Just tried your test program on my IVB machine, and it works fine as below:
gongzg@ivb-gt2-rev4:/tmp/testclbug$ ./testclbug OpenCL platform name: Intel Gen OCL Driver OpenCL version: OpenCL 1.2 beignet 1.0 (git-07809dc) OpenCL device: Intel(R) HD Graphics IvyBridge GT2 going to compile CL prog prog compiled gongzg@ivb-gt2-rev4:/tmp/testclbug$ llvm-config --version 3.5.0 As the error message is definitely from LLVM internal, I doubt there is still something wrong with your LLVM installation. It may be helpful if you can share the llvm ir output. You can try as below: # export OCL_OUTPUT_LLVM_AFTER_LINK=1 # ./testclbug > llvm.log 2>&1 If you can't get any llvm ir by this method, you can try: # export OCL_OUTPUT_LLVM_BEFORE_LINK=1 # ./testclbug > llvm.log 2>&1 > -----Original Message----- > From: Nicolas Bourdaud [mailto:nicolas.bourd...@gmail.com] > Sent: Tuesday, December 2, 2014 6:11 PM > To: Song, Ruiling; Zhigang Gong; beignet@lists.freedesktop.org > Subject: Re: [Beignet] Program compilation crash on Ivy Bridge > > Thanks for your answer. > > I am using the attached file version. I have retried (recompiled beignet 1.0, > install and rerun the test) making sure that there is absolutely only > llvm/clang > 3.5 on the system and I still get the same on IvyBridge: > > > OpenCL platform name: Intel Gen OCL Driver OpenCL version: OpenCL 1.2 > > beignet 1.0.0 OpenCL device: Intel(R) HD Graphics IvyBridge GT2 going > > to compile CL prog Call parameter type does not match function > > signature! > > i128 8404990 > > float %8 = call i32 @_Z6islessff(float %7, i128 8404990) #3 Both > > operands to a binary operator are not of the same type! > > %18 = fmul ppc_fp128 0xM0000000000FF40060000000000000000, > float %17 > > Invalid insertelement operands! > > %20 = insertelement <4 x float> %19, ppc_fp128 %18, i32 1 Invalid > > insertelement operands! > > %26 = insertelement <4 x float> %25, ppc_fp128 > > 0xM00000000000000000000000000000000, i32 3 LLVM ERROR: Broken > function found, compilation aborted! > > I don't whether it can help but I have provided the full test program (opencl > + C) > in attachment (testclbug.tar.gz). > > I am than willing to share any build log, intermediate file, or information > about > the problematic systems that could help to understand the issue. I just don't > know which one could be relevant and helpful. > > Cheers, > > Nicolas > > > On 12/02/2014 02:52 AM, Song, Ruiling wrote: > > It is really strange, I tried you kernel on my Ivybridge machine. With > > beignet > 1.0 and LLVM/clang 3.5 It compiled successfully. > > Did you uninstall llvm/clang 3.4 before you install llvm/clang 3.5? > > Also you need to make sure you really use the attached file version. > > > > From your attached kernel code (isless(C, 0.5f);), it only could generate > > below > call instruction. > > %call = call i32 @_Z6islessff(float %4, float 5.000000e-01) But you > > get below call instruction, which is obviously wrong. > > float %8 = call i32 @_Z6islessff(float %7, i128 8404990) > > > >> -----Original Message----- > >> From: Beignet [mailto:beignet-boun...@lists.freedesktop.org] On > >> Behalf Of Nicolas Bourdaud > >> Sent: Tuesday, December 02, 2014 2:17 AM > >> To: Zhigang Gong; beignet@lists.freedesktop.org > >> Subject: Re: [Beignet] Program compilation crash on Ivy Bridge > >> > >> Hi! > >> > >> Two weeks ago, I mentioned a problem of OpenCL program crashing > >> during compilation on IvyBridge while compiling fine on Haswell. I > >> have been advised to test using llvm/clang 3.5. So I have compiled > >> beignet 1.0.0 using llvm/clang 3.5 but it turns out that I still get > >> the same problem on IvyBridge (Haswell still works): > >> > >>> OpenCL platform name: Intel Gen OCL Driver OpenCL version: OpenCL > >>> 1.2 beignet 1.0.0 OpenCL device: Intel(R) HD Graphics IvyBridge GT2 > >>> Running OpenCL 1.2 version going to compile CL prog Call parameter > >>> type does not match function