Changelog for beignet-git-20180820T153219-45.122.x86_64.rpm :

* Sat Nov 17 2018 opensuse-packagingAATTopensuse.org- Update to version 20180820T153219:
* Allow creating out-of-order queues with clCreateCommandQueue
* Make in-order command queues actually be in-order
* Add preliminary LLVM 7 support
* Add LLVM 6.0 support
* More user-friendly \"type not supported\" errors
* Don\'t leak memory on long chains of events
* Enable Coffee Lake support
* Fix enabling of fp64 extension
* Ensure that DRM device uses the i915 driver
* Runtime: Remove X11 dri2 connection failed warning message.
* Mon Sep 28 2015 mardnhAATTgmx.de- Update to version 1.1.0.git1442978823.f48b4f6: + Fix clLinkProgram error. + GBE: fix ub1grf(nr, subnr) issue. + GBE: Minor refine uw1grf(nr, subnr). + should check the return value of cl_program_new. + return 32 could gain 0.2% performance on opencv optical flow case. + enable create image 2d from buffer in clCreateImage. + add utest for creating 2d image from buffer. + fix bswap bug. + add bswap64 for gen7/gen75 and gen8 seperately. + add bswap64 in utest. + Calculate appropriate timestamps for cl profile + Driver: fix the annoying \"Failed to release userptr...\" error message
* Sun Sep 20 2015 mardnhAATTgmx.de- Update to version 1.0.2.git1442563800.0ff7417: + Reorder GBE_BIN_GENERATER arguments. + GBE: Fix a bug in phicopy coaleasing. + runtime bug: brw GT3 devices reported to GT2. + runtime: fix a builtin-kernel related thread safe bug. + Remove deprecated function cl_context_get_static_kernel(). + Runtime: correct event and the wait events compare when check event. + Use a separate pattern for simd shuffle instead of binary pattern + GBE: Fix a bug in assigning image bti. + Bump master version to 1.2. + Update Release 1.1.0 NEWS.mdwn. + backend/src/backend: Handle -dump-opt-llvm=[PATH] + Utest: fix a builtin_powr_float fail when OCL_STRICT_CONFORMANCE=0. + GBE: Fix build error. + runtime: add Broadwell deviceID 0x162B + runtime: add detailed broadwell device name. + Use Integer for U32/S32 Immediate load. + libocl: fix degrees function precision issue. + Update last event status in clFinish. + runtime: always try to update event status in clGetEventProfilingInfo(). + fix a powr function issue in cpu compiler math + fix utest fail. + GBE/IR: add collectInsnNum to collect block instruction number. + generate sub_group_id inside kernel instead of payload + Set proper Vendor ID + backend: Handle (but ignore) -dump-opt-asm=[PATH]. + backend: Move ASM printing to a helper function. + backend: Convert outputAssembly to C file I/O. + backend, src: Add ASM file name to gbe_program_new_from_llvm + backend: Add ASM file name to GenProgram object. + backend: Add ASM file name to GenContext object. + backend: Turn on ASM dump. + fix issue when build against llvm3.3 + GBE/PRINTF: store variable instead of pointer in \"slots\". + backend/src/backend: Handle -dump-opt-llvm=[PATH] in clCompileProgram and clBuildProgram OpenCL API + remove GBE_CURBE_STACK_POINTER in payload + correct simd width when dst of simd_shuffle is scalar + GBE: fix the broken image_1d_buffer write. + utests: refine image 1d buffer test case. + GBE: one minor bug in OP_SIMD_XXX. + GBE: a potential bug in instruction scheduling. + Fix clGetKernelArgInfo fail on piglit + Runtime: Add NULL pointer check in clGetKernelArgInfo + Utest: Add -cl-kernel-arg-info to the utest test_get_arg_info + utests: Added unit tests to test LLVM and ASM dump generation. + GBE: Use addRemappedFile to avoid creating temporary cl source file. + GBE: add check dumpASMFileName.empty() + GBE: fix build error with LLVM 3.5 and previous version. + Fix piglit clLinkProgram fail. + Don\'t use cl_buffer_get_subdata in clEnqueueReadBuffer.
