On Jul 6, 2012, at 3:48 PM, Bill Wendling wrote:> On Jul 5, 2012, at 1:33 AM, Duncan Sands wrote: >> Hi Bill, >> >>> Nothing that I'm aware of has changed with EH. Is it possible to bisect the problem? >> >> I don't see any relevant LLVM changes, so I guess clang C++ compilation slowed >> down due to some clang changes. I'm not going to investigate this. >> > Crumbs. > > John, Do you know of anything that went into clang that could account for EH slowdowns?Is this Linux? Is it possibly just Objective-C tests? There have been several patches about ObjC runtimes recently that affected exceptions. John.
Hi John, On 07/07/12 02:20, John McCall wrote:> On Jul 6, 2012, at 3:48 PM, Bill Wendling wrote: >> On Jul 5, 2012, at 1:33 AM, Duncan Sands wrote: >>> Hi Bill, >>> >>>> Nothing that I'm aware of has changed with EH. Is it possible to bisect the problem? >>> >>> I don't see any relevant LLVM changes, so I guess clang C++ compilation slowed >>> down due to some clang changes. I'm not going to investigate this. >>> >> Crumbs. >> >> John, Do you know of anything that went into clang that could account for EH slowdowns? > > Is this Linux? Is it possibly just Objective-C tests? There have been several patches about ObjC runtimes recently that affected exceptions.yes, it is linux. While the worst slowdowns were in Objective-C tests, there were lots of non-objc slowdowns too (see list in quoted message in Bill's earlier email). Ciao, Duncan.
On Jul 7, 2012, at 12:10 AM, Duncan Sands wrote:> On 07/07/12 02:20, John McCall wrote: >> On Jul 6, 2012, at 3:48 PM, Bill Wendling wrote: >>> On Jul 5, 2012, at 1:33 AM, Duncan Sands wrote: >>>> Hi Bill, >>>> >>>>> Nothing that I'm aware of has changed with EH. Is it possible to bisect the problem? >>>> >>>> I don't see any relevant LLVM changes, so I guess clang C++ compilation slowed >>>> down due to some clang changes. I'm not going to investigate this. >>>> >>> Crumbs. >>> >>> John, Do you know of anything that went into clang that could account for EH slowdowns? >> >> Is this Linux? Is it possibly just Objective-C tests? There have been several patches about ObjC runtimes recently that affected exceptions. > > yes, it is linux. While the worst slowdowns were in Objective-C tests, there > were lots of non-objc slowdowns too (see list in quoted message in Bill's > earlier email).In that case, I do not know of anything that would have caused a global slowdown in exception handling. John.