aaron/ispc
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Improve atomic_swap_global() to take advantage of associativity.

Browse Source 
We now do a single atomic hardware swap and then effectively do 
swaps between the running program instances such that the result
is the same as if they had happened to run a particular ordering
of hardware swaps themselves.

Also cleaned up __atomic_swap_uniform_* built-in implementations
to not take the mask, which they weren't using anyway.

Finishes Issue #56.
This commit is contained in:
Matt Pharr
2012-01-20 10:37:33 -08:00
parent 4388338dad
commit 1bba9d4307
3 changed files with 113 additions and 82 deletions
Download Patch File Download Diff File Expand all files Collapse all files

68
builtins/util.m4
View File

@@ -760,14 +760,12 @@ define(`global_atomic_uniform', `
ifelse(LLVM_VERSION, `LLVM_2_9',`
declare $3 @llvm.atomic.load.$2.$3.p0$3($3 * %ptr, $3 %delta)
define $3 @__atomic_$2_uniform_$4_global($3 * %ptr, $3 %val,
<$1 x MASK> %mask) nounwind alwaysinline {
define $3 @__atomic_$2_uniform_$4_global($3 * %ptr, $3 %val) nounwind alwaysinline {
%r = call $3 @llvm.atomic.load.$2.$3.p0$3($3 * %ptr, $3 %val)
ret $3 %r
}
', `
define $3 @__atomic_$2_uniform_$4_global($3 * %ptr, $3 %val,
<$1 x MASK> %mask) nounwind alwaysinline {
define $3 @__atomic_$2_uniform_$4_global($3 * %ptr, $3 %val) nounwind alwaysinline {
%r = atomicrmw $2 $3 * %ptr, $3 %val seq_cst
ret $3 %r
}
@@ -786,26 +784,7 @@ declare i32 @llvm.atomic.swap.i32.p0i32(i32 * %ptr, i32 %val)
declare i64 @llvm.atomic.swap.i64.p0i64(i64 * %ptr, i64 %val)')
define(`global_swap', `
define <$1 x $2> @__atomic_swap_$3_global($2* %ptr, <$1 x $2> %val,
<$1 x MASK> %mask) nounwind alwaysinline {
%rptr = alloca <$1 x $2>
%rptr32 = bitcast <$1 x $2> * %rptr to $2 *
per_lane($1, <$1 x MASK> %mask, `
%val_LANE_ID = extractelement <$1 x $2> %val, i32 LANE
ifelse(LLVM_VERSION, `LLVM_2_9',`
%r_LANE_ID = call $2 @llvm.atomic.swap.$2.p0$2($2 * %ptr, $2 %val_LANE_ID)', `
%r_LANE_ID = atomicrmw xchg $2 * %ptr, $2 %val_LANE_ID seq_cst')
%rp_LANE_ID = getelementptr $2 * %rptr32, i32 LANE
store $2 %r_LANE_ID, $2 * %rp_LANE_ID')
%r = load <$1 x $2> * %rptr
ret <$1 x $2> %r
}
define $2 @__atomic_swap_uniform_$3_global($2* %ptr, $2 %val,
<$1 x MASK> %mask) nounwind alwaysinline {
define $2 @__atomic_swap_uniform_$3_global($2* %ptr, $2 %val) nounwind alwaysinline {
ifelse(LLVM_VERSION, `LLVM_2_9',`
%r = call $2 @llvm.atomic.swap.$2.p0$2($2 * %ptr, $2 %val)', `
%r = atomicrmw xchg $2 * %ptr, $2 %val seq_cst')
@@ -845,7 +824,7 @@ ifelse(LLVM_VERSION, `LLVM_2_9',`
}
define $2 @__atomic_compare_exchange_uniform_$3_global($2* %ptr, $2 %cmp,
$2 %val, <$1 x MASK> %mask) nounwind alwaysinline {
$2 %val) nounwind alwaysinline {
ifelse(LLVM_VERSION, `LLVM_2_9',`
%r = call $2 @llvm.atomic.cmp.swap.$2.p0$2($2 * %ptr, $2 %cmp, $2 %val)', `
%r = cmpxchg $2 * %ptr, $2 %cmp, $2 %val seq_cst')
@@ -1997,38 +1976,18 @@ global_atomic_uniform(WIDTH, umax, i64, uint64)
global_swap(WIDTH, i32, int32)
global_swap(WIDTH, i64, int64)
define <WIDTH x float> @__atomic_swap_float_global(float * %ptr, <WIDTH x float> %val,