* Wed Jul 15 2015 mardnhAATTgmx.de- Update to version 1.0.2.git1436951394.55c8538: + Fix comparison in if() check. + fix utest bug. + Runtime: Add default extension for platforms before BDW. + libocl: Add macro define for fp16. + GBE: optimize phi elimination. + runtime: Need to separate atomic in L3 test and SLM test in self_test(). + Fixed a thread safe bug. + Need to check eventWaitList in clEnqueueNDRangeKernel. + BDW: override GenContext::patchBranches. + correct ULP value in utests + GBE/IMM: Temporarily allow integer type in getFloatValue()
* Fri Jul 10 2015 mardnhAATTgmx.de- Update to version 1.0.2.git1436193612.9d4fd34: + Fix a indirect argument load bug. + GBE: Fix a typo that cause gpu hang. + GBE: add offsetID in SelectionVector. + GBE: fix one potential register spilling bug. + GBE: Support storing/loading pointers to/from private array + reimplement structurize algorithm. + fix global variable out of boundary writing in libocl. + Fix error in CMakeLists.txt of examples. + Add example to show v4l2 buffer sharing with extension clGetMemObjectFdIntel. + Add document to describe the detials of v4l2 buffer sharing. + enable CL_RG + CL_UNORM_INT8 for image + Turn on OCL_STRICT_CONFORMANCE as default. + Use NP2 stack size to avoid cache line conflict. + Use the Byte Gather after HSW when read byte/shor. + use self test to determine enable/or disable atomics in L3 for HSW. + don\'t merge serial blocks with barrier. + Backend: Add half float as a new type. + Backend: Add half float support for immediate. + backend: Add half float type into the instruction. + Backend: Add half float support in gen backend. + Backend: Add half to insn selection. + Backend: Add half float ASM output support. + libocl: Enable fp16 extension in the header. + libocl: Add half builtin functions for relational module. + libocl: Add half builtin functions to math module. + libocl: Add half builtin functions for common module. + libocl: Add the builtin xxx_as functions for half. + libocl: Add builtin convert functions for half. + libocl: Add vector generation for half type. + backend: Add convert_sat functions for half in the gen_backend. + Backend: Add support for half\'s div and rem. + backend: Add conversion support between half and other type. + libocl: Add all the half defines for script generation. + utest: Add test cases for half. + runtime: Add fp16 extension to BDW later platform. + Set the loop unroll\'s threshold to 1024. + enable CL_UNSIGNED_INT8 for CL_RG to fix regression + structuralize bug fix: check the loop successor before merge. + runtime: Use cl_get_platform_default to replace global value. + runtime: Add cl device\'s standalone extension. + Backend: Add half support for CHV and SKL.
* Fri May 29 2015 mardnhAATTgmx.de- removed patch beignet-icd-name-fix.diff (included upstream)- Update to version 1.0.2.git1432616628.b03a780: + runtime: fix a conformance bug in cl_get_kernel_arg_info. + Generate NAN for UNDEF value in printf parser. + Build: fix the beignet icd name when CMAKE_INSTALL_FULL_LIBDIR is undefined. + Add driver callback for updating device info + Query the driver directly for compute units and subslice + BUGFIX: Prohibit \'make package\' from doing system install of ICD vendor file + BDW: Refine I64HADD and I64RHADD. + Use matching versions of clang/llvm and libclang/libllvm + FindLLVM: allow LLVM/Clang 3.6 + Don\'t crash if device inaccessible + Fix: (v3) Event callback that were not executed when command was already CL_COMPLETE + thread safety for callbacks + update document. + GBE: correct some temporary virtual register\'s simdWidth. + GBE: avoid to use the GenRegister::xxxgrf(simdWidth,xxx). + runtime: Enhance the error handling when flush gpgpu command queue. + strip unsupported attributes and calling conventions. + GBE: fix safe type definition. + GBE: extend registers/tuples/immediates to 32bit wide. + GBE: extend backend label to 32 bit. + GBE: don\'t type cast register/labelindex to integer. + GBE: Extend front label ip to 32 bit on