llvm dev - Jul 2012 - [LLVMdev] FYI: Planning to remove ProfileInfo and related passes from LLVM

On Sun, Jul 15, 2012 at 8:32 AM, Alastair Murray <alastairmurray42 at
gmail.com
> wrote:
> Hi Chandler,
>
> I'm a GSoC student working on profiling support (mentor CC'ed). 
I'm no
> stranger to the issues with the current system: my original proposal was
> written without knowledge of the limitations.  This is why this list
> hasn't heard much from me yet.
>
It would be better to actually email the list with the problems you are
hitting sooner. We would have directed you to the BranchProbabilityInfo and
BlockFrequencyInfo infrastructure that is now actively in use by several
optimization passes in LLVM.

Can you start a thread with what your specific goals are? The only email I
have from you is from May 1st, which was your project outline for the GSoC
application, to which Evan responded, but to which nothing else ever
materialized.

I think if any of us knew that you were actively working on the existing
ProfileInfo and profile loading infrastructure we would have jumped in
quickly to redirect things. I'll provide a bit of feedback below, but we
should really hold the main part of this discussion on a separate thread,
as ultimately *how* you can use the BPI and BFI analyses and the metadata
they are predicated on is independent from whether we remove the existing
code... The important thing is that you can use them, and that the existing
code won't help much (if at all).

I would like to continue working on profiling support but I'm
not
> attached to ProfileInfo and wouldn't be distraught if it gets removed.
> I'd rather work on a useful solution than a dead one and nothing
I've
> done so far couldn't be ported to a different interface.  I know some
> people are definitely interested in profiling support.
>
I am one of them. ;] I've been working on the BPI- and BFI-based
optimizations in LLVM.

My reason for doing this GSoC project was to gain experience with
LLVM
> and that aim doesn't disappear after one summer's worth of work. 
My
> personal plan was to keep on working on profiling support beyond the end
> date, long-term (though not full time).  I do compiler research, but not
> currently with LLVM -- so this is a genuine desire and plan.
>
The last thing I want is for GSoC work to come to nothing. I've been a GSoC
student, and so it's important to me that we give you useful things to work
on that the overarching LLVM community benefits from.

> Perhaps the best way to proceed is to remove ProfileInfo etc and then I
> can work on re-adding/re-writing it with support for BranchProbability
> and BlockFrequency.  Then I can maintain what I add.  (As an unknown
> figure offering to maintain the existing code seems a bit hollow.)
>
> Naturally, any changes to my GSoC plans need discussion with my mentor.
>
Yes, I think you and your mentor need to discuss this and to start engaging
with the LLVM community through the mailing list about the design of the
summer-of-code work, how it should be structured, etc. I think that the
community at large is extremely interested in improved PGO support in LLVM,
and so we'd be really interested in giving feedback and direction advice...


I'd like to hear an explicit OK from you and your mentor before I remove
anything, as I don't want to get in the way of any immediate progress or
work you're doing as part of GSoC.


A few specifics below.
>
> On 15/07/12 20:07, Chandler Carruth wrote:
> > Hello folks,
> >
> > I'd like to remove all of the old and defunct profile info passes
from
> > LLVM. These have been almost entirely supplanted by the
> > BranchProbability and BlockFrequency systems, which are actually on by
> > default, and in use in optimization passes.
> >
> > The old system is not on, and hasn't been touched in years except
to do
> > minor build fixes and updates.
> >
> > As far as I'm aware, the only thing the old system supported which
the
> > new one does not is loading profile data. However, it didn't
support
> > doing anything useful with that data once loaded,
>
> There is also the llvm-prof tool, but yes: there is only one
> optimization pass that uses profiling data and it is not on by default.
>   The hindrance to using ProfileInfo in other passes is that almost no
> passes preserve it and it can not be recalculated (just estimated by
> ProfileEstimate).
>
> As far as I can tell BranchProbability and BlockFrequency would also be
> invalidated by any CFG altering passes.  Preserving ProfileInfo was my
> primary long-term (post-GSoC) task, preserving BranchProbability and
> BlockFrequency instead should be no harder.
>
Details for another thread, but essentially, BranchProbabilityInfo is going
to "recompute" the probabilities, but only insofar as they stem from
heuristics that can be computed. If you look at the pass, there is also
support for loading branch weight metadata and using that to form
probability estimates. This metadata is attached to the IR and is the
primary mechanism for preserving information.

The overarching expected strategy is that a profile-loading pass would load
a profile and convert it to metadata annotations on the IR itself. Then
passes would be taught to preserve this metadata wherever possible, and
BranchProbabilityInfo will compute probabilities based on the metadata,
falling back to static heuristics only when the metadata is absent.

The missing pieces are:
1) Ensuring that passes such as SimplifyCFG preserve as much branch weight
metadata as possible. This is already an issue for metadata that comes from
__builtin_expect() source code annotations, so it is something you can
write tests for today and observe problems.
2) A pass to load profile data and attach metadata to branch instructions
based on it.
3) Ensuring that #2 works with the profiles produced by an instrumented
binary.

