Accepting request 691606 from home:mnhauke:branches:science
- Update to version 1.3
* Support for Clang/LLVM 8.0.
Bug Fixes:
* Fixed kernel debug symbol generation.
* HSA: fix kernel caching.
* Fix clCreateImage doesn't fail with unsupported image type.
* Fix handle non-kernel functions with barriers properly.
* Fix Unable to build pocl with CUDA support with LLVM 7 and host
GCC 8.2.
* Fix image format/size handling with multiple devices in context.
* Fix padding issue with context arrays that manifested as unaligned
access errors after autovectorization.
Notable Internal Changes
* Add group ids as hidden kernel arguments instead of digging
them up from the context struct.
* Ability to generate the final binary via separate assembly text +
assembler call. Useful for supporting LLVM targets without direct
binary emission support.
* Use Clang's Driver API for launching the final linkage step. This
way we utilize the toolchain registry with correct linkage steps
required for the target at hand.
* Add 'device_aux_functions' to the driver layer attributes. This
can be used to retain device-specific functions required by the
target across the pruning of unused globals.
* The "default kernels" hack which was used to store kernel metadata,
has been removed. Kernel metadata are now stored only once, in
cl_program struct; every new cl_kernel structs holds only a pointer.
* Major 'pthread' CPU driver cleanup.
* Major Workgroup.cc cleanup.
- Remove reproducible.patch (fixed upstream)
OBS-URL: https://build.opensuse.org/request/show/691606
OBS-URL: https://build.opensuse.org/package/show/science/pocl?expand=0&rev=45
This commit is contained in:
Martin Hauke
Git OBS Bridge
parent
39fe8c4c9b
commit
cb98fe7909
@ -1,3 +0,0 @@
|
||||
version https://git-lfs.github.com/spec/v1
|
||||
oid sha256:0c43e68f336892f3a64cba19beb99d9212f529bedb77f7879c0331450b982d46
|
||||
size 1603374
|
||||
3
pocl-1.3.tar.gz
Normal file
3
pocl-1.3.tar.gz
Normal file
@ -0,0 +1,3 @@
|
||||
version https://git-lfs.github.com/spec/v1
|
||||
oid sha256:6527e3f47fab7c21e96bc757c4ae3303901f35e23f64642d6da5cc4c4fcc915a
|
||||
size 1414807
|
||||
34
pocl.changes
34
pocl.changes
@ -1,3 +1,37 @@
|
||||
-------------------------------------------------------------------
|
||||
Thu Apr 4 19:21:26 UTC 2019 - Martin Hauke <mardnh@gmx.de>
|
||||
|
||||
- Update to version 1.3
|
||||
* Support for Clang/LLVM 8.0.
|
||||
Bug Fixes:
|
||||
* Fixed kernel debug symbol generation.
|
||||
* HSA: fix kernel caching.
|
||||
* Fix clCreateImage doesn't fail with unsupported image type.
|
||||
* Fix handle non-kernel functions with barriers properly.
|
||||
* Fix Unable to build pocl with CUDA support with LLVM 7 and host
|
||||
GCC 8.2.
|
||||
* Fix image format/size handling with multiple devices in context.
|
||||
* Fix padding issue with context arrays that manifested as unaligned
|
||||
access errors after autovectorization.
|
||||
Notable Internal Changes
|
||||
* Add group ids as hidden kernel arguments instead of digging
|
||||
them up from the context struct.
|
||||
* Ability to generate the final binary via separate assembly text +
|
||||
assembler call. Useful for supporting LLVM targets without direct
|
||||
binary emission support.
|
||||
* Use Clang's Driver API for launching the final linkage step. This
|
||||
way we utilize the toolchain registry with correct linkage steps
|
||||
required for the target at hand.
|
||||
* Add 'device_aux_functions' to the driver layer attributes. This
|
||||
can be used to retain device-specific functions required by the
|
||||
target across the pruning of unused globals.
|
||||
* The "default kernels" hack which was used to store kernel metadata,
|
||||
has been removed. Kernel metadata are now stored only once, in
|
||||
cl_program struct; every new cl_kernel structs holds only a pointer.
|
||||
* Major 'pthread' CPU driver cleanup.
|
||||
* Major Workgroup.cc cleanup.
|
||||
- Remove reproducible.patch (fixed upstream)
|
||||
|
||||
-------------------------------------------------------------------
|
||||
Wed Oct 31 12:13:35 UTC 2018 - Bernhard Wiedemann <bwiedemann@suse.com>
|
||||
|
||||
|
||||
@ -1,7 +1,7 @@
|
||||
#
|
||||
# spec file for package pocl
|
||||
#
|
||||
# Copyright (c) 2018 SUSE LINUX GmbH, Nuernberg, Germany.
|
||||
# Copyright (c) 2019 SUSE LINUX GmbH, Nuernberg, Germany.
|
||||
# Copyright (c) 2014 Guillaume GARDET <guillaume@opensuse.org>
|
||||
#
|
||||
# All modifications and additions to the file contributed by third parties
|
||||
@ -17,9 +17,9 @@
|
||||
#
|
||||
|
||||
|
||||
%define sover 2.2.0
|
||||
%define sover 2.3.0
|
||||
Name: pocl
|
||||
Version: 1.2
|
||||
Version: 1.3
|
||||
Release: 0
|
||||
Summary: Portable Computing Language - an OpenCL implementation
|
||||
# The whole code is under MIT
|
||||
@ -30,7 +30,6 @@ Group: Development/Tools/Other
|
||||
URL: http://portablecl.org/
|
||||
Source0: https://github.com/pocl/pocl/archive/v%{version}.tar.gz#/%{name}-%{version}.tar.gz
|
||||
Source99: pocl-rpmlintrc
|
||||
Patch0: reproducible.patch
|
||||
BuildRequires: clang < 8
|
||||
BuildRequires: clang-devel < 8
|
||||
BuildRequires: cmake
|
||||
@ -95,7 +94,6 @@ This subpackage provides the development files needed for pocl.
