<div dir="ltr"><br><div class="gmail_extra"><br><div class="gmail_quote">On Wed, May 25, 2016 at 1:54 PM, Xinliang David Li <span dir="ltr"><<a href="mailto:davidxl@google.com" target="_blank">davidxl@google.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-style:solid;border-left-color:rgb(204,204,204);padding-left:1ex"><div dir="ltr">Please commit this one.</div></blockquote><div><br></div><div>
<p class=""><span class="">r270762.</span></p><p class=""><span class="">-- Sean Silva</span></p></div><div> </div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-style:solid;border-left-color:rgb(204,204,204);padding-left:1ex"><div dir="ltr"><div><br></div><div>Thanks,</div><div><br></div><div>David</div></div><div class=""><div class="h5"><div class="gmail_extra"><br><div class="gmail_quote">On Wed, May 25, 2016 at 1:45 PM, Sean Silva <span dir="ltr"><<a href="mailto:chisophugis@gmail.com" target="_blank">chisophugis@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-style:solid;border-left-color:rgb(204,204,204);padding-left:1ex"><div dir="ltr">Does something like this fix it?<div><br></div><div>-- Sean Silva<br><div class="gmail_extra"><br><div class="gmail_quote"><div><div>On Wed, May 25, 2016 at 1:39 PM, Vedant Kumar via llvm-commits <span dir="ltr"><<a href="mailto:llvm-commits@lists.llvm.org" target="_blank">llvm-commits@lists.llvm.org</a>></span> wrote:<br></div></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-style:solid;border-left-color:rgb(204,204,204);padding-left:1ex"><div><div><span><br>
> On May 25, 2016, at 1:20 PM, Xinliang David Li <<a href="mailto:davidxl@google.com" target="_blank">davidxl@google.com</a>> wrote:<br>
><br>
><br>
> On Wed, May 25, 2016 at 1:17 PM, Steven Wu <<a href="mailto:stevenwu@apple.com" target="_blank">stevenwu@apple.com</a>> wrote:<br>
> It is -O3 PGO for arm64 with training input.<br>
><br>
> I am looking as well. Seems PData->Values is NULL but I don't know how yet.<br>
<br>
</span>Steven sent me a core dump. I'll take a look at this to double-check the value of PData.<br>
<span><br>
><br>
> With static allocation, this should not happen -- the field is statically initialized.<br>
<br>
</span>But, we know static allocation is enabled because we get a warning from allocateOneNode().<br>
<br>
If enough warnings occur, is it possible for CurrentVNode to wrap around s.t:<br>
<br>
COMPILER_RT_PTR_FETCH_ADD(ValueProfNode, CurrentVNode, 1) < EndVNode ?<br>
<br>
If this happens, allocateOneNode() can eventually return an incorrect VNode which will then be stored into PData->Values.<br>
<span><font color="#888888"><br>
vedant<br>
</font></span><div><div><br>
><br>
> David<br>
><br>
><br>
><br>
> Steven<br>
><br>
>> On May 25, 2016, at 1:06 PM, Xinliang David Li <<a href="mailto:davidxl@google.com" target="_blank">davidxl@google.com</a>> wrote:<br>
>><br>
>> I could not reproduce the issue (only saw out of nodes warning). What additional option is used (I used O3)? What is the input (train, ref, test)?<br>
>><br>
>> David<br>
>><br>
>> On Wed, May 25, 2016 at 11:24 AM, Vedant Kumar <<a href="mailto:vsk@apple.com" target="_blank">vsk@apple.com</a>> wrote:<br>
>> Hi David,<br>
>><br>
>> We're seeing another issue that we think is related to the recent static VP node allocation changes. Instrumented programs are hitting a segfault:<br>
>><br>
>> ```<br>
>> 458.sjeng is fixed but 445.gobmk is broken.<br>
>> LLVM Profile Warning: Unable to track new values: Running out of static counters. Consider using option -mllvm -vp-counters-per-site=<n> to allocate more value profile counters at compile time.<br>
>> /Users/buildslave/jenkins/workspace/Performance_ARM64_SPEC2006_INT-O3_LTO_PGO-master/spec2006/cur_run/nt/build/LNTBased/speccpu2006/int/445.gobmk/tools/timeit-target: error: child terminated by signal 11<br>
>> ```<br>
>><br>
>> The backtrace is:<br>