demand. + GBE: Use actual bti information to determine a pointer\'s addressspace. + GBE: refine error handling for private libva buffer sharing extension. + GBE: correct the instruction replacement logic in scalarize pass. + GBE: fix an potential assertion in constant expanding pass. + GBE: fix a bug in byte scatter write. + BDW: Refine unpacked_ud in the gen8_context.cpp. + runtime: don\'t try to open nonexistent render nodes or device files. + utests: don\'t continue to run any case when fail to initialize device. + Doc: add a command to install dependencies. + GBE: should initialize useDWLabel to false by default. + Doc: update cmd parser issue for HSW platforms. + Kill the A0 cache in GenContext. + Doc: update known issue for the store/load pointer issue. + add simd level function __gen_ocl_get_simd_size + add utest for __gen_ocl_get_simd_size + add simd level function __gen_ocl_get_simd_id + add utest for __gen_ocl_get_simd_id + utest_pow: don\'t fail on declared lack of denormals. + Make tgamma meet the accuracy standard. + Allow building with Python 3 + utests: fix test case builtin_tgamma. + CHV: Add cherryview support in the runtime. + Chv: Add chv backend support. + Optimization of clEnqueueCopyBufferToImage for 16 aligned case. + add benckmark for copy data from buffer to image. + Revert \"CL/Driver: enable atomics in L3 for HSW.\" + Fix two argument lowering bug. + add introduction to build Beignet with yocto + GBE: fix LOD initialization for typed write instruction. + rename __gen_ocl_get_simd_id/size to get_sub_group_id/size + add sub group functions intel_sub_group_shuffle + rename __gen_ocl_simd_any/all to sub_group_any/all + add utest for intel_sub_group_shuffle + Add Indirect struct argument read support. + Add stuct argument indirect load test. + add environment variable OCL_OUTPUT_KERNEL_SOURCE. + correct the src output of alu3 when OCL_OUTPUT_ASM=1 + GBE: make all memory operation share same bti dependency. + GBE: Fix the immediate data type + GBE: remove unnecessary assert + CHV: Fix a chv long convert bug. + Fix a indirect register bug. + Add a sanity test in clGetDeviceIDs + Docs: update/clarify Haswell issues + Remove some LGPL incompatible code. + gbe: Implement a new BTI solution to support dynamic bti + build: use EXECUTE_PROCESS to replace the deprecated EXEC_PROGRAM.
* Tue Mar 17 2015 mardnhAATTgmx.de- Update to version 1.0.2.git1426575594.413a725: + GBE: fix build error for llvm 3.6. + GBE: fix build error for LLVM 3.4/3.3. + Add benchmark of clEnqueueCopyImageToBuffer(copy 2d image to buffer). + Optimization of clEnqueueCopyImageToBuffer for 16 aligned case. + Correct the error llvm link msg copy in function genProgramLinkFromLLVM. + libocl: define NULL to zero + libocl: Directly scalarize built-in with vector input. + GBE: unify element type before insertelement in legalize pass. + Crash when hardware inaccessible + build: use AATTBEIGNET_INSTALL_DIRAATT for the icd file. + GBE: expand constant expressions in constant vector + GBE: remove constant expression handling code in gen writer pass. + GBE: Support unaligned load/store of dword/qword in GenIR. + Enable multiarch (32/64-bit co-installation) + GBE: Fix fast-math issue under llvm 3.6. + GBE: remove the unecessary type check for SEL instructio. + GBE: support compare two bool variables. + GBE: add fastcall support. + Build: use -Bsymbolic to fix conflicts with other LLVM users. + Fix llvm3.6 build error. + GBE: add a new incompatible compile option -cl-finite-math-only. + enable cl_khr_spir extension to build and run from SPIR binary. + change the workitem related api to OVERLOABABLE. + SPIR binary support for printf function. + add utest for load spir binary. + Backend: Add the indirect fields and functions for gen register. + Backend: Add functions to set a0 register. + Backend: Correct indirect mode encoder setting for Gen7. + Backend: Correct indirect mode encoder setting for Gen8. + Backend: Handle the bswap using indirect mode access. + Add a0 setting and bswap