K-ballo has quit [Quit: K-ballo]
jaafar has joined #ste||ar
jaafar has quit [Ping timeout: 276 seconds]
weilewei has joined #ste||ar
weilewei has quit [Remote host closed the connection]
diehlpk has joined #ste||ar
<
diehlpk>
The Sc19 paper is published
<
diehlpk>
So update your CV
<
hkaiser>
diehlpk: let's update our publications pages as well
<
hkaiser>
remind me tomorrow, pls
<
diehlpk>
hkaiser, sure
<
diehlpk>
I already updated the preprint with the DOI
<
diehlpk>
it will be updated tomorrow
diehlpk has quit [Ping timeout: 246 seconds]
hkaiser has quit [Quit: bye]
jaafar has joined #ste||ar
jaafar has quit [Ping timeout: 240 seconds]
jaafar has joined #ste||ar
jaafar has quit [Ping timeout: 268 seconds]
rori has joined #ste||ar
jbjnr has quit [Read error: Connection reset by peer]
simbergm has quit [Write error: Connection reset by peer]
K-ballo has joined #ste||ar
simbergm has joined #ste||ar
hkaiser has joined #ste||ar
weilewei has joined #ste||ar
hkaiser has quit [Quit: bye]
hkaiser has joined #ste||ar
jaafar has joined #ste||ar
rori has quit [Ping timeout: 245 seconds]
<
jaafar>
K-ballo: can you tell me how the bloat reports are generated?
<
K-ballo>
jaafar: "the bloat reports" sounds so fancy :)
<
K-ballo>
I have a bunch of tools that I run, for the evaluate_assert/lambda report I only got to run one of them, Sizer
* K-ballo
will run the rest of the reports if he finds some extra free time before the release
* jaafar
is interested in tools like this in general
<
jaafar>
There's a slow-to-build OSS tool I'm curious about
<
jaafar>
lots of template instantiations thanks to cartesian product of types
<
K-ballo>
two other tools are linked from sizer
<
K-ballo>
and there's gcc -ftime-report, clang -ftime-trace, and msvc 's
<
K-ballo>
msvc's /d1ReportTime and /d2cgsummary
<
jaafar>
I've been using -ftime-trace, -ftime-report doesn't seem to give useful user-level info
<
jaafar>
I may need to analyze the -ftime-trace output with something other than the chrome tracer view
<
K-ballo>
-ftime-trace is very new.. msvc ones too.. I've been working with just -ftime-report for years
<
jaafar>
there's far too much detail, it just looks like "code generation... 80%" made up of countless tiny things
<
jaafar>
K-ballo: anything special you look for in time-report?
<
K-ballo>
memory usage, or suspiciously high parse times
<
K-ballo>
then I would tweak code a bit and run again.. and see what the impact was... and again and again
<
K-ballo>
for one project I went even further, and I put clang with debug info under a traditional profiler while it was compiling
<
jaafar>
I guess that's the ultimate approach
<
K-ballo>
yeah... though I ended up submitting clang patches instead of actually improving the code in question
<
K-ballo>
but at least I had a very clear understanding of what was taking time where
<
jaafar>
I imagine so :)
<
jaafar>
phase opt and generate : 174.46 ( 88%) 15.47 ( 71%) 189.99 ( 87%) 6002507 kB ( 65%)
khuck has joined #ste||ar
weilewei has quit [Remote host closed the connection]
<
khuck>
hkaiser: any chance you're around?
weilewei has joined #ste||ar
<
K-ballo>
jaafar: what's the name of that OSS tool?
hkaiser has quit [Quit: bye]
weilewei has quit [Remote host closed the connection]
hkaiser has joined #ste||ar
K-ballo1 has joined #ste||ar
K-ballo has quit [Read error: Connection reset by peer]
K-ballo has joined #ste||ar
K-ballo1 has quit [Ping timeout: 276 seconds]
K-ballo has quit [Remote host closed the connection]
K-ballo has joined #ste||ar