<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html; charset=ISO-8859-1"
http-equiv="Content-Type">
</head>
<body bgcolor="#ffffff" text="#000000">
Sorry to resurrect an old thread, but I finally got around to
testing this approach (round tripping through bitcode in memory) and
it works beautifully - and isn't that much slower than cloning.<br>
<br>
I have noticed however that the copy process isn't thread-safe. The
problem is that in Function, there is lazy initialization code for
arguments:<br>
<br>
void CheckLazyArguments() const {<br>
if (hasLazyArguments())<br>
BuildLazyArguments();<br>
}<br>
void BuildLazyArguments() const;<br>
<br>
I've worked around this by calling Function::getArgumentList()
outside the threaded code to harden it before the threaded copies.
Are there other lazy data structures that need to be solidified
before threading? Should I be playing it safe and put a thread lock
around the entire copy procedure?<br>
<br>
In case you are interested, here is the algorithm I'm using to copy
a Module to a same (or different) LLVMContext:<br>
<br>
if (&context == &other.myContext)<br>
{<br>
// If the context is shared, we can use cloning<br>
ValueToValueMapTy map;<br>
myModule = CloneModule(other.myModule, map);<br>
}<br>
else<br>
{<br>
// Otherwise, round trip the module to a stream and then
back<br>
// into the new context. This approach allows for
duplication<br>
// and optimization to proceed in parallel for different<br>
// modules.<br>
std::string str;<br>
raw_string_ostream stream(str);<br>
WriteBitcodeToFile(other.myModule, stream);<br>
<br>
StringRef ref(stream.str());<br>
UT_ScopedPtr<MemoryBuffer>
buf(MemoryBuffer::getMemBuffer(ref));<br>
myModule = ParseBitcodeFile(buf.get(), myContext);<br>
<br>
UT_ASSERT(myModule);<br>
}<br>
<br>
<br>
On 06/18/2013 01:29 PM, Reid Kleckner wrote:
<blockquote
cite="mid:CACs=ty+ot6actnh944Z0Ft1w0CMWWRNzS+N4X+Q7B9b=GQX-PA@mail.gmail.com"
type="cite">
<div dir="ltr">You could probably round trip it through bitcode in
memory. I think all of the IR cloning functionality assumes
that only one context is being used. Even if the serialization
isn't efficient as a clone could be, it should give you very
high confidence that everything Just Works. :)</div>
<div class="gmail_extra"><br>
<br>
<div class="gmail_quote">On Tue, Jun 18, 2013 at 1:16 PM, Andrew
Clinton <span dir="ltr"><<a moz-do-not-send="true"
href="mailto:andrew@sidefx.com" target="_blank">andrew@sidefx.com</a>></span>
wrote:<br>
<blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt
0.8ex; border-left: 1px solid rgb(204, 204, 204);
padding-left: 1ex;">
I have a Module/LLVMContext that I'd like to clone and
manipulate in different threads (each thread may perform
different translation / optimization, so they need unique
copies). Currently this process has to be locked, since
each clone of the Module still refers to the same
LLVMContext. Is there a way to clone both the Module and
LLVMContext so that the copies can be manipulated
independently in different threads?<br>
<br>
Andrew<br>
_______________________________________________<br>
LLVM Developers mailing list<br>
<a moz-do-not-send="true" href="mailto:LLVMdev@cs.uiuc.edu"
target="_blank">LLVMdev@cs.uiuc.edu</a> <a
moz-do-not-send="true" href="http://llvm.cs.uiuc.edu"
target="_blank">http://llvm.cs.uiuc.edu</a><br>
<a moz-do-not-send="true"
href="http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev"
target="_blank">http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev</a><br>
</blockquote>
</div>
<br>
</div>
</blockquote>
<br>
</body>
</html>