[NVPTX] Fix issues in ptxas integration to LIT tests

[asavonic]

Summary

1. Fixed a typo in PTXAS_EXECUTABLE CMake variable (PXTAS -> PTXAS).

2. Version check was implemented incorrectly, now version (major, minor) is converted to int for comparison.

3. ptxas -arch argument was missing in 2 tests.

[tra]

LGTM in general.

llvm/test/CodeGen/NVPTX/atomicrmw-expand.ll

3–4

Do we have it documented somewhere that detection of ptxas version X.Y, enables ptxas-A.B feature for all A.B <= X.Y?

llvm/test/CodeGen/NVPTX/atomicrmw-expand.ll

 ; RUN: llc < %s -march=nvptx64 -mcpu=sm_30 | FileCheck %s --check-prefixes=ALL,SM30
 ; RUN: llc < %s -march=nvptx64 -mcpu=sm_60 | FileCheck %s --check-prefixes=ALL,SM60
-; RUN: %if ptxas %{ llc < %s -march=nvptx64 -mcpu=sm_30 | %ptxas-verify %}
-; RUN: %if ptxas %{ llc < %s -march=nvptx64 -mcpu=sm_60 | %ptxas-verify %}
+; RUN: %if ptxas %{ llc < %s -march=nvptx64 -mcpu=sm_30 | %ptxas-verify %if !ptxas-11.0 %{-arch=sm_30%} %}
+; RUN: %if ptxas %{ llc < %s -march=nvptx64 -mcpu=sm_60 | %ptxas-verify -arch=sm_60 %}
 
 ; CHECK-LABEL: fadd_double
 define void @fadd_double(ptr %0, double %1) {
 entry:
   ; SM30: atom.cas.b64
   ; SM60: atom.add.f64
   %2 = atomicrmw fadd ptr %0, double %1 monotonic, align 8
   ret void

llvm/test/CodeGen/NVPTX/bug52623.ll

-; RUN: llc < %s -march=nvptx -mcpu=sm_75 -verify-machineinstrs
-; RUN: %if ptxas %{ llc < %s -march=nvptx -mcpu=sm_75 | %ptxas-verify %}
+; RUN: llc < %s -march=nvptx -verify-machineinstrs
+; RUN: %if ptxas %{ llc < %s -march=nvptx | %ptxas-verify %}
 
 ; Check that llc will not crash even when first MBB doesn't contain
 ; any instruction.
 
 target datalayout = "e-i64:64-i128:128-v16:16-v32:32-n16:32:64"
 target triple = "nvptx64"
 
 %printf_args.0.8 = type { i8* }

llvm/test/CodeGen/NVPTX/ldu-ldg.ll

-; RUN: llc < %s -march=nvptx -mcpu=sm_20 | FileCheck %s
-; RUN: %if ptxas %{ llc < %s -march=nvptx -mcpu=sm_20 | %ptxas-verify %}
+; RUN: llc < %s -march=nvptx -mcpu=sm_32 | FileCheck %s
+; RUN: %if ptxas %{ llc < %s -march=nvptx -mcpu=sm_32 | %ptxas-verify %if !ptxas-11.0 %{-arch=sm_32%} %}
 
 
 declare i8 @llvm.nvvm.ldu.global.i.i8.p1i8(i8 addrspace(1)* %ptr, i32 %align)
 declare i32 @llvm.nvvm.ldu.global.i.i32.p1i32(i32 addrspace(1)* %ptr, i32 %align)
 declare i8 @llvm.nvvm.ldg.global.i.i8.p1i8(i8 addrspace(1)* %ptr, i32 %align)
 declare i32 @llvm.nvvm.ldg.global.i.i32.p1i32(i32 addrspace(1)* %ptr, i32 %align)
 
 

llvm/test/lit.cfg.py

if version:
         # versions, so add a feature for every known version prior to
         # the current one.
         ptxas_known_versions = [
             (9, 0), (9, 1), (9, 2),
             (10, 0), (10, 1), (10, 2),
             (11, 0), (11, 1), (11, 2), (11, 3), (11, 4), (11, 5), (11, 6),
         ]
 
+        def version_int(ver):
+            return ver[0] * 100 + ver[1]
+
         # ignore ptxas if its version is below the minimum supported
         # version
         min_version = ptxas_known_versions[0]
-        if version[0] < min_version[0] or version[1] < min_version[1]:
+        if version_int(version) < version_int(min_version):
             print(
                 'Warning: ptxas version {}.{} is not supported'.format(
                     version[0], version[1]))
             return
 
-        for known_major, known_minor in ptxas_known_versions:
-            if known_major <= version[0] and known_minor <= version[1]:
+        for known_version in ptxas_known_versions:
+            if version_int(known_version) <= version_int(version):
+                major, minor = known_version
                 config.available_features.add(
-                    'ptxas-{}.{}'.format(known_major, known_minor))
+                    'ptxas-{}.{}'.format(major, minor))
 
     config.available_features.add('ptxas')
     tools.extend([ToolSubst('%ptxas', ptxas_executable),
                   ToolSubst('%ptxas-verify', '{} -c -o /dev/null -'.format(
                       ptxas_executable))])
 
 ptxas_executable = \
     os.environ.get('LLVM_PTXAS_EXECUTABLE', None) or config.ptxas_executable

llvm/test/lit.site.cfg.py.in

 config.gold_executable = "@GOLD_EXECUTABLE@"
 config.ld64_executable = "@LD64_EXECUTABLE@"
 config.osx_sysroot = path(r"@CMAKE_OSX_SYSROOT@")
 config.ocamlfind_executable = "@OCAMLFIND@"
 config.have_ocamlopt = @HAVE_OCAMLOPT@
 config.ocaml_flags = "@OCAMLFLAGS@"
 config.include_go_tests = @LLVM_INCLUDE_GO_TESTS@
 config.go_executable = "@GO_EXECUTABLE@"
-config.ptxas_executable = "@PXTAS_EXECUTABLE@"
+config.ptxas_executable = "@PTXAS_EXECUTABLE@"
 config.enable_shared = @ENABLE_SHARED@
 config.enable_assertions = @ENABLE_ASSERTIONS@
 config.targets_to_build = "@TARGETS_TO_BUILD@"
 config.native_target = "@LLVM_NATIVE_ARCH@"
 config.llvm_bindings = "@LLVM_BINDINGS@".split(' ')
 config.host_os = "@HOST_OS@"
 config.host_cc = "@HOST_CC@"
 config.host_cxx = "@HOST_CXX@"