<html><head><meta http-equiv="Content-Type" content="text/html charset=utf-8"></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space;" class=""><div class="">Hi,</div><div class=""><br class=""></div><div class="">I wonder if this might be in any way related to the HUGE amount of new terms/types/coercions created by the specialiser as documented in:</div><div class=""><a href="https://ghc.haskell.org/trac/ghc/ticket/9630#comment:10" class="">https://ghc.haskell.org/trac/ghc/ticket/9630#comment:10</a></div><div class=""><br class=""></div><div class="">I don’t have a profiled version of GHC, so I was wondering if you could run your tests with a ‘-fno-specialise’, and see how everything performs then?</div><div class=""><br class=""></div><div class="">Cheers,</div><div class=""><br class=""></div><div class="">Christiaan</div><br class=""><div><blockquote type="cite" class=""><div class="">On 12 Apr 2015, at 18:28, Michal Terepeta <<a href="mailto:michal.terepeta@gmail.com" class="">michal.terepeta@gmail.com</a>> wrote:</div><br class="Apple-interchange-newline"><div class=""><div dir="ltr" class=""><div class="">So I've tried to compile Idris/Agda with prof compilers but this</div><div class="">didn't quite work out due to deps not compiling (apparently it's not</div><div class="">possible to use template haskell with a profiled compiler).</div><div class=""><br class=""></div><div class="">Out of curiosity I had a look at compiling haskell-src-exts since that</div><div class="">takes quite a while. I've used ghc HEAD and 7.8.4 (both built with</div><div class="">BuildFlavour=prof & bootstrapped with a standard ghc 7.8.4) and it's</div><div class="">interesting -- the current HEAD takes quite a bit longer and allocates</div><div class="">way more than 7.8.4. One of the main things that stand out is the</div><div class="">CallArity analysis (which IIRC was not there in 7.8.4). So unless I</div><div class="">messed something up with measuring, the analysis seem to be</div><div class="">pretty expensive.</div><div class=""><br class=""></div><div class="">Anyway, the results are below.</div><div class=""><br class=""></div><div class="">Cheers,</div><div class="">Michal</div><div class=""><br class=""></div><div class=""><br class=""></div><div class="">** HEAD</div><div class=""><br class=""></div><div class=""><span class="Apple-tab-span" style="white-space:pre"> </span>Sun Apr 12 15:52 2015 Time and Allocation Profiling Report (Final)</div><div class=""><br class=""></div><div class=""><span class="Apple-tab-span" style="white-space:pre"> </span> ghc +RTS -p -RTS [...]</div><div class=""><br class=""></div><div class=""><span class="Apple-tab-span" style="white-space:pre"> </span>total time = 147.84 secs (147841 ticks @ 1000 us, 1 processor)</div><div class=""><span class="Apple-tab-span" style="white-space:pre"> </span>total alloc = 172,378,600,408 bytes (excludes profiling overheads)</div><div class=""><br class=""></div><div class="">COST CENTRE MODULE %time %alloc</div><div class=""><br class=""></div><div class="">SimplTopBinds SimplCore 32.4 28.8</div><div class="">CallArity SimplCore 18.4 25.6</div><div class="">lintAnnots CoreLint 4.5 4.6</div><div class="">CoreTidy HscMain 4.5 5.1</div><div class="">pprNativeCode AsmCodeGen 3.2 3.4</div><div class="">OccAnal SimplCore 3.2 3.1</div><div class="">occAnalBind.assoc OccurAnal 2.6 2.5</div><div class="">StgCmm HscMain 2.3 1.9</div><div class="">Simplify SimplCore 2.1 0.2</div><div