logic for GEN8 + Backend: Delete bswap logic in the llvm_to_gen stage. + Modify the utest case for bswap. + GBE: Only emit multiply when immediate is not one. + Backend: Add the logic to handle uniform src for BSwap Gen8. + Backend: Fix errors in disasm for indirect instruction Gen8. + replace fabs with llvm intrinsic. + replace rndz with llvm intrinsic. + replace rnde with llvm intrinsic. + replace rndu with llvm intrinsic. + replace rndd with llvm intrinsic. + replace mad with llvm intrinsic. + replace pow with llvm intrinsic. + Revert \"libocl: using mad() to implement dot()\" + reset the SPIR target datalayout. + only support spir extension for beignet build with llvm 3.5 or later. + simple return if spir extension not supported. + GBE: fix an image related bugs. + Build: set 3.5 as the stable LLVM version for beignet. + Docs: update team information. + Bump to 1.1 + Add LLVM_INCLUDE_DIR to CMakeList of src. + strip PointerCast for call instructions before use. + correct env var to output llvm IR
* Thu Feb 12 2015 mardnhAATTgmx.de- Update to version 1.0.1.git1423738879.2259a37: + Add submodule libva for examples. + Add example to show libva buffer sharing with extension clCreateImageFromLibvaIntel. + Add document to describe the detials of libva buffer sharing. + update document. + Remove useless old legalize related files. + GBE: Need to check invalid register to determine the actual src num. + GBE: fix the hacky usage of invalid register. + runtime: fix a potential null pointer dereference. + runtime: don\'t free the host_ptr for a subbuffer. + SKL: fix some 3D and 2D array image fail. + SKL: fix some failed piglit tests, caused by read constant error. + Fix the bug of ulong mad sat + Correct the bit fields error for indirect address of Gen8 + SKL: Fix opencv perf hang. + prefer newer llvm versions over 3.3 + remove unsafe define -D__$(USER)__ + GBE: Load/store should use same address space as before. + GBE: Import PromoteIntegers pass from pNaCl + Change the KB and MB define to enum. + libocl: refine implementation of sign(). + GBE: We need use exiting block here. + Use llvm-c\'s LLVMLinkModules instead of llvm::Linker::LinkModules. + Add llvm3.6 build support. + Remove useless llvm head file FindUsedTypes.h.
* Fri Feb 06 2015 mardnhAATTgmx.de- Update to version 1.0.1.git1423201207.66a4850: + update document. + Import the native long type of ul1 ul8 and ul16 + Add long type support for disasm. + Add long imm value in gen8 instruction. + Add unpacked ud and unpacked uw for long type. + Add the long unpacked ud uw into the instruction schedule consideration + Modify the split logic in encoder + Add the u64 imm type in register + Add long support flag into gen selection + Modify the load IMM 64 function. + Disasm supports to print long imm value in instruction. + Add functions for conversion between native and fake long. + Overload the READ64 and WRITE64 function for Gen8 + Add the canHandleLong virtual function into gen encoder + Overload all the alu1 and alu1withtemp functions. + Overload all the simple binary functions. + Modify the convert logic in gen selection. + Modify the fake dst register for CMP in register allocation. + Overload I64MUL function. + Overload the i64MULHI function. + Overload the I64HADD function. + Overload I64RHADD function. + Overload the I64MADSAT function. + Overlaod I64 Div and Rem function. + Modify the cmp and sel for I64 and cleanup all virtual functions. + Add test case for long mul_sat and mul_hi + Add test case for i64 div and rem. + Add long NOT test case. + Add PackLong and UnpackLong functions + Modify the bitcast to support native long + Add the logic for UPSAMPLE_LONG + Add test case for long bitcast. + GBE: fix an ACC register related instruction scheduling bug + Add the missing LOAD_INT64_IMM virtual function for GEN8. + Fix bug for bitcast test case because of long type. + GBE: fix popcount bugs. + correct the cache line size to be 64 + loose the alignment limitation for host_ptr of CL_MEM_USE_HOST_PTR + update utest to loose userptr limitation + add builtin function atan2pi and ldexp + Add the check for src and dst span different registers. + GBE: add GEN_TYPE_HF to getTypeSize. + reimplement the LZD instruction in backend. + libocl: reimplement clz with lzd instruction instead of fbh. + fix clz utest issue. + Fix the long bitcast post schedule bug. + Add the logic for pack/unpack long for scalar. + check the predication in case of endless loop. + Correct the wrong type annotation for long in disasm. + Fix bug for scalar long conversion. + libocl: Refine char/short abs() implementation. + libocl: refine implementation of abs_diff() + libocl: using mad() to implement dot() + GBE: Only add non-zero offset in gep lowering pass. + libocl: refine length() implementation. + libocl: refine implementation of normalize(). + libocl: refine implementation of logb(). + SKL: Add skl pci ids and device. + SKL: enable skl device. + SKL: Use TILE_Y as default TILING mode in skl. + SKL: correct the pipe control struct. + SKL: Add the function gen9\' intel_build_idrt. + SKL: add skl select_pipeline and cache_control functions. + SKL: Add function intel_gpgpu_bind_image_gen9. + SKL: fix skl LD fail. + remove the libFunc interface. + Backend: fix one bug of long mad_sat. + Re-format the asm print for long imm + GBE: Import constantexpr lower pass from pNaCl + GBE: expand large integer instructions + GBE: Fix a build error against llvm release version + Fix a bug of 1d image array test case. + Backend: Fix one bug of printf because of ir reorder. + Refine benchmark output. + Refine copy_buf benchmark and rename the file. + replace sin/cos with llvm intrinsic. + replace sqrt with llvm intrinsic. + replace log with llvm intrinsic. + replace exp with llvm intrinsic. + Return error, don\'t crash, on allocation failure + Implement 1D/2D image array related cl_mem_kernel_copy_image in cl way instead of cpu way. + GBE: Fix a bug in legalize pass.
* Mon Jan 19 2015 mardnhAATTgmx.de- Update to version 1.0.1.git1421375489.786da41: + utests: reduce test count. + runtime: tweak max memory allocation size. + runtime: fix max work group size for IVBGT1. + add edge case detection for powr in utests + Don\'t check some edge condtion in non-strict mode. + Remove obsolete MI_FLUSH + do not include llvm/clang headers for libgbeinterp + change Immediate::operator= from private to public + do not use C++11 features inside libgbeinterp + fix utest build for some old gcc version + libocl: remove useless code. + CL/Driver: enable atomics in L3 for HSW. + libocl: Reimplement trigonometric functions. + refine gbe_bin_generater usage to add -t option + CL/Driver: quick fix regression caused by remove MI_FLUSH. + remove useless dependency libocl + Fix loop condition of PrintfSet constructor. + utests: skip one test when it fail to open XDisplay. + CL/Driver/HSW: Convert L3 cycle for texture to uncachable. + Add read buffer/image benchmark. + add option BUILD_STANDALONE_GBE_COMPILER to build static compiler + add CMake option USE_STANDALONE_GBE_COMPILER and STANDALONE_GBE_COMPILER_DIR + only build tests that do not need compiler when standalone compiler is provided + add howto for old gcc version + Fix the printf buffer size bug. + add LZD IR instruction. + add clz(count leading zero) utest. + add llvm intrinsic call translate. + fix llvm.trunc.float instruction bug. + fix the wrong implementation of popcount.
* Tue Dec 30 2014 mardnhAATTgmx.de- Update to version 1.0.0.git1419831858.c724546: + GBE: use sr0.1\'s SLM Offset to eliminate the software SLM offset for HSW. + GBE: remove software maintained SLM offset related code. + libocl: implement high precision pown() + fix min_max_read_image_args and min_max_parameter_size issue. + add collectImageArgs to handle image count limitations. + replace hash_map with map + Remove deprecated fulsim code + Add aub dump support + Use libdrm interface to get device id + Separate flush and invalidate in function intel_gpgpu_pipe_control. + Fix PrintfState copying.
