Re: Killing old dead bugs

2017-07-17 Thread Jeff Law
On 07/17/2017 10:33 PM, NightStrike wrote: > On Mon, Jul 17, 2017 at 11:23 AM, Martin Sebor wrote: >> you did for the bugs below is ideal. Adding a test case if one >> doesn't exist in the test suite is also very useful, though quite >> a bit more work. > > Isn't a testcase always required? Requ

Re: Killing old dead bugs

2017-07-17 Thread NightStrike
On Mon, Jul 17, 2017 at 11:23 AM, Martin Sebor wrote: > you did for the bugs below is ideal. Adding a test case if one > doesn't exist in the test suite is also very useful, though quite > a bit more work. Isn't a testcase always required?

Re: [patch] RFC: Hook for insn costs?

2017-07-17 Thread Richard Henderson
On 07/17/2017 12:20 AM, Richard Biener wrote: On Sun, Jul 16, 2017 at 12:51 AM, Segher Boessenkool Now what should it take as input? An rtx_insn, or just the pattern (as insn_rtx_cost does)? Is there any useful info on the other operands of an rtx_insn? If not then passing in the pattern (a

Re: Killing old dead bugs

2017-07-17 Thread Yuri Gribov
On Mon, Jul 17, 2017 at 4:23 PM, Martin Sebor wrote: > On 07/17/2017 02:14 AM, Yuri Gribov wrote: >> >> Hi Mikhail, >> >> On Sun, Jul 2, 2017 at 6:39 PM, Mikhail Maltsev >> wrote: >>> >>> Hi. Yes, bug maintenance is appreciated. See this message and replies >>> to it: https://gcc.gnu.org/ml/gcc/2

Re: Complex multiplication in gcc

2017-07-17 Thread Sean McAllister
Ah OK, thank you, I wasn't aware of that particular mechanism. If I run the code and break on __mulsc3 it disassembles as I'd expect. On Mon, Jul 17, 2017 at 12:32 PM, Gabriel Paubert wrote: > On Mon, Jul 17, 2017 at 10:51:21AM -0600, Sean McAllister wrote: >> When generating code for a simple i

Re: Complex multiplication in gcc

2017-07-17 Thread Gabriel Paubert
On Mon, Jul 17, 2017 at 10:51:21AM -0600, Sean McAllister wrote: > When generating code for a simple inner loop (instantiated with > std::complex) > > template > void __attribute__((noinline)) benchcore(const cx* __restrict__ aa, > const cx* __restrict__ bb, const cx* __restrict__ cc, cx* __restr

Complex multiplication in gcc

2017-07-17 Thread Sean McAllister
When generating code for a simple inner loop (instantiated with std::complex) template void __attribute__((noinline)) benchcore(const cx* __restrict__ aa, const cx* __restrict__ bb, const cx* __restrict__ cc, cx* __restrict__ dd, cx uu, cx vv, size_t nn) { for (ssize_t ii=0; ii < nn; ii++) {

Re: Killing old dead bugs

2017-07-17 Thread Martin Sebor
On 07/17/2017 02:14 AM, Yuri Gribov wrote: Hi Mikhail, On Sun, Jul 2, 2017 at 6:39 PM, Mikhail Maltsev wrote: Hi. Yes, bug maintenance is appreciated. See this message and replies to it: https://gcc.gnu.org/ml/gcc/2016-04/msg00258.html . Replies in your link suggest to leave a final comment

Re: [patch] RFC: Hook for insn costs?

2017-07-17 Thread Segher Boessenkool
Him On Mon, Jul 17, 2017 at 03:36:13PM +0200, Georg-Johann Lay wrote: > >On Wed, Jul 12, 2017 at 03:15:09PM +0200, Georg-Johann Lay wrote: > >>the current cost computations in rtlanal.c and maybe other places > >>suffer from the fact that they are hiding parts of the expressions > >>from the back-

Re: Could preprocessor warn for unsafe macros and side-effects?

2017-07-17 Thread Eric Gallager
On 7/12/17, Martin Sebor wrote: > On 07/11/2017 11:50 PM, sa...@hederstierna.com wrote: >> Hi >> >> Reading about macro pitfalls and eg duplication side-effects >> https://gcc.gnu.org/onlinedocs/cpp/Macro-Pitfalls.html#Macro-Pitfalls >> >> would it be possible to let the preprocessor generate warn

Re: [patch] RFC: Hook for insn costs?