class="">RegAlloc AsmCodeGen 2.1 2.4</div><div class="">FloatOutwards SimplCore 2.0 1.6</div><div class="">regLiveness AsmCodeGen 1.9 1.9</div><div class="">tc_rn_src_decls TcRnDriver 1.8 1.3</div><div class="">sink CmmPipeline 1.7 1.5</div><div class="">NewStranal SimplCore 1.3 1.5</div><div class="">genMachCode AsmCodeGen 1.1 1.0</div><div class="">layoutStack CmmPipeline 1.0 1.0</div><div class=""><br class=""></div><div class=""><br class=""></div><div class="">** HEAD with -fno-call-arity</div><div class=""><br class=""></div><div class=""><span class="Apple-tab-span" style="white-space:pre"> </span>Sun Apr 12 18:16 2015 Time and Allocation Profiling Report (Final)</div><div class=""><br class=""></div><div class=""><span class="Apple-tab-span" style="white-space:pre"> </span> ghc +RTS -p -RTS [...] -fno-call-arity</div><div class=""><br class=""></div><div class=""><span class="Apple-tab-span" style="white-space:pre"> </span>total time = 113.71 secs (113714 ticks @ 1000 us, 1 processor)</div><div class=""><span class="Apple-tab-span" style="white-space:pre"> </span>total alloc = 121,884,896,720 bytes (excludes profiling overheads)</div><div class=""><br class=""></div><div class="">COST CENTRE MODULE %time %alloc</div><div class=""><br class=""></div><div class="">SimplTopBinds SimplCore 37.2 36.6</div><div class="">CoreTidy HscMain 6.0 7.3</div><div class="">lintAnnots CoreLint 5.8 6.5</div><div class="">pprNativeCode AsmCodeGen 4.1 4.8</div><div class="">OccAnal SimplCore 3.6 3.8</div><div class="">occAnalBind.assoc OccurAnal 2.9 3.2</div><div class="">StgCmm HscMain 2.9 2.6</div><div class="">RegAlloc AsmCodeGen 2.6 3.4</div><div class="">FloatOutwards SimplCore 2.6 2.3</div><div class="">regLiveness AsmCodeGen 2.5 2.8</div><div class="">tc_rn_src_decls TcRnDriver 2.4 1.9</div><div class="">Simplify SimplCore 2.4 0.3</div><div class="">sink CmmPipeline 2.1 2.2</div><div class="">NewStranal SimplCore 1.7 2.1</div><div class="">genMachCode AsmCodeGen 1.4 1.4</div><div class="">layoutStack CmmPipeline 1.4 1.4</div><div class="">NativeCodeGen CodeOutput 1.1 1.2</div><div class="">FloatInwards SimplCore 1.1 1.4</div><div class="">do_block Hoopl.Dataflow 1.0 0.6</div><div class="">Digraph.scc Digraph 0.8 1.3</div><div class=""><br class=""></div><div class=""><br class=""></div><div class="">** 7.8.4</div><div class=""><br class=""></div><div class=""><span class="Apple-tab-span" style="white-space:pre"> </span>Sun Apr 12 15:41 2015 Time and Allocation Profiling Report (Final)</div><div class=""><br class=""></div><div class=""><span class="Apple-tab-span" style="white-space:pre"> </span> ghc +RTS -p -RTS [...]</div><div class=""><br class=""></div><div class=""><span class="Apple-tab-span" style="white-space:pre"> </span>total time = 93.11 secs (93112 ticks @ 1000 us, 1 processor)</div><div class=""><span class="Apple-tab-span" style="white-space:pre"> </span>total alloc = 103,135,975,120 bytes (excludes profiling overheads)</div><div class=""><br class=""></div><div class="">COST CENTRE MODULE %time %alloc</div><div class=""><br class=""></div><div class="">SimplTopBinds SimplCore 38.5 37.4</div><div class="">pprNativeCode AsmCodeGen 6.2 7.2</div><div class="">StgCmm HscMain 3.9 4.2</div><div class="">RegAlloc AsmCodeGen 3.7 5.1</div><div class="">occAnalBind.assoc OccurAnal 3.3 3.6</div><div class="">OccAnal SimplCore 3.3 3.6</div><div class="">regLiveness AsmCodeGen 3.1 3.4</div><div class="">FloatOutwards SimplCore 2.9 