Re: Scheduler questions (related to PR17808)

2005-07-05 Thread Steven Bosscher
On Tuesday 05 July 2005 20:30, Vladimir Makarov wrote: > Andrey Belevantsev wrote: > > Vladimir Makarov wrote: > >> I'll look at this PR today. > > > > We've looked today at this issue. We think the problem is that > > proposed patch of sched_get_condition() treats conditional jumps > > likely to C

Re: Scheduler questions (related to PR17808)

2005-07-05 Thread Vladimir Makarov
Andrey Belevantsev wrote: Vladimir Makarov wrote: I'll look at this PR today. We've looked today at this issue. We think the problem is that proposed patch of sched_get_condition() treats conditional jumps likely to COND_EXECs, but it doesn't fix other places in sched-deps, where COND_EX

Re: Scheduler questions (related to PR17808)

2005-06-30 Thread Steven Bosscher
On Friday 01 July 2005 01:05, Vladimir Makarov wrote: > Andrey Belevantsev wrote: > > Vladimir Makarov wrote: > >> I'll look at this PR today. > > > > We've looked today at this issue. We think the problem is that > > proposed patch of sched_get_condition() treats conditional jumps > > likely to CO

Re: Scheduler questions (related to PR17808)

2005-06-30 Thread Steven Bosscher
On Friday 01 July 2005 01:05, Vladimir Makarov wrote: > Andrey Belevantsev wrote: > > Vladimir Makarov wrote: > >> I'll look at this PR today. > > > > We've looked today at this issue. We think the problem is that > > proposed patch of sched_get_condition() treats conditional jumps > > likely to CO

Re: Scheduler questions (related to PR17808)

2005-06-30 Thread Vladimir Makarov
Andrey Belevantsev wrote: Vladimir Makarov wrote: I'll look at this PR today. We've looked today at this issue. We think the problem is that proposed patch of sched_get_condition() treats conditional jumps likely to COND_EXECs, but it doesn't fix other places in sched-deps, where COND_EX

Re: Scheduler questions (related to PR17808)

2005-06-30 Thread James E Wilson
Andrey Belevantsev wrote: We've also found that current mainline ICEs compiling the testcase with "-O0 -fschedule-insns -fschedule-insns2". I suspect this is a bug in ia64_reorg in ia64.c. It shouldn't be trying to schedule during a non-optimizing compile. So the line if (ia64_flag_schedu

Re: Scheduler questions (related to PR17808)

2005-06-30 Thread James E Wilson
Steven Bosscher wrote: Are predicate checks free, or should there be some post-pass to clean up this kind of useless predication? On IA-64, predicate checks are free in terms of run-time. There is the problem that unnecessary predicate checks might increase register lifetimes, causing extra

Re: Scheduler questions (related to PR17808)

2005-06-30 Thread Steven Bosscher
On Thursday 30 June 2005 16:18, Andrey Belevantsev wrote: > Vladimir Makarov wrote: > > I'll look at this PR today. > > We've looked today at this issue. We think the problem is that proposed > patch of sched_get_condition() treats conditional jumps likely to > COND_EXECs, but it doesn't fix other

Re: Scheduler questions (related to PR17808)

2005-06-30 Thread Steven Bosscher
On Thursday 30 June 2005 16:18, Andrey Belevantsev wrote: > We don't know reload well enough to know for sure which place should be > fixed in reload, or maybe in update_life_info(). Is this issue worth > opening another PR? This is a separate PR. Actually, I'm surprised that -O0 -fschedule-insns

Re: Scheduler questions (related to PR17808)

2005-06-30 Thread Steven Bosscher
On Thursday 30 June 2005 16:19, Mostafa Hagog wrote: > Hi, > > Steven Bosscher <[EMAIL PROTECTED]> wrote on 30/06/2005 01:46:22: > > Hi, > > [snip] > > > Then the ia64 machine-reorg scheduler gets to work, and it produces: > > > > (insn:TI 8 70 12 0 (set (reg:BI 262 p6 [353]) > > (ne:BI (re

Re: Scheduler questions (related to PR17808)

2005-06-30 Thread Mostafa Hagog
Hi, Steven Bosscher <[EMAIL PROTECTED]> wrote on 30/06/2005 01:46:22: > > Hi, > [snip] > Then the ia64 machine-reorg scheduler gets to work, and it produces: > > (insn:TI 8 70 12 0 (set (reg:BI 262 p6 [353]) > (ne:BI (reg/v:SI 15 r15 [orig:348 b1 ] [348]) > (const_int 0 [0

Re: Scheduler questions (related to PR17808)

2005-06-30 Thread Andrey Belevantsev
Vladimir Makarov wrote: I'll look at this PR today. We've looked today at this issue. We think the problem is that proposed patch of sched_get_condition() treats conditional jumps likely to COND_EXECs, but it doesn't fix other places in sched-deps, where COND_EXECs are considered. Maxim Kuvy

Re: Scheduler questions (related to PR17808)

2005-06-30 Thread Vladimir Makarov
Steven Bosscher wrote: Notice how the conditional sets of r14 and r17 in insns 9 and 10 have been moved past insn 14, which uses these registers. Shouldn't there be true dependencies on insns 9 and 10 for insn 14? In theory, yes. But the scheduler (it is ebb scheduler as I understand) fre

Re: Scheduler questions (related to PR17808)

2005-06-30 Thread Steven Bosscher
On Thursday 30 June 2005 12:53, Richard Earnshaw wrote: > On Thu, 2005-06-30 at 11:22, Steven Bosscher wrote: > > For your problem, I think the jump should just be forced somehow to > > be the last insn in the block. You can't move things into other > > basic blocks if you are not doing interblock

Re: Scheduler questions (related to PR17808)

2005-06-30 Thread Richard Earnshaw
On Thu, 2005-06-30 at 11:22, Steven Bosscher wrote: > For your problem, I think the jump should just be forced somehow to > be the last insn in the block. You can't move things into other > basic blocks if you are not doing interblock scheduling, after all. > For the PR I was looking at, someone w

Re: Scheduler questions (related to PR17808)

2005-06-30 Thread Steven Bosscher
On Thursday 30 June 2005 11:32, Richard Earnshaw wrote: > Hmm, certainly looks suspicious, but it's not really what the original > bug report was about. This might be some interblock scheduling problem > that is also triggered by the extra scheduling freedom we get with the > disabled change. Hm

Re: Scheduler questions (related to PR17808)

2005-06-30 Thread Richard Earnshaw
On Wed, 2005-06-29 at 23:46, Steven Bosscher wrote: > Hi, > > I have a question about the scheduler. Forgive me if I'm totally > missing the point here, this scheduling business is not my thing ;-) > > Consider the following snippet that I've derived from PR17808 with a > few hacks in the compil

Re: Scheduler questions (related to PR17808)

2005-06-29 Thread Dale Johannesen
On Jun 29, 2005, at 3:46 PM, Steven Bosscher wrote: I have a question about the scheduler. Forgive me if I'm totally missing the point here, this scheduling business is not my thing ;-) Consider the following snippet that I've derived from PR17808 with a few hacks in the compiler to renumber in

Scheduler questions (related to PR17808)

2005-06-29 Thread Steven Bosscher
Hi, I have a question about the scheduler. Forgive me if I'm totally missing the point here, this scheduling business is not my thing ;-) Consider the following snippet that I've derived from PR17808 with a few hacks in the compiler to renumber insns and dump RTL with all the dependencies befor