<WIDTH x MASK> %mask) nounwind alwaysinline {
%iptr = bitcast float * %ptr to i32 *
%ival = bitcast <WIDTH x float> %val to <WIDTH x i32>
%iret = call <WIDTH x i32> @__atomic_swap_int32_global(i32 * %iptr, <WIDTH x i32> %ival, <WIDTH x MASK> %mask)
%ret = bitcast <WIDTH x i32> %iret to <WIDTH x float>
ret <WIDTH x float> %ret
}
define <WIDTH x double> @__atomic_swap_double_global(double * %ptr, <WIDTH x double> %val,
<WIDTH x MASK> %mask) nounwind alwaysinline {
%iptr = bitcast double * %ptr to i64 *
%ival = bitcast <WIDTH x double> %val to <WIDTH x i64>
%iret = call <WIDTH x i64> @__atomic_swap_int64_global(i64 * %iptr, <WIDTH x i64> %ival, <WIDTH x MASK> %mask)
%ret = bitcast <WIDTH x i64> %iret to <WIDTH x double>
ret <WIDTH x double> %ret
}
define float @__atomic_swap_uniform_float_global(float * %ptr, float %val,
<WIDTH x MASK> %mask) nounwind alwaysinline {
define float @__atomic_swap_uniform_float_global(float * %ptr, float %val) nounwind alwaysinline {
%iptr = bitcast float * %ptr to i32 *
%ival = bitcast float %val to i32
%iret = call i32 @__atomic_swap_uniform_int32_global(i32 * %iptr, i32 %ival, <WIDTH x MASK> %mask)
%iret = call i32 @__atomic_swap_uniform_int32_global(i32 * %iptr, i32 %ival)
%ret = bitcast i32 %iret to float
ret float %ret
}
define double @__atomic_swap_uniform_double_global(double * %ptr, double %val,
<WIDTH x MASK> %mask) nounwind alwaysinline {
define double @__atomic_swap_uniform_double_global(double * %ptr, double %val) nounwind alwaysinline {
%iptr = bitcast double * %ptr to i64 *
%ival = bitcast double %val to i64
%iret = call i64 @__atomic_swap_uniform_int64_global(i64 * %iptr, i64 %ival, <WIDTH x MASK> %mask)
%iret = call i64 @__atomic_swap_uniform_int64_global(i64 * %iptr, i64 %ival)
%ret = bitcast i64 %iret to double
ret double %ret
}
@@ -2058,24 +2017,23 @@ define <WIDTH x double> @__atomic_compare_exchange_double_global(double * %ptr,
ret <WIDTH x double> %ret
}
define float @__atomic_compare_exchange_uniform_float_global(float * %ptr, float %cmp, float %val,
<WIDTH x MASK> %mask) nounwind alwaysinline {
define float @__atomic_compare_exchange_uniform_float_global(float * %ptr, float %cmp,
float %val) nounwind alwaysinline {
%iptr = bitcast float * %ptr to i32 *
%icmp = bitcast float %cmp to i32
%ival = bitcast float %val to i32
%iret = call i32 @__atomic_compare_exchange_uniform_int32_global(i32 * %iptr, i32 %icmp,
i32 %ival, <WIDTH x MASK> %mask)
i32 %ival)
%ret = bitcast i32 %iret to float
ret float %ret
}
define double @__atomic_compare_exchange_uniform_double_global(double * %ptr, double %cmp,
double %val, <WIDTH x MASK> %mask) nounwind alwaysinline {
double %val) nounwind alwaysinline {
%iptr = bitcast double * %ptr to i64 *
%icmp = bitcast double %cmp to i64
%ival = bitcast double %val to i64
%iret = call i64 @__atomic_compare_exchange_uniform_int64_global(i64 * %iptr, i64 %icmp,
i64 %ival, <WIDTH x MASK> %mask)
%iret = call i64 @__atomic_compare_exchange_uniform_int64_global(i64 * %iptr, i64 %icmp, i64 %ival)
%ret = bitcast i64 %iret to double
ret double %ret
}