2.4</div><div class="">sink CmmPipeline 2.8 2.8</div><div class="">Simplify SimplCore 2.6 0.3</div><div class="">tc_rn_src_decls TcRnDriver 2.4 2.1</div><div class="">genMachCode AsmCodeGen 1.9 2.0</div><div class="">NewStranal SimplCore 1.8 2.1</div><div class="">layoutStack CmmPipeline 1.8 1.8</div><div class="">Core2Core HscMain 1.3 1.2</div><div class="">deSugar HscMain 1.1 1.1</div><div class="">do_block Hoopl.Dataflow 1.1 0.7</div><div class="">CoreTidy HscMain 1.0 1.1</div><div class="">CorePrep HscMain 1.0 1.1</div><div class="">Digraph.scc Digraph 0.9 1.5</div><div class="">versioninfo MkIface 0.9 1.0</div><div class="">zonkEvBndr_zonkTcTypeToType TcHsSyn 0.6 1.4</div><div class=""><br class=""></div></div><br class=""><div class="gmail_quote">On Fri, Apr 3, 2015 at 4:49 PM David Feuer <<a href="mailto:david.feuer@gmail.com" class="">david.feuer@gmail.com</a>> wrote:<br class=""><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="ltr" class="">On a machine with an SSD instead of a hard disk, swapping greatly reduces the lifespan of the storage device.<br class=""></div><div class="gmail_extra"><br class=""><div class="gmail_quote">On Fri, Apr 3, 2015 at 10:14 AM, Bertram Felgenhauer <span dir="ltr" class=""><<a href="mailto:bertram.felgenhauer@googlemail.com" target="_blank" class="">bertram.felgenhauer@googlemail.com</a>></span> wrote:<br class=""><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><span class="">George Colpitts wrote:<br class="">
> I'm curious why the amount of RAM is relevant as all of our OS have virtual<br class="">
> memory so it is only the size of the heap and the amount of swap that<br class="">
> should be relevant for an Out Of Memory error, right?<br class="">
<br class="">
</span>The computer may not be your own. VPSs are essentially priced based on<br class="">
RAM available to the virtual server, and have limited swapping space,<br class="">
so this is an area where increased memory consumption hurts. Building<br class="">
binaries elsewhere is usually an option, but more painful than doing<br class="">
it on the VPS itself.<br class="">
<br class="">
Cheers,<br class="">
<br class="">
Bertram<br class="">
<div class=""><div class="">_______________________________________________<br class="">
Glasgow-haskell-users mailing list<br class="">
<a href="mailto:Glasgow-haskell-users@haskell.org" target="_blank" class="">Glasgow-haskell-users@haskell.org</a><br class="">
<a href="http://mail.haskell.org/cgi-bin/mailman/listinfo/glasgow-haskell-users" target="_blank" class="">http://mail.haskell.org/cgi-bin/mailman/listinfo/glasgow-haskell-users</a><br class="">
</div></div></blockquote></div><br class=""></div>
_______________________________________________<br class="">
Glasgow-haskell-users mailing list<br class="">
<a href="mailto:Glasgow-haskell-users@haskell.org" target="_blank" class="">Glasgow-haskell-users@haskell.org</a><br class="">
<a href="http://mail.haskell.org/cgi-bin/mailman/listinfo/glasgow-haskell-users" target="_blank" class="">http://mail.haskell.org/cgi-bin/mailman/listinfo/glasgow-haskell-users</a><br class="">
</blockquote></div>
_______________________________________________<br class="">Glasgow-haskell-users mailing list<br class=""><a href="mailto:Glasgow-haskell-users@haskell.org" class="">Glasgow-haskell-users@haskell.org</a><br class="">http://mail.haskell.org/cgi-bin/mailman/listinfo/glasgow-haskell-users<br class=""></div></blockquote></div><br class=""></body></html>