2017-07-17 Thread Georg-Johann Lay
On 16.07.2017 00:51, Segher Boessenkool wrote: Hi! On Wed, Jul 12, 2017 at 03:15:09PM +0200, Georg-Johann Lay wrote: the current cost computations in rtlanal.c and maybe other places suffer from the fact that they are hiding parts of the expressions from the back-end, like SET_DESTs of single_s

Re: expmed costs and i386.c cost for widening mul (PR81444)

2017-07-17 Thread Georg-Johann Lay
On 17.07.2017 10:53, Georg-Johann Lay wrote: Hi, while testing a patch to fix PR81444, I came across a new FAIL due to the patch in i386.c/pr71321.c PR81444 is about wrong modes used by expmed.c as it computes costs for widening operations like widening mul. It uses GET_MODE_WIDER_MODE for the

Re: Loop reversal

2017-07-17 Thread Segher Boessenkool
On Mon, Jul 17, 2017 at 11:56:50AM +0200, Richard Biener wrote: > On Thu, Jul 13, 2017 at 2:18 AM, Kugan Vivekanandarajah > wrote: > > I am looking into reversing loop to increased efficiency. There is > > already a PR22041 for this and an old patch > > https://gcc.gnu.org/ml/gcc-patches/2006-01/m

Re: whereis PLUGIN_REGISTER_GGC_CACHES? how to migrate it for GCC v6.x?

2017-07-17 Thread Trevor Saunders
On Wed, Jul 12, 2017 at 10:12:03AM +0800, Leslie Zhai wrote: > PS: Trevor's email is not available? thanks! Sorry about that, I've left Mozilla and been vacationing for a month, so didn't get to updating MAINTAINERS yet. Here's a patch doing that. Trev commit ff900f40d23f765fd59047a90a7e3ff18cb

Re: [patch] RFC: Hook for insn costs?

2017-07-17 Thread Segher Boessenkool
On Mon, Jul 17, 2017 at 12:20:16PM +0200, Richard Biener wrote: > > So I argue that we want to have a hook for insn cost. > > And I think at previous GNU Cauldrons we agreed to that. Excellent! So now it just needs to happen ;-) > > Now what should it take as input? An rtx_insn, or just the pa

Re: [patch] RFC: Hook for insn costs?

2017-07-17 Thread Richard Biener
On Sun, Jul 16, 2017 at 12:51 AM, Segher Boessenkool wrote: > Hi! > > On Wed, Jul 12, 2017 at 03:15:09PM +0200, Georg-Johann Lay wrote: >> the current cost computations in rtlanal.c and maybe other places >> suffer from the fact that they are hiding parts of the expressions >> from the back-end, l

Re: Loop reversal

2017-07-17 Thread Richard Biener
On Thu, Jul 13, 2017 at 2:18 AM, Kugan Vivekanandarajah wrote: > I am looking into reversing loop to increased efficiency. There is > already a PR22041 for this and an old patch > https://gcc.gnu.org/ml/gcc-patches/2006-01/msg01851.html by Zdenek > which never made it to mainline. > > For constant

Re: Linux and Windows generate different binaries

2017-07-17 Thread Segher Boessenkool
On Mon, Jul 17, 2017 at 11:10:01AM +0300, Alexander Monakov wrote: > On Sun, 16 Jul 2017, Segher Boessenkool wrote: > > > How? There's no stable sort in libc and switching over to std::stable_sort > > > would be problematic. > > > > Why? > > - you'd need to decide if the build time cost of extra

expmed costs and i386.c cost for widening mul (PR81444)

2017-07-17 Thread Georg-Johann Lay
Hi, while testing a patch to fix PR81444, I came across a new FAIL due to the patch in i386.c/pr71321.c PR81444 is about wrong modes used by expmed.c as it computes costs for widening operations like widening mul. It uses GET_MODE_WIDER_MODE for the wider mode where is should use GET_MODE_2XWID

Re: Killing old dead bugs

2017-07-17 Thread Yuri Gribov
Hi Mikhail, On Sun, Jul 2, 2017 at 6:39 PM, Mikhail Maltsev wrote: > Hi. Yes, bug maintenance is appreciated. See this message and replies > to it: https://gcc.gnu.org/ml/gcc/2016-04/msg00258.html . Replies in your link suggest to leave a final comment in bugs with explanatory suggestion to clos

Re: Linux and Windows generate different binaries

2017-07-17 Thread Alexander Monakov
On Sun, 16 Jul 2017, Segher Boessenkool wrote: > > How? There's no stable sort in libc and switching over to std::stable_sort > > would be problematic. > > Why? - you'd need to decide if the build time cost of extra 8000+ lines lines brought in by (per each TU) is acceptable; - you'd need to