* Tue Dec 23 2014 mardnhAATTgmx.de- Update to version 1.0.0.git1419319480.a336aaf: + Change the IVB/HSW\'s max_work_group_size to 512, and BYT to 256. + GBE: disable spill register under simd16 mode. + utests: fix work group size issue in compiler_fill_image_2d_array. + utests: fix a typo in test cases. + GBE: don\'t split instruction for some special case. + GBE: Place loop exits after loop blocks when sorting basic blocks. + utests: reduce work group size to 256 to satisfy BYT platform. + fix issue to create cl image from libva with non-zero offset + add test for clCreateImageFromLibvaIntel + utests: fix indent in CMakeLists.txt + fix issue to pass utest of runtime_climage_from_boname for BDW + GBE: Output CFG of Gen IR to dot file. + GBE: Fix bug with negative constant GEP index. + utests: Add one case to test negative index array access. + Fix based on piglit OpenCL falied case (cl-api-compile-program). + GBE: fix a regression caused by the negative index handling patch. + GBE: optimize GEP constant offset calculation. + fix dnetc overflow issue. + fix bswap implementation issue. + clean code, the logic is already at the beginning of function + add test of cl_mem_use_host_ptr into benchmark + refine utest of cl_mem_use_host_ptr + enable CL_MEM_ALLOC_HOST_PTR with user_ptr to avoid copy between GPU/CPU + add utest of CL_MEM_ALLOC_HOST_PTR + utests: make utests maths ULP values consistent with specification + GBE: Re-implement BTI logic in backend + Change CL_DEVICE_PREFERRED_VECTOR_WIDTH_CHAR from 8 to 16. + GBE: support const private array initialization. + utests: Add const private array initialization test. + GBE: Fix a disassembly bug. + GBE: Fix the printf issue caused by new bti implementation + refine bswap utest to cover nsetc fail cases. + Fix a makefile bug for gcc is not the default compiler + refine overflow utest to cover nsetc fail cases. + libocl: Fix precision of builtin tanpi. + libocl: Move spec required macro to header file. + GBE: remove useless code. + GBE: eliminate duplicate GEP handling logic. + GBE: Add constant pointer in the memcpy intrinsic. + CL: Don\'t find mesa source code. + GBE: Add some missing constant expression cases. + Update optimization tips. + Fix delete operator using. + change the utest summary code + disable overflow utest test before llvm-3.5 + GBE: don\'t always treat a multiple destination instruction as root. + Refactor all image builtin functions. + GBE: switch to use CLANG native image types. + GBE: switch to CLANG native sampler_t. + GBE: remove some image1d_buffer related builtin functions. + GBE/CL: use 2D image to implement large image1D_buffer. + GBE: code cleanup. + add half math function support. + libocl: Improve precision of pow/powr. + libocl: Imporve precision of exp() + libocl: Flush denorm input into zero in rootn() + libocl: flush denorm into zero in ldexp() + libocl: Correctly handle -inf in exp10. + libocl: flush denorm to zero in remquo() + fix max_parameter_size not correct on x86 platforms. + GBE: fix an image regression.
* Fri Nov 21 2014 mardnhAATTgmx.de- Update to version 1.0.0.git1416556620.9c18f05: + Revert \"add test for clCreateImageFromLibvaIntel\" + Revert \"fix issue to create cl image from libva with non-zero offset\" + docs: update mixed_buffer_pointer document. + utests: remove all shader toy test cases. + License: adjust all license version to LGPL v2.1+. + GBE: fix relocatable issue for pch file. + Revert \"BDW: Change the default tiling mode to TILING_Y on BDW.\" + GBE: fix one double related bugs for post register scheduling. + update some documents. + runtime: fix one bug in BDW image. + Update documents. + runtime: refine version handling. + runtime: fix bug in cl_enqueue_read_buffer. + runtime: disable userptr due to random fail. + Fix the bug of multi-thread crash + GBE: work around error reporting for unresolved symbols + Bump to 1.0.0. + Remove patch version on master branch. + Change the IVB/HSW L3 SQC credit setting. + reuse the loop info from llvm. + re-enable userptr with fix: CPU access after GPU finishes the rendering + Fix NO_TILING alignment bug. + add the reduced self loop node detection. + BDW: Change the default tiling mode to TILING_Y on BDW. + Fix the opencv_test_core/OCL_Arithm random segment fault.
