similar to: [LLVMdev] Specifying Additional Compilation Passes to lli

Hi, Is this the right mailing list for sending in diffs by irregular contributers? Should I send diffs directly to the code owner instead? Tom ----- Original Message ----- From: "Thomas B. Jablin" <tjablin at CS.Princeton.EDU> To: "LLVM Developers Mailing List" <llvmdev at cs.uiuc.edu> Sent: Thursday, September 11, 2008 1:55:09 PM GMT -05:00 US/Canada Eastern

Hi, I'm interested in specifying some additional passes to the JIT via the command-line. The enclosed patch allows lli to take compiler passes as command-line arguments in the same way opt does. This is my first submission, so any comments, criticisms, or observations would be very welcome. Thanks. Tom Jablin -------------- next part -------------- A non-text attachment was scrubbed... Name:

Evan, My overall goal is to support dynamic optimization in LLVM. In order to do so, I must gather profiling information at runtime, then recompile the profiled functions. Currently, I'm just adding and removing calls into my profiler in a custom pass. What is the advantage of giving the JIT a second profile manager over my current implementation? Thanks. Tom ----- Original Message -----

----- "Evan Cheng" <evan.cheng at apple.com> wrote: > On Sep 16, 2008, at 8:44 AM, Thomas B. Jablin wrote: > > > Evan, > > So, if I understand you correctly, the design you have in mind is > > to: create a PassManager, pass it to the JIT on construction, and > > modify runJITOnFunction to run the second PassManager on the > > Function

Evan, So, if I understand you correctly, the design you have in mind is to: create a PassManager, pass it to the JIT on construction, and modify runJITOnFunction to run the second PassManager on the Function being jit'd before running the codegen PassManager. Thanks. Tom ----- Original Message ----- From: "Evan Cheng" <evan.cheng at apple.com> To: "LLVM Developers Mailing

On Sep 16, 2008, at 12:17 PM, Thomas B. Jablin wrote: > > ----- "Evan Cheng" <evan.cheng at apple.com> wrote: > >> On Sep 16, 2008, at 8:44 AM, Thomas B. Jablin wrote: >> >>> Evan, >>> So, if I understand you correctly, the design you have in mind is >>> to: create a PassManager, pass it to the JIT on construction, and >>>

Does any real code benefit from dead code eliminating read-only functions? Tom ----- Original Message ----- From: "Reid Kleckner" <rnk at mit.edu> To: "Thomas B. Jablin" <tjablin at CS.Princeton.EDU> Cc: "LLVM Developers Mailing List" <llvmdev at cs.uiuc.edu> Sent: Monday, May 10, 2010 9:38:47 PM GMT -05:00 US/Canada Eastern Subject: Re: [LLVMdev]

What do you think of this patch? I have added a check for the case I mentioned in the previous email as well as a similar situation I discovered later. Tom ----- Original Message ----- From: "Dan Gohman" <gohman at apple.com> To: "Thomas B. Jablin" <tjablin at CS.Princeton.EDU> Cc: "LLVM Developers Mailing List" <llvmdev at cs.uiuc.edu> Sent:

Sorry. Here is the correct version. ----- Original Message ----- From: "Thomas B. Jablin" <tjablin at CS.Princeton.EDU> To: "LLVM Developers Mailing List" <llvmdev at cs.uiuc.edu> Sent: Friday, December 5, 2008 1:08:57 PM GMT -05:00 US/Canada Eastern Subject: Re: [LLVMdev] MachineCodeEmitter Patch Evan, Here are the modifications you asked for. I have, for the

On Thu, Sep 25, 2008 at 5:04 PM, Thomas B. Jablin <tjablin at cs.princeton.edu> wrote: > Hi, > The call target pass in the poolalloc suite yields an incorrect output for the following short test program: The DSA results are now (r56847) correct for this test case. The call is marked incomplete. Doing better is actually a pathological case in DSA which is hard to fix without

