http://ave1.org/2018/gnat-and-pragma-inline_always-a-report/ Tue, 06 Oct 2020 17:03:02 +0000 http://polimedia.us hourly 1 http://ave1.org/2018/gnat-and-pragma-inline_always-a-report/#comment-86 Loper OS » “Finite Field Arithmetic.” Chapter 12A: Karatsuba Redux. (Part 1 of 2) Sat, 17 Nov 2018 23:33:06 +0000 http://ave1.org/?p=45#comment-86 [...] ave1 carefully analyzed the executables generated by AdaCore’s x86-64 GNAT, and showed that this compiler is prone to ignore Inline pragmas unless they are specified in a particular [...] [...] ave1 carefully analyzed the executables generated by AdaCore’s x86-64 GNAT, and showed that this compiler is prone to ignore Inline pragmas unless they are specified in a particular [...]

]]> http://ave1.org/2018/gnat-and-pragma-inline_always-a-report/#comment-80 Loper OS » “Finite Field Arithmetic.” Chapter 11: Tuning and Unified API. Fri, 09 Nov 2018 18:47:48 +0000 http://ave1.org/?p=45#comment-80 [...] ave1 carefully analyzed the executables generated by AdaCore’s x86-64 GNAT, and showed that this compiler is prone to ignore Inline pragmas unless they are specified in a particular [...] [...] ave1 carefully analyzed the executables generated by AdaCore’s x86-64 GNAT, and showed that this compiler is prone to ignore Inline pragmas unless they are specified in a particular [...]

]]> http://ave1.org/2018/gnat-and-pragma-inline_always-a-report/#comment-20 ave1 Tue, 06 Feb 2018 09:43:45 +0000 http://ave1.org/?p=45#comment-20 Which calls are you looking at? for methods in the <i>same</i> module the inlining pragma works. For methods outside of the module it seems to not work. Which calls are you looking at? for methods in the same module the inlining pragma works. For methods outside of the module it seems to not work.

]]> http://ave1.org/2018/gnat-and-pragma-inline_always-a-report/#comment-19 Stanislav Datskovskiy Mon, 05 Feb 2018 14:32:20 +0000 http://ave1.org/?p=45#comment-19 FTR I see exactly the same effect on the dump when toggling the pragmas in their current (i.e. in the .adb bodies) locations. (This is on gnat 2016). FTR I see exactly the same effect on the dump when toggling the pragmas in their current (i.e. in the .adb bodies) locations. (This is on gnat 2016).

]]> http://ave1.org/2018/gnat-and-pragma-inline_always-a-report/#comment-18 Stanislav Datskovskiy Mon, 05 Feb 2018 14:29:26 +0000 http://ave1.org/?p=45#comment-18 How then to explain the massive timing variance when my "incorrect" inline pragmas are selectively removed? How then to explain the massive timing variance when my "incorrect" inline pragmas are selectively removed?

]]>