hkaiser changed the topic of #ste||ar to: STE||AR: Systems Technology, Emergent Parallelism, and Algorithm Research | stellar-group.org | HPX: A cure for performance impaired parallel applications | github.com/STEllAR-GROUP/hpx | This channel is logged: irclog.cct.lsu.edu
K-ballo1 has joined #ste||ar
K-ballo has quit [Ping timeout: 264 seconds]
K-ballo1 is now known as K-ballo
hkaiser has quit [Quit: Bye!]
Yorlik_ has joined #ste||ar
Yorlik has quit [Ping timeout: 265 seconds]
HHN93 has joined #ste||ar
<HHN93> I am having some trouble setting up HPX on rostam
<HHN93> what sholud be the boost include dir?
HHN93 has quit [Client Quit]
<satacker[m]> <HHN93> "what sholud be the boost include..." <- Don't do that, module load boost llvm
<satacker[m]> <Arnav-Negi> "satacker will you be available..." <- We can start atleast you can work on early, let me know if you can meet this sunday any time?
<satacker[m]> Oh wait
<satacker[m]> I forgot about Alireza
K-ballo1 has joined #ste||ar
K-ballo has quit [Ping timeout: 265 seconds]
K-ballo1 is now known as K-ballo
hkaiser has joined #ste||ar
HHN93 has joined #ste||ar
<HHN93> gonidelis[m] can you please clarify about boost?
<HHN93> ok got it
<HHN93> I am receiving the following error when trying to compile a program
<HHN93> unable to initialize decompress status for section .debug_info
<HHN93> ok it does compile, but on trying to execute it results in an error
<hkaiser> HHN93: uhh - on rostam?
<HHN93> multiple libraries not found
<HHN93> yes
<HHN93> for the newer GCC compiler
<hkaiser> strange, we run multiple CIs on rostam with various configurations, all successful
<hkaiser> what 'newer gcc'?
<HHN93> I think we do face lot of issues with newer gcc compiler, I remember this was also the case when I tried add module support
<HHN93> gcc/12.3
<hkaiser> use a different compiler and report the issues to Alireza, please
<HHN93> with gcc/11.2 it's fine no issues
<HHN93> also tried fixing a boost deprecated warning, that's fine right?
<hkaiser> in boost?
<hkaiser> where did you fix it?
<HHN93> we include a boost header, which is deprecated in boost/1.78
<HHN93> changed the header
<hkaiser> ahh, sure
<hkaiser> just make sure it still works with older versions of Boost
<HHN93> also I get the following log when executing
<HHN93> tcmalloc: large alloc 4294967296 bytes == 0x2a0a000 @
<HHN93> did I mess up the build? shouldn't the log slow down execution?
<HHN93> `just make sure it still works with older versions of Boost`
<HHN93> yeah that's what I wanted to know, will the CIs take care of it?
<hkaiser> have an eye on it
<HHN93> `tcmalloc: large alloc 4294967296 bytes == 0x2a0a000 @`
<HHN93> are these logs supposed to happen in Release build
<hkaiser> never seen those
<HHN93> ok apparently its kinda a warning message for large allocations, and there's an env variable to increase the threshold
<hkaiser> HHN93: I'm working on WIndows - no tcmalloc ;-)
leohoo_sdu[m] has joined #ste||ar
leohoo_sdu[m] has left #ste||ar [#ste||ar]
HHN93 has quit [Quit: Client closed]
hkaiser has quit [Quit: Bye!]
HHN93 has joined #ste||ar
HHN93 has quit [Quit: Client closed]
Yorlik_ is now known as Yorlik
Arnav-Negi has joined #ste||ar
Arnav-Negi has quit [Client Quit]
HHN93 has joined #ste||ar
HHN93 has quit [Client Quit]
K-ballo has quit [Ping timeout: 264 seconds]
K-ballo1 has joined #ste||ar
K-ballo1 is now known as K-ballo
hkaiser has joined #ste||ar
hkaiser_ has joined #ste||ar
hkaiser has quit [Ping timeout: 240 seconds]
HHN93 has joined #ste||ar
<HHN93> is there any way to force a loop to not be vectorized?
<HHN93> any gcc/clang compiler directives?
<hkaiser_> HHN93: compile with -O0
<HHN93> I need to test performance
<HHN93> want to make sure vectorization is not causing the slowdown
<HHN93> the loop is getting auto vectorized
<gnikunj[m]> HHN93: You’re looking for -fno-tree-vectorize
<gnikunj[m]> It’s a gcc flag that disables vectorization
<HHN93> that disables vectorization for all files right?
<HHN93> I want to disable specfically for a loop
<hkaiser_> HHN93: you should be able to use all options in pragmas
<gnikunj[m]> HHN93: there were pragmas in OpenMP land but I’m not sure if they work without OpenMP threading.
<gnikunj[m]> ICC has #pragma novector as well
HHN93 has quit [Quit: Client closed]