MAX fusion.
Sebastian -----Original Message----- From: Tobias Grosser [mailto:tob...@grosser.es] Sent: Tuesday, September 29, 2015 5:14 PM To: Sebastian Pop; gcc-patches@gcc.gnu.org Cc: seb...@gmail.com; aditya...@samsung.com; richard.guent...@gmail.com Subject: Re: [PATCH] use MIN fusion for ISL-14 On 09/30/2015 12:10 AM, Sebastian Pop wrote: > This patch fixes PR66754 by reverting an earlier unintended change. > We now generate a much simpler AST for interchange-1.c: > > ISL AST generated by ISL: > { > for (int c1 = 0; c1 <= 1334; c1 += 1) { > S_7(c1); > for (int c3 = 0; c3 <= 1334; c3 += 1) > S_4(c1, c3); > S_5(c1); > } > for (int c1 = 0; c1 <= 1334; c1 += 1) > S_10(c1); > S_8(); > } Sure. Out of interest. How did the AST look before? Also, does it make sense to add a test case? Best, Tobias