|
||||
|
||||
%prep
|
||||
%setup -q
|
||||
%patch0 -p1
|
||||
|
||||
%build
|
||||
%define __builder ninja
|
||||
|
||||
@ -1,80 +0,0 @@
|
||||
https://bugzilla.opensuse.org/show_bug.cgi?id=1110722
|
||||
commit e99f4ca98bee7c59d8869a1b1d0aa764ffd5ec28
|
||||
Author: Michal Babej <michal.babej@tut.fi>
|
||||
Date: Thu Oct 11 13:16:43 2018 +0300
|
||||
|
||||
Fix PR #667
|
||||
|
||||
diff --git a/lib/CL/devices/basic/basic.c b/lib/CL/devices/basic/basic.c
|
||||
index a64680f0..048f0008 100644
|
||||
--- a/lib/CL/devices/basic/basic.c
|
||||
+++ b/lib/CL/devices/basic/basic.c
|
||||
@@ -354,9 +354,6 @@ pocl_init_cpu_device_infos (cl_device_id dev)
|
||||
dev->llvm_cpu = get_llvm_cpu_name ();
|
||||
#endif
|
||||
|
||||
- if(dev->llvm_cpu && (!strcmp(dev->llvm_cpu, "(unknown)")))
|
||||
- dev->llvm_cpu = OCL_KERNEL_TARGET_CPU;
|
||||
-
|
||||
#else /* No compiler, no CPU info */
|
||||
dev->llvm_cpu = NULL;
|
||||
dev->llvm_target_triplet = "";
|
||||
diff --git a/lib/CL/pocl_llvm_build.cc b/lib/CL/pocl_llvm_build.cc
|
||||
index fda6e3a2..7b7001f8 100644
|
||||
--- a/lib/CL/pocl_llvm_build.cc
|
||||
+++ b/lib/CL/pocl_llvm_build.cc
|
||||
@@ -839,16 +839,14 @@ static llvm::Module* getKernelLibrary(cl_device_id device)
|
||||
kernellib += device->llvm_target_triplet;
|
||||
if (is_host) {
|
||||
kernellib += '-';
|
||||
+#ifdef KERNELLIB_HOST_DISTRO_VARIANTS
|
||||
+ kernellib += getX86KernelLibName();
|
||||
+#else
|
||||
kernellib_fallback = kernellib;
|
||||
kernellib_fallback += OCL_KERNEL_TARGET_CPU;
|
||||
kernellib_fallback += ".bc";
|
||||
-#ifdef KERNELLIB_HOST_DISTRO_VARIANTS
|
||||
- if (triple.getArch() == Triple::x86_64 ||
|
||||
- triple.getArch() == Triple::x86)
|
||||
- kernellib += getX86KernelLibName();
|
||||
- else
|
||||
+ kernellib += device->llvm_cpu;
|
||||
#endif
|
||||
- kernellib += device->llvm_cpu;
|
||||
}
|
||||
kernellib += ".bc";
|
||||
|
||||
@@ -861,6 +859,7 @@ static llvm::Module* getKernelLibrary(cl_device_id device)
|
||||
}
|
||||
else
|
||||
{
|
||||
+#ifndef KERNELLIB_HOST_DISTRO_VARIANTS
|
||||
if (is_host && pocl_exists(kernellib_fallback.c_str()))
|
||||
{
|
||||
POCL_MSG_WARN("Using fallback %s as the built-in lib.\n",
|
||||
@@ -868,6 +867,7 @@ static llvm::Module* getKernelLibrary(cl_device_id device)
|
||||
lib = parseModuleIR(kernellib_fallback.c_str());
|
||||
}
|
||||
else
|
||||
+#endif
|
||||
POCL_ABORT("Kernel library file %s doesn't exist.\n", kernellib.c_str());
|
||||
}
|
||||
assert (lib != NULL);
|
||||
diff --git a/lib/CL/pocl_llvm_utils.cc b/lib/CL/pocl_llvm_utils.cc
|
||||
index feb88424..163304fe 100644
|
||||
--- a/lib/CL/pocl_llvm_utils.cc
|
||||
+++ b/lib/CL/pocl_llvm_utils.cc
|
||||
@@ -134,11 +134,13 @@ get_llvm_cpu_name ()
|
||||
}
|
||||
#endif
|
||||
|
||||
+#ifndef KERNELLIB_HOST_DISTRO_VARIANTS
|
||||
if (r.str() == "generic") {
|
||||
POCL_MSG_WARN("LLVM does not recognize your cpu, trying to use "
|
||||
OCL_KERNEL_TARGET_CPU " for -target-cpu\n");
|
||||
r = llvm::StringRef(OCL_KERNEL_TARGET_CPU);
|
||||
}
|
||||
+#endif
|
||||
|
||||
assert(r.size() > 0);
|
||||
char *cpu_name = (char *)malloc(r.size() + 1);
|
||||
Loading…
Reference in New Issue
Block a user