It's important to note that #2 is easy. It's #1 and #3 that are really
hard. When I have talked to others about #3, there has been the feeling
that we would probably want to write a custom instrumentation pass that
would add instrumentation to the right LLVM IR formations, in order to make
sure that when it is read back in, the profile data matches up with the IR
correctly, and is available early enough in the optimization passes to be
used.
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20120715/0baec10d/attachment.html>

Hi Chandler and Alastair,

I have been using the Profile.pl and the related passes and optimizations for
about 4 years now. With every new release lately, the support for the profile
scripts and their framework seemed to be downgrading. Hence, I used my own tiny
one line fixes to keep them working. I offered to send these small patches to
keep these scripts working, to the LLVM dev so that others can use it if they
want. I think my email got lost that time, in the sea of emails we see on the
dev mail list.

I must confess that I was not aware of the BPI and BFI infrastructures. The
breaking of the profiling infrastructure always baffled me. Now it makes sense,
since it has been superseded by these new frameworks.

Anyway, if you guys decide to keep the old profiling framework, it would be
good. As Alastair has mentioned, the llvm-prof helps in a way for instrumenting
the code with the profile data. Maybe this is also part of the BPI/BFI in which
case it's great.

Cheers,
Alok
PhD Candidate, NTU Singapore.

From: llvmdev-bounces at cs.uiuc.edu [mailto:llvmdev-bounces at cs.uiuc.edu] On
Behalf Of Chandler Carruth
Sent: Monday, July 16, 2012 6:34 AM
To: Alastair Murray
Cc: csdavec at swan.ac.uk; llvmdev at cs.uiuc.edu
Subject: Re: [LLVMdev] FYI: Planning to remove ProfileInfo and related passes
from LLVM

On Sun, Jul 15, 2012 at 8:32 AM, Alastair Murray <alastairmurray42 at
gmail.com<mailto:alastairmurray42 at gmail.com>> wrote:
Hi Chandler,

I'm a GSoC student working on profiling support (mentor CC'ed).  I'm
no
stranger to the issues with the current system: my original proposal was
written without knowledge of the limitations.  This is why this list
hasn't heard much from me yet.

It would be better to actually email the list with the problems you are hitting
sooner. We would have directed you to the BranchProbabilityInfo and
BlockFrequencyInfo infrastructure that is now actively in use by several
optimization passes in LLVM.

Can you start a thread with what your specific goals are? The only email I have
from you is from May 1st, which was your project outline for the GSoC
application, to which Evan responded, but to which nothing else ever
materialized.

I think if any of us knew that you were actively working on the existing
ProfileInfo and profile loading infrastructure we would have jumped in quickly
to redirect things. I'll provide a bit of feedback below, but we should
really hold the main part of this discussion on a separate thread, as ultimately
*how* you can use the BPI and BFI analyses and the metadata they are predicated
on is independent from whether we remove the existing code... The important
thing is that you can use them, and that the existing code won't help much
(if at all).

I would like to continue working on profiling support but I'm not
attached to ProfileInfo and wouldn't be distraught if it gets removed.
I'd rather work on a useful solution than a dead one and nothing I've
done so far couldn't be ported to a different interface.  I know some
people are definitely interested in profiling support.

I am one of them. ;] I've been working on the BPI- and BFI-based
optimizations in LLVM.

My reason for doing this GSoC project was to gain experience with LLVM
and that aim doesn't disappear after one summer's worth of work.  My
personal plan was to keep on working on profiling support beyond the end
date, long-term (though not full time).  I do compiler research, but not
currently with LLVM -- so this is a genuine desire and plan.

The last thing I want is for GSoC work to come to nothing. I've been a GSoC
student, and so it's important to me that we give you useful things to work
on that the overarching LLVM community benefits from.

Perhaps the best way to proceed is to remove ProfileInfo etc and then I
can work on re-adding/re-writing it with support for BranchProbability
and BlockFrequency.  Then I can maintain what I add.  (As an unknown
figure offering to maintain the existing code seems a bit hollow.)

Naturally, any changes to my GSoC plans need discussion with my mentor.

Yes, I think you and your mentor need to discuss this and to start engaging with
the LLVM community through the mailing list about the design of the
summer-of-code work, how it should be structured, etc. I think that the
community at large is extremely interested in improved PGO support in LLVM, and
so we'd be really interested in giving feedback and direction advice...


I'd like to hear an explicit OK from you and your mentor before I remove
anything, as I don't want to get in the way of any immediate progress or
work you're doing as part of GSoC.


A few specifics below.

On 15/07/12 20:07, Chandler Carruth wrote:
> Hello folks,
>
> I'd like to remove all of the old and defunct profile info passes from
> LLVM. These have been almost entirely supplanted by the
> BranchProbability and BlockFrequency systems, which are actually on by
> default, and in use in optimization passes.
>
> The old system is not on, and hasn't been touched in years except to do
> minor build fixes and updates.
>
> As far as I'm aware, the only thing the old system supported which the
> new one does not is loading profile data. However, it didn't support
> doing anything useful with that data once loaded,
There is also the llvm-prof tool, but yes: there is only one
optimization pass that uses profiling data and it is not on by default.
  The hindrance to using ProfileInfo in other passes is that almost no
passes preserve it and it can not be recalculated (just estimated by
ProfileEstimate).

