<div dir="ltr"><br><div class="gmail_extra"><br><div class="gmail_quote">On Mon, May 23, 2016 at 8:23 PM, Sean Silva <span dir="ltr"><<a href="mailto:chisophugis@gmail.com" target="_blank">chisophugis@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="ltr">Jake and I have been integrating IRPGO on PS4, and we've identified 3 remaining work items.<div><br></div><div><br></div><div>- Driver changes</div><div><br></div><div>We'd like to make IRPGO the default on PS4. We also think that it would be beneficial to make IRPGO the default PGO on all platforms (coverage would continue to use FE instr as it does currently, of course). In previous conversations (e.g. <a href="http://reviews.llvm.org/D15829" target="_blank">http://reviews.llvm.org/D15829</a>) it has come up that Apple have requirements that would prevent them from moving to IRPGO as the default PGO, at least without a deprecation period of one or two releases.</div><div><br></div><div>I'd like to get consensus on a path forward.</div><div>As a point of discussion, how about we make IRPGO the default on all platforms except Apple platforms. I really don't like fragmenting things like this (e.g. if a third-party tests "clang's" PGO they will get something different depending on the platform), but I don't see another way given Apple's constraints.</div><div><br></div><div><br></div><div>- Pre-instrumentation passes</div><div><br></div><div>Pre-instrumentation optimization has been critical for reducing the overhead of PGO for the PS4 games we tested (as expected). However, in our measurements (and we are glad to provide more info) the main benefit was inlining (also as expected). A simple pass of inlining at threshold 100 appeared to give all the benefits. Even inlining at threshold 0 gave almost all the benefits. For example, the passes initially proposed in <a href="http://reviews.llvm.org/D15828" target="_blank">http://reviews.llvm.org/D15828</a> did not improve over just inlining with threshold 100.</div><div><br></div><div>(due to PR27299 we also need to add simplifycfg after inlining to clean up, but this doesn't affect the instrumentation overhead in our measurements)</div><div><br></div><div>Bottom line: for our use cases, inlining does all the work, but we're not opposed to having more passes, which might be beneficial for non-game workloads (which is most code).</div><div><br></div><div><br></div><div>- Warnings</div><div><br></div><div>We identified 3 classes of issues which manifest as spammy warnings when applying profile data with IRPGO (these affect FEPGO also I believe, but we looked in depth at IRPGO):</div><div><br></div><div>1. The main concerning one is that getPGOFuncName mangles the filename into the counter name. This causes us to get instrprof_error::unknown_function when the pgo-use build is done in a different build directory from the training build (which is a reasonable thing to support). In this situation, PGO data is useless for all `static` functions (and as a byproduct results in a huge volume of warnings).</div></div></blockquote><div><br></div><div><br></div><div>Rong, I was just looking at implementing a fix for this, but noticed something. Can we get rid of the "InLTO" argument to getPGOFuncName if we unconditionally apply the funcname metadata to all functions?</div><div><br></div><div>-- Sean Silva</div><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="ltr"><div><br></div><div>2. In different TU's, pre-instr inlining might make different inlining decisions (for example, different functions may be available for inlining), causing hash mismatch errors (instrprof_error::hash_mismatch). In building a large game, we only saw 8 instance of this, so it is not as severe as 1, but would be good to fix.</div><div><br></div><div>3. A .cpp file may be compiled and put into an archive, but then not selected by the linker and will therefore not result in a counter in the profraw. When compiling this file with pgo-use, instrprof_error::unknown_function will result and a warning will be emitted.</div><div><br></div><div>Case 1 can be fixed using a function hash or other unique identifier instead of a file path. David, in D20195 you mentioned that Rong was working on a patch that would fix 2; we are looking forward to that.</div><div><br></div><div>For 3, I unfortunately do not know of any solution. I don't think there is a way for us to make this warning reliable in the face of this circumstance. So my conclusion is that instrprof_error::unknown_function at least must be defaulted to off unfortunately.</div><span class="HOEnZb"><font color="#888888"><div><div><br></div><div>-- Sean Silva</div></div></font></span></div>
</blockquote></div><br></div></div>