* Mon Nov 10 2014 mardnhAATTgmx.de- Update to version 0.9.3.git1415604223.48e2da8: + add llvm Intrinsic call support. + add utest compiler_overflow for llvm intrinsic function. + Fix the compare_image_2d_and_1d_array test case bug + Refine the intel gpgpu delete. + Fix a size assert when setup bti. + BDW: Fix bwd 32
*32 scalar multiplication bug. + IVB/HSW/BYT: Revert the Dynamic state Base Addr and relative buffers address setting. + Remove intel_gpgpu_check_binded_buf_address() + fix some typos + use env to set environment variables for GBE_BIN_GENERATER + enable llvm intrinsic call usub_with_overflow funtion. + add utest for llvm intrinsic call usub_with_overflow funtion. + GBE: workaround register allocation fail caused by custom loop unroll. + GBE: Fix live range for temporary register in replaceReg + enable llvm intrinsic call bswap function. + add utest function bswap. + utest: change the box_blur_image to be identical to box_blur. + utests: replace the nodistriutable picture. + GBE: fix disassembly bug. + fix bswap kernel function type issue. + add building dependency GIT_SHA1 + utest: add new test that trigger an assignment operation bug in if. + GBE: fix a bool handling bug when SEL on a uniform bool variable. + GBE: Support more instructions for constant expression handling. + GBE: fix bug in pow()/pown(). + GBE: fix bug in erf()/erfc(). + GBE: fix bug in tgamma(). + utests: fix bugs in builtin_pow(). + utests: fix bugs in builtin_tgamma(). + GBE: remove useless debug info. + GBE: Fix kernel argument size for vector3 + utests: add a test to trigger cl_float3 bug in clSetKernelArg. + BDW: Set the URB/REST size to 384K/384K when SLM disable. + support CL_MEM_USE_HOST_PTR with userptr for cl buffer + enable CL_DEVICE_HOST_UNIFIED_MEMORY when userptr is supported + add test for cl buffer created with CL_MEM_USE_HOST_PTR + GBE: Fix a bitcast from float vector to wide interger issue in legalize pass. + GBE: Do topological sorting of basicblocks. + fix piglit clCreateProgramWithBinary fail. + fix a bug in clCompileProgram(). + fix issue to create cl image from libva with non-zero offset + add test for clCreateImageFromLibvaIntel + BDW: Change the default tiling mode to TILING_Y on BDW. + use posix_memalign instead of aligned_alloc to be more compatible
* Mon Oct 27 2014 mardnhAATTgmx.de- Update to version 0.9.3.git1414386114.6eb5d67: + use global flag 0.0 to control unstructured simple block. + Fix the bug of 1D array slice pitch + Add the test case for image 1d array fill + Add the test case for image 2d array fill + GBE: add Selection instruction handler at legalize pass. + GBE: increase maximum src/dst operands to 32. + GBE: add basic PHINode support in legalize pass. + add beignet GIT_HAL1 if there is .git directory + GBE: fix regression caused by simple block optimization. + Add the disasm support for Gen8 + create GIT_SHA1 without any dependency + GBE: handle dead loop BBs in liveness analysis. + Make use of write enable flag for mem bo map + Clear batch buffer pointer after unmap + Use pread/pwrite for buffer enqueue read/write + Fix AUX buffer for page alignment + GBE: set default address space to -1 to avoid incorrect unroll hint. + GBE: fix a wrong type of cl_device_info.
* Mon Oct 20 2014 mardnhAATTgmx.de- Update to version 0.9.3.git1413776840.8fe98a5: + GBE: disable custom loop unroll for LLVM 3.3/3.4.
* Sat Oct 18 2014 mardnhAATTgmx.de- add ncurses-devel to BuildRequires: + linker was complainining about missing -ltinfo- fix build error: + set CMAKE_BUILD_TYPE=Release
* Fri Oct 17 2014 mardnhAATTgmx.de- Update to version 0.9.3.git1413529844.0ccfdf5: + Fix a HSW regression. + Refine the shared function ID define. + Add the libdrm version check. + Fix memcpy and memset bug. + Let the failure of intel_drm lib\'s check as a FATAL_ERROR + add opencl-1.2 builtin function popcount. + add utest popcount for all types. + Fix HSW thread_n <= 64 assert. + Fit the printf bug in loop + GBE: fix a bug in legalize pass. + Fix a HSW constant buffer regression. + BDW: Change BDW\'s max work group size to 512. + BDW: Fix load/store half error. + BDW: Also need set Shader Channel Select for constant buffer in BDW. + Fix a upsample regression. + GBE: do intrinsics lowering pass earlier. + Fix a HSW regression. + Refine the the error handling in function cl_command_queue_ND_range_gen7. + GBE: fix a legalize pass bug when bitcast wide integer to incompaitble vector. + GBE: Add a customized loop unrolling handling mechanism.