<div dir="ltr"><div>Cmm is too low level, I've implemented the primops in haskell in a high level way, including the out of line primops with the rts related parts (scheduler, io manager).</div><div>see:</div><div><a href="https://github.com/grin-compiler/ghc-whole-program-compiler-project/blob/master/external-stg-interpreter/lib/Stg/Interpreter/ThreadScheduler.hs">https://github.com/grin-compiler/ghc-whole-program-compiler-project/blob/master/external-stg-interpreter/lib/Stg/Interpreter/ThreadScheduler.hs</a></div><div><a href="https://github.com/grin-compiler/ghc-whole-program-compiler-project/tree/master/external-stg-interpreter/lib/Stg/Interpreter/PrimOp">https://github.com/grin-compiler/ghc-whole-program-compiler-project/tree/master/external-stg-interpreter/lib/Stg/Interpreter/PrimOp</a></div><div><br></div><div>STM is still missing though, but IMO it would be similar to concurrency/exception related primops.</div><div>Regarding the ghcjs STM implementation, IMO the primops needs to be implemented at least in Haskell in a pure way with ADTs to be easy for reasoning.<br></div><div>But thanks for the reference.</div><div><br></div><div>Currently, I'm in the design phase. I.e. I need to design the abstract domain of the STG machine values.</div><div><br></div><div>If this approach succeeds then it would be interesting to apply the calculating correct compilers method on the stg interpreter to get a compiler form it.</div><div>With this level of automation it would be extremely easy to support new target platforms, because the RTS would be generated automatically. <br></div></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Tue, Jun 8, 2021 at 10:51 PM Carter Schonwald <<a href="mailto:carter.schonwald@gmail.com">carter.schonwald@gmail.com</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div dir="auto">The stm impl In ghcjs might be a helpful comparative example on that front. </div><div dir="auto"><br></div><div dir="auto">Though I guess more broadly does this necessitate having a model of the Cmm semantics for the out of line primops ? </div><div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Tue, Jun 8, 2021 at 5:10 AM Simon Peyton Jones via ghc-devs <<a href="mailto:ghc-devs@haskell.org" target="_blank">ghc-devs@haskell.org</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
<div style="overflow-wrap: break-word;" lang="EN-GB">
<div>
<p class="MsoNormal" style="margin-right:0cm;margin-bottom:6pt;margin-left:36pt">
I wonder if there was an attempt in the past to create an abstract interpreter for GHC Core or STG to approximate the program runtime behaviour?<u></u><u></u></p>
<p class="MsoNormal"><span><u></u> <u></u></span></p>
<p class="MsoNormal"><span>No, not that I know of. Because of all the primops, concurrency, STM, etc, it would be something of a challenge. The AAM story could be interesting…</span></p></div></div><div style="overflow-wrap: break-word;" lang="EN-GB"><div><p class="MsoNormal"><span><u></u><u></u></span></p>
<p class="MsoNormal"><span><u></u> <u></u></span></p>
<p class="MsoNormal"><span>Simon<u></u><u></u></span></p>
<p class="MsoNormal"><span><u></u> <u></u></span></p>
<div style="border-color:currentcolor currentcolor currentcolor blue;border-style:none none none solid;border-width:medium medium medium 1.5pt;padding:0cm 0cm 0cm 4pt">
<div>
<div style="border-color:rgb(225,225,225) currentcolor currentcolor;border-style:solid none none;border-width:1pt medium medium;padding:3pt 0cm 0cm">
<p class="MsoNormal"><b><span lang="EN-US">From:</span></b><span lang="EN-US"> ghc-devs <<a href="mailto:ghc-devs-bounces@haskell.org" target="_blank">ghc-devs-bounces@haskell.org</a>>
<b>On Behalf Of </b>Csaba Hruska<br>
<b>Sent:</b> 07 June 2021 15:18<br>
<b>To:</b> GHC developers <<a href="mailto:ghc-devs@haskell.org" target="_blank">ghc-devs@haskell.org</a>><br>
<b>Subject:</b> abstract interpreter for GHC Core or STG<u></u><u></u></span></p>
</div>
</div>
<p class="MsoNormal"><u></u> <u></u></p>
<div>
<div>
<p class="MsoNormal" style="margin-right:0cm;margin-bottom:6pt;margin-left:0cm">
Hello,<u></u><u></u></p>
</div>
<div>
<p class="MsoNormal" style="margin-right:0cm;margin-bottom:6pt;margin-left:0cm">
<u></u> <u></u></p>
</div>
<div>
<p class="MsoNormal" style="margin-right:0cm;margin-bottom:6pt;margin-left:0cm">
I wonder if there was an attempt in the past to create an abstract interpreter for GHC Core or STG to approximate the program runtime behaviour?<u></u><u></u></p>
</div>
<div>
<p class="MsoNormal" style="margin-right:0cm;margin-bottom:6pt;margin-left:0cm">
I'm curious because I'd like to turn my external STG interterpreter to an abstract interpreter using the AAM (Abstracting Abstract Machines) method.<u></u><u></u></p>
</div>
<div>
<p class="MsoNormal" style="margin-right:0cm;margin-bottom:6pt;margin-left:0cm">
This approach seems promising to me because a single Haskell code base (ext STG interpreter) could be the specification of the Haskell operational semantics and also be a detailed static analyzer that could help optimization transformations.<u></u><u></u></p>
</div>
<div>
<p class="MsoNormal" style="margin-right:0cm;margin-bottom:6pt;margin-left:0cm">
I'm interested in any attempt that happened during GHC/Haskell evolution.<u></u><u></u></p>
</div>
<div>
<p class="MsoNormal" style="margin-right:0cm;margin-bottom:6pt;margin-left:0cm">
<u></u> <u></u></p>
</div>
<div>
<p class="MsoNormal" style="margin-right:0cm;margin-bottom:6pt;margin-left:0cm">
Regards,<u></u><u></u></p>
</div>
<div>
<p class="MsoNormal" style="margin-right:0cm;margin-bottom:6pt;margin-left:0cm">
Csaba Hruska<u></u><u></u></p>
</div>
</div>
</div>
</div>
</div>
_______________________________________________<br>
ghc-devs mailing list<br>
<a href="mailto:ghc-devs@haskell.org" target="_blank">ghc-devs@haskell.org</a><br>
<a href="http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs" rel="noreferrer" target="_blank">http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs</a><br>
</blockquote></div></div>
</blockquote></div>