signature! > >>> i128 8404990 > >>> float %8 = call i32 @_Z6islessff(float %7, i128 8404990) #3 Both > >>> operands to a binary operator are not of the same type! > >>> %18 = fmul ppc_fp128 0xM0000000000FF40060000000000000000, > >> float %17 > >>> Invalid insertelement operands! > >>> %20 = insertelement <4 x float> %19, ppc_fp128 %18, i32 1 Invalid > >>> insertelement operands! > >>> %26 = insertelement <4 x float> %25, ppc_fp128 > >>> 0xM00000000000000000000000000000000, i32 3 LLVM ERROR: Broken > >> function found, compilation aborted! > >> > >> > >> I don't know what I could check. So any lead is welcome. For > >> reference the offending CPUs are: > >> - Intel(R) Core(TM) i7-3770 CPU @ 3.40GHz > >> - Intel(R) Core(TM) i5-3470 CPU @ 3.20GHz I have put the program > >> provoking the crash in attachment (test.cl) > >> > >> Thanks in advance to anyone who could shed some light on these issues. > >> > >> Cheers, > >> > >> Nicolas > >> > >> > >> On 11/18/2014 10:00 AM, Nicolas Bourdaud wrote: > >>> Thanks! I will try this > >>> > >>> Nicolas > >>> > >>> On 18/11/2014 02:52, Zhigang Gong wrote: > >>>> The following error log is from LLVM internal. It's likely a > >>>> llvm/clang 3.4's > >> bug. > >>>> The kernel works fine with both beignet 0.9.3 and 1.0.0 by using > >> llvm/clang 3.5. > >>>> The recommended llvm version for beignet is LLVM 3.5. > >>>> > >>>> OpenCL platform name: Intel Gen OCL Driver OpenCL version: OpenCL > >>>> 1.2 beignet 0.9.3 OpenCL device: Intel(R) HD Graphics IvyBridge GT2 > >>>> Running OpenCL 1.2 version going to compile CL prog Call parameter > >>>> type does not match function signature! > >>>> i128 8404990 > >>>> float %8 = call i32 @_Z6islessff(float %7, i128 8404990) #7 Both > >>>> operands to a binary operator are not of the same type! > >>>> %18 = fmul ppc_fp128 0xM0000000000FF40060000000000000000, > >> float %17 > >>>> Invalid insertelement operands! > >>>> %20 = insertelement <4 x float> %19, ppc_fp128 %18, i32 1 Invalid > >>>> insertelement operands! > >>>> %26 = insertelement <4 x float> %25, ppc_fp128 > >>>> 0xM00000000000000000000000000000000, i32 3 Broken module found, > >> compilation aborted! > >>>> Aborted > >>>> > >>>>> -----Original Message----- > >>>>> From: Beignet [mailto:beignet-boun...@lists.freedesktop.org] On > >>>>> Behalf Of Nicolas Bourdaud > >>>>> Sent: Monday, November 17, 2014 7:27 PM > >>>>> To: beignet@lists.freedesktop.org > >>>>> Subject: [Beignet] Program compilation crash on Ivy Bridge > >>>>> > >>>>> Hi! > >>>>> > >>>>> In beignet 0.9.3, beignet aborts in the middle on compilation with > >>>>> compiling for ivybridge microarchitecture, while it compiles and > >>>>> runs > >> fine on Haswell. > >>>>> > >>>>> I have provided a test program (test.cl) that shows the problem > >>>>> and the generate log on haswell (haswell.log) and ivybridge > >>>>> (ivybridge.log). I have tested on different ivybridge CPU > >>>>> (HD4000, > >>>>> HD2500) and the issue is the same. I have also tested on different > >>>>> haswell CPU (only HD4600) and it always goes fine. > >>>>> > >>>>> For information beignet has been tested for: > >>>>> - ivybridge using linux 3.14 compiled with llvm 3.4. > >>>>> - haswell using linux 3.14 compiled with llvm 3.4. > >>>>> - haswell using linux 3.16 compiled with llvm 3.5 In all case, > >>>>> the program compiled and runs correctly with haswell, not with > ivybridge. > >>>>> > >>>>> I could give a try with beignet 1.0.0, but maybe you the problem > >>>>> would be obvious to you. > >>>>> > >>>>> Any idea about the problem? > >>>>> > >>>>> Cheers, > >>>>> > >>>>> Nicolas Bourdaud > >>>> > >>> > >>> > > > > _______________________________________________ > > Beignet mailing list > > Beignet@lists.freedesktop.org > > http://lists.freedesktop.org/mailman/listinfo/beignet > > _______________________________________________ Beignet mailing list Beignet@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/beignet