Dear All, Should this patch be committed now? I just did a quick read-over of it when it was first submitted. -- John T. -------------- next part -------------- An embedded message was scrubbed... From: "Thomas B. Jablin" <tjablin at CS.Princeton.EDU> Subject: Re: [LLVMdev] insertAfter method Patch Date: Mon, 12 Jan 2009 15:58:34 -0500 (EST) Size: 4813 URL:

Nevermind. I misread the documentation. On Sat, Apr 30, 2011 at 8:12 PM, Thomas Jablin <tjablin at gmail.com> wrote: > Hi, > > Is it valid to query ScalarEvolution::getSCEVAtScope with a Value > which is an Instruction inside the provided Loop? If so, I believe > there is a bug in computeSCEVAtScope. Specifically: > > 04716     // If the scope is outside the

Thanks. I do not have commit privilege. Tom ----- Original Message ----- From: "Evan Cheng" <evan.cheng at apple.com> To: "LLVM Developers Mailing List" <llvmdev at cs.uiuc.edu> Sent: Monday, December 8, 2008 5:39:33 PM GMT -05:00 US/Canada Eastern Subject: Re: [LLVMdev] MachineCodeEmitter Patch Looks good. Do you have commit privilege? Evan On Nov 22, 2008, at

Hi, I saw that Nick Lewycky removed the LLVM portable hash_map and hash_sets. My research group was relying on those classes. What replaces hash_map and hash_set? The LLVM implementation was very convenient as there is no equivalent in STL, and Microsoft's implementation had a different name, as does the C++ 0X implementation. Thanks. Tom Jablin

That's a good point.  But it's worth noting that recompileAndRelinkFunction() and freeMachineCodeForFunction() are both vestiges of the old JIT (i.e. the "JIT" as opposed to the "MCJIT").  The old JIT is no longer actively supported. -Phil On April 26, 2014 at 9:47:05 AM, Sri (emdcdeveloper at gmail.com) wrote: Hi Fillip                  Addition to my previous

Looks good. Do you have commit privilege? Evan On Nov 22, 2008, at 1:19 PM, Thomas Jablin wrote: > Here is the corrected version. > > Thomas Jablin wrote: >> Actually, there is a problem with the patch. Please delay review. >> >> Thomas Jablin wrote: >> >>> Hi, >>> The following code: >>> >>> #include<stdio.h>

On Tue, Dec 22, 2009 at 6:14 AM, Chris Lattner <clattner at apple.com> wrote: > On Dec 19, 2009, at 3:36 PM, Gianluca Guida wrote: >> Attached patch makes runJITOnFunction more reliable. > > When would MCI be null? Everytime you call recompileAndRelinkFunction. It calls runJITOnFunction without specifying the MCI argument, which get defaulted to NULL. Gianluca -- It was a

Hi All, I just implemented an often-requested feature: the ability to delete machine code out of the JIT's code buffer (the ExecutionEngine::freeMachineCodeForFunction(F) method). The implementation uses a very general free-list mechanism for tracking free ranges in the buffer, and it works perfectly in my simple test cases designed to cover all of the code paths. However, I don't

Evan, Here are the modifications you asked for. I have, for the most part, not changed intptr_t to uintptr_t inside the JITInfo classes, because the pointer arithmetic there can sometimes legitimately yield negative numbers. Tom ----- Original Message ----- From: "Evan Cheng" <evan.cheng at apple.com> To: "LLVM Developers Mailing List" <llvmdev at cs.uiuc.edu>

Here is the corrected version. Thomas Jablin wrote: > Actually, there is a problem with the patch. Please delay review. > > Thomas Jablin wrote: > >> Hi, >> The following code: >> >> #include<stdio.h> >> >> char bigArray[0x1000000]; >> >> int main(int argc, char **argv) { >> printf("mem: 0x%x\n", (unsigned)