As far as I can tell BranchProbability and BlockFrequency would also be
invalidated by any CFG altering passes.  Preserving ProfileInfo was my
primary long-term (post-GSoC) task, preserving BranchProbability and
BlockFrequency instead should be no harder.

Details for another thread, but essentially, BranchProbabilityInfo is going to
"recompute" the probabilities, but only insofar as they stem from
heuristics that can be computed. If you look at the pass, there is also support
for loading branch weight metadata and using that to form probability estimates.
This metadata is attached to the IR and is the primary mechanism for preserving
information.

The overarching expected strategy is that a profile-loading pass would load a
profile and convert it to metadata annotations on the IR itself. Then passes
would be taught to preserve this metadata wherever possible, and
BranchProbabilityInfo will compute probabilities based on the metadata, falling
back to static heuristics only when the metadata is absent.

The missing pieces are:
1) Ensuring that passes such as SimplifyCFG preserve as much branch weight
metadata as possible. This is already an issue for metadata that comes from
__builtin_expect() source code annotations, so it is something you can write
tests for today and observe problems.
2) A pass to load profile data and attach metadata to branch instructions based
on it.
3) Ensuring that #2 works with the profiles produced by an instrumented binary.

It's important to note that #2 is easy. It's #1 and #3 that are really
hard. When I have talked to others about #3, there has been the feeling that we
would probably want to write a custom instrumentation pass that would add
instrumentation to the right LLVM IR formations, in order to make sure that when
it is read back in, the profile data matches up with the IR correctly, and is
available early enough in the optimization passes to be used.

-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20120716/d97fcdd2/attachment.html>

On 16/07/12 07:33, Chandler Carruth wrote:
> Can you start a thread with what your specific goals are?
This is to follow.

> I'd like to hear an explicit OK from you and your mentor before I
remove
> anything, as I don't want to get in the way of any immediate progress
or
> work you're doing as part of GSoC.
This is an explicit OK for you to remove ProfileInfo and related code. 
My mentor and I agree that it is better to work on code that will be 
sticking around.

The existing EdgeProfiling.cpp and the runtime code may be worth saving 
(these require ProfilingUtils.* and ProfileInfoTypes.h respectively).

Without a profile loader they aren't useful (even llvm-prof is based on 
ProfileInfoLoader), but as they count edge frequency it intuitively 
seems like the profiling data they generate could be used by a rewritten 
BranchProbability profile loader.

One note, if I plan on submitting patches to restore rewritten parts is 
destroying `svn blame` history a problem?

> The missing pieces are:
> 1) Ensuring that passes such as SimplifyCFG preserve as much branch
> weight metadata as possible. This is already an issue for metadata that
> comes from __builtin_expect() source code annotations, so it is
> something you can write tests for today and observe problems.
> 2) A pass to load profile data and attach metadata to branch
> instructions based on it.
> 3) Ensuring that #2 works with the profiles produced by an instrumented
> binary.
>
> It's important to note that #2 is easy. It's #1 and #3 that are
really
> hard. When I have talked to others about #3, there has been the feeling
> that we would probably want to write a custom instrumentation pass that
> would add instrumentation to the right LLVM IR formations, in order to
> make sure that when it is read back in, the profile data matches up with
> the IR correctly, and is available early enough in the optimization
> passes to be used.
I understand that #1 is hard, but #3 isn't so clear to me.  The current 
ProfileInfo based -insert-edge-profiling and -profile-loader passes can 
already do this.  The instrumentation and loading passes have to occur 
at exactly the same point of their respective pass pipelines, but that 
is unavoidable.


Regards,
Alastair.

Hi Alok,

On 16/07/12 12:41, #ALOK PRAKASH# wrote:
> I have been using the Profile.pl and the related passes and
> optimizations for about 4 years now. With every new release lately, the
> support for the profile scripts and their framework seemed to be
> downgrading. Hence, I used my own tiny one line fixes to keep them
> working. I offered to send these small patches to keep these scripts
> working, to the LLVM dev so that others can use it if they want. I think
> my email got lost that time, in the sea of emails we see on the dev mail
> list.
profile.pl has seen a non-cleanup related commit in 5 years ...  It 
seems so simple I'm not sure I see a need for it.  But clearly you use it!
> I must confess that I was not aware of the BPI and BFI infrastructures.
> The breaking of the profiling infrastructure always baffled me. Now it
> makes sense, since it has been superseded by these new frameworks.
>
> Anyway, if you guys decide to keep the old profiling framework, it would
> be good. As Alastair has mentioned, the llvm-prof helps in a way for
> instrumenting the code with the profile data. Maybe this is also part of
> the BPI/BFI in which case it's great.
If a new profiler loader based on BPI gets written then llvm-prof could 
be rewritten to use the BFI system without too much difficulty.

I found llvm-prof useful for examining the behaviour of profiling 
support within LLVM, but I'm curious about what you use it for.  It 
seems quite limited for performance analysis.

Regards,
Alastair.