[llvm-dev] Should we split llvm Support and ADT?

Tue May 30 11:05:22 PDT 2017

In my experience the buildsystem works fine in combination with tablegen (at least this aspect of it). The real problem here is that tablegen is just slow. Some of the X86 tables take indeed 

Last time I looked at it tablegen had still room to optimize the way it resolves class hierarchies and the variables within which it did basically one at a time, so it needed to traverse the hierarchies again for each variables.

> On May 27, 2017, at 12:54 PM, David Blaikie via llvm-dev <llvm-dev at lists.llvm.org> wrote:
> 
> Maybe we do and build systems aren't respecting/noticing this? I'm not sure.
> 
> On Sat, May 27, 2017 at 12:50 PM Craig Topper <craig.topper at gmail.com <mailto:craig.topper at gmail.com>> wrote:
> I thought we already did write tablegen output to temporary files like X86GenAsmWriter1.inc.tmp first and then diffed them with the real .inc file and conditionally copied.
> 
> ~Craig
> 
> On Sat, May 27, 2017 at 11:02 AM, David Blaikie via llvm-dev <llvm-dev at lists.llvm.org <mailto:llvm-dev at lists.llvm.org>> wrote:
> 
> 
> On Fri, May 26, 2017 at 8:06 PM Zachary Turner via llvm-dev <llvm-dev at lists.llvm.org <mailto:llvm-dev at lists.llvm.org>> wrote:
> It would be better, because a debug tablegen is slower than an optimized tablegen, but it's still slow and it doesn't address the problem that tablegen runs *at all* when it doesn't really need to. I think if tablegen wasn't running all the time we could incremental builds down from 15 minutes (and that's on my really powerful machine) to under 5, which seemed like a big productivity boost averaged out over all users x time
> 
> Is that 10 minutes of running tablegen, though? I think this touches on Hal's comment:
> "Maybe we should use a diff-and-update approach (I thought, however, that we already did that)."
> 
> In that the 10 minutes is presumably from timestamp-based invalidation chaining down from running tablegen, to then rebuilding everything that depends on tablegen's output.
> 
> That could be addressed by either the build system or tablegen writing output to a temporary file, comparing the temp to the current file, and not modifying the current file if they're the same - thus not invalidating everything down the chain that depends on the tablegen output (not sure if that's actually enough to convince a build system not to rebuild other things - but worth a shot). This is a bit of a hackaround what the build system itself could/should be doing, but in the absence of a content based invalidation scheme - addressing this one particularly critical point in LLVM's build seems like it could be worth such a workaround. (this would speed up the build for even the times when tablegen's real dependencies change - add a new function to ArrayRef, or STLExtras, etc... and tablegen runs, but it doesn't cascade to everything else). 
> 
> (plus, opt build of tablegen, so that conclusion can be reached more quickly)
> 
> As Hal also mentioned, I'd be OK with this if there's a good logical division that can be taken more than "whatever is/isn't used in tablegen".
>  
> 
> We've got a lot of random stuff in Support, not all really related and not all relevant to every project. It seems like we should be able to do better.
> On Fri, May 26, 2017 at 6:48 PM Hal Finkel <hfinkel at anl.gov <mailto:hfinkel at anl.gov>> wrote:
> 
> On 05/26/2017 07:59 PM, Zachary Turner wrote:
>> It's that TableGen depends on Support, so if you change one file in support, support gets recompiled into a new static archive, which triggers a rerun of tablegen on all the tablegen inputs, which is extremely slow.
> 
> I thought that we had a "build an optimized tablegen even in debug mode" setting to work around this problem. Would that avoid this problem?
> 
> 
>  -Hal
> 
> 
>> 
>> On Fri, May 26, 2017 at 5:56 PM Hal Finkel <hfinkel at anl.gov <mailto:hfinkel at anl.gov>> wrote:
>> 
>> 
>> On 05/26/2017 07:47 PM, Zachary Turner via llvm-dev wrote:
>>> Changing a header file somewhere and having to spend 10 minutes waiting for a build leads to a lot of wasted developer time.
>>> 
>>> The real culprit here is tablegen.  Can we split support and ADT into two - the parts that tablegen depends on and the parts that it doesn't?
>> 
>> What's the actual problem here? Is it that TableGen regenerates different files and so we then need to rebuild all dependencies of those files? Maybe we should use a diff-and-update approach (I thought, however, that we already did that).
>> 
>>  -Hal
>> 
>>> 
>>> From what I can gather, Tablegen currently depends on these headers and all of their transitive dependencies.
>>> 
>>> #include "llvm/Support/Casting.h"
>>> #include "llvm/Support/CommandLine.h"
>>> #include "llvm/Support/Compiler.h"
>>> #include "llvm/Support/DataTypes.h"
>>> #include "llvm/Support/Debug.h"
>>> #include "llvm/Support/Error.h"
>>> #include "llvm/Support/ErrorHandling.h"
>>> #include "llvm/Support/Format.h"
>>> #include "llvm/Support/FormattedStream.h"
>>> #include "llvm/Support/LEB128.h"
>>> #include "llvm/Support/LowLevelTypeImpl.h"
>>> #include "llvm/Support/ManagedStatic.h"
>>> #include "llvm/Support/MathExtras.h"
>>> #include "llvm/Support/MemoryBuffer.h"
>>> #include "llvm/Support/PrettyStackTrace.h"
>>> #include "llvm/Support/Regex.h"
>>> #include "llvm/Support/SMLoc.h"
>>> #include "llvm/Support/ScopedPrinter.h"
>>> #include "llvm/Support/Signals.h"
>>> #include "llvm/Support/SourceMgr.h"
>>> #include "llvm/Support/raw_ostream.h"
>>> 
>>> #include "llvm/ADT/APInt.h"
>>> #include "llvm/ADT/ArrayRef.h"
>>> #include "llvm/ADT/BitVector.h"
>>> #include "llvm/ADT/CachedHashString.h"
>>> #include "llvm/ADT/DenseSet.h"
>>> #include "llvm/ADT/IndexedMap.h"
>>> #include "llvm/ADT/IntEqClasses.h"
>>> #include "llvm/ADT/MapVector.h"
>>> #include "llvm/ADT/Optional.h"
>>> #include "llvm/ADT/PointerUnion.h"
>>> #include "llvm/ADT/STLExtras.h"
>>> #include "llvm/ADT/SetVector.h"
>>> #include "llvm/ADT/SmallPtrSet.h"
>>> #include "llvm/ADT/SmallSet.h"
>>> #include "llvm/ADT/SmallVector.h"
>>> #include "llvm/ADT/SparseBitVector.h"
>>> #include "llvm/ADT/Statistic.h"
>>> #include "llvm/ADT/StringExtras.h"
>>> #include "llvm/ADT/StringMap.h"
>>> #include "llvm/ADT/StringRef.h"
>>> #include "llvm/ADT/StringSet.h"
>>> #include "llvm/ADT/StringSwitch.h"
>>> #include "llvm/ADT/TinyPtrVector.h"
>>> #include "llvm/ADT/Twine.h"
>>> 
>>> 
>>> Is this something worth putting effort into?  If so, I volunteer.
>>> 
>>> 
>>> 
>> 
>>> _______________________________________________
>>> LLVM Developers mailing list
>>> llvm-dev at lists.llvm.org <mailto:llvm-dev at lists.llvm.org>
>>> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev <http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev>
>> 
>> 
>>  -- 
>> Hal Finkel
>> Lead, Compiler Technology and Programming Languages
>> Leadership Computing Facility
>> Argonne National Laboratory
> 
> -- 
> Hal Finkel
> Lead, Compiler Technology and Programming Languages
> Leadership Computing Facility
> Argonne National Laboratory
> _______________________________________________
> LLVM Developers mailing list
> llvm-dev at lists.llvm.org <mailto:llvm-dev at lists.llvm.org>
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev <http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev>
> 
> _______________________________________________
> LLVM Developers mailing list
> llvm-dev at lists.llvm.org <mailto:llvm-dev at lists.llvm.org>
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev <http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev>
> 
> 
> _______________________________________________
> LLVM Developers mailing list
> llvm-dev at lists.llvm.org
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20170530/b8e6804e/attachment.html>