>><br>
>> ```<br>
>> * frame #0: 0x0000000100112f10 445.gobmk.simple`__llvm_profile_instrument_target + 132 at InstrProfilingValue.c:137<br>
>> frame #1: 0x000000010007d528 445.gobmk.simple`shapes_callback + 2352<br>
>> frame #2: 0x0000000100035b0c 445.gobmk.simple`matchpat_loop + 2088<br>
>> frame #3: 0x0000000100034b38 445.gobmk.simple`matchpat_goal_anchor + 1124<br>
>> frame #4: 0x000000010007cb0c 445.gobmk.simple`shapes + 384<br>
>> frame #5: 0x0000000100029e4c 445.gobmk.simple`do_genmove + 2344<br>
>> frame #6: 0x00000001000a5a48 445.gobmk.simple`gtp_gg_genmove + 216<br>
>> frame #7: 0x0000000100099a6c 445.gobmk.simple`gtp_main_loop + 660<br>
>> frame #8: 0x000000010009bd00 445.gobmk.simple`main + 7188<br>
>> ```<br>
>><br>
>> In __llvm_profile_instrument_target(), it looks like we're failing the condition: CounterIndex < NumVSites.<br>
>><br>
>> Do you mind taking a look?<br>
>><br>
>> thanks,<br>
>> vedant<br>
>><br>
>><br>
>> > On May 23, 2016, at 12:29 PM, Xinliang David Li <<a href="mailto:davidxl@google.com" target="_blank">davidxl@google.com</a>> wrote:<br>
>> ><br>
>> > Fix is on the way.<br>
>> ><br>
>> > David<br>
>> ><br>
>> > On Mon, May 23, 2016 at 12:27 PM, Vedant Kumar <<a href="mailto:vsk@apple.com" target="_blank">vsk@apple.com</a>> wrote:<br>
>> ><br>
>> > > On May 23, 2016, at 12:19 PM, Xinliang David Li <<a href="mailto:davidxl@google.com" target="_blank">davidxl@google.com</a>> wrote:<br>
>> > ><br>
>> > ><br>
>> > ><br>
>> > > On Mon, May 23, 2016 at 12:15 PM, Vedant Kumar <<a href="mailto:vsk@apple.com" target="_blank">vsk@apple.com</a>> wrote:<br>
>> > > Hi David,<br>
>> > ><br>
>> > > I think one of the SPEC2006 tests doesn't have enough statically-allocated VP nodes per site. We're seeing:<br>
>> > ><br>
>> > > > "Child terminated by signal 25" (SIGXFSZ) after:<br>
>> > ><br>
>> > > Is this related?<br>
>> ><br>
>> > I don't think so. I suspect that's happening because the device is overloaded with logging info.<br>
>> ><br>
>> ><br>
>> > > ><br>
>> > > > LLVM Profile Warning: Running out of nodes: site_0@func_12822962448227433604, value=4295054468<br>
>> > > > ...<br>
>> > > > LLVM Profile Warning: Running out of nodes: site_0@func_12822962448227433604, value=4295052980<br>
>> > > > LLVM Profile Warning: Running out of nodes: site_0@func_12822962/Users/buildslave/jenkins/workspace/Performance_ARM64_SPEC2006_INT-O3_LTO_PGO-master/spec2006/cur_run/nt/build/LNTBased/speccpu2006/int/458.sjeng<br>
>> > ><br>
>> > > It seems like the fix for now is to either tweak vp-counters-per-site for the test or to set -vp-static-alloc=false.<br>
>> > ><br>
>> > > In the long term, do you think it's worth adjusting vp-counters-per-site s.t we can run SPEC without modifications? If so, is SPEC the right testbed?<br>
>> > ><br>
>> > > Rong had helped collecting SPEC related statistics, but looks like there is something missing. I will investigate. In the meantime, can you try the workaround?<br>
>> ><br>
>> > Yes, we'll do that.<br>
>> ><br>
>> > Here's all the logging info from our bot:<br>
>> ><br>
>> ><br>
>> ><br>
>> ><br>
>> ><br>
>> > thanks,<br>
>> > vedant<br>
>> ><br>
>> ><br>
>> > ><br>
>> > > thanks,<br>
>> > ><br>
>> > > David<br>
>> > ><br>
>> > ><br>
>> > > thanks,<br>
>> > > vedant<br>
>> > ><br>
>> ><br>
>> ><br>
>> ><br>
>><br>
>><br>
><br>
><br>
<br>
</div></div></div></div><div><div>_______________________________________________<br>
llvm-commits mailing list<br>
<a href="mailto:llvm-commits@lists.llvm.org" target="_blank">llvm-commits@lists.llvm.org</a><br>
<a href="http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits" rel="noreferrer" target="_blank">http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits</a><br>
</div></div></blockquote></div><br></div></div></div>
</blockquote></div><br></div>
</div></div></blockquote></div><br></div></div>