From a10d9cd8767e88a064719d71cc97144ba8b112d1 Mon Sep 17 00:00:00 2001 From: jsmall-nvidia Date: Tue, 10 Mar 2020 12:31:25 -0400 Subject: WIP Prefix Sum for CUDA (#1268) * Fix some typos. * Add wave-prefix-sum.slang test * First pass at implementing prefixSum. * Small improvments to prefixSum CUDA. * Small improvement to prefix sum. * Enable prefix sum in stdlib. --- tests/hlsl-intrinsic/wave-prefix-sum.slang | 16 ++++++++++++++++ tests/hlsl-intrinsic/wave-prefix-sum.slang.expected.txt | 8 ++++++++ 2 files changed, 24 insertions(+) create mode 100644 tests/hlsl-intrinsic/wave-prefix-sum.slang create mode 100644 tests/hlsl-intrinsic/wave-prefix-sum.slang.expected.txt (limited to 'tests') diff --git a/tests/hlsl-intrinsic/wave-prefix-sum.slang b/tests/hlsl-intrinsic/wave-prefix-sum.slang new file mode 100644 index 000000000..f8d9bb560 --- /dev/null +++ b/tests/hlsl-intrinsic/wave-prefix-sum.slang @@ -0,0 +1,16 @@ +//DISABLE_TEST(compute):COMPARE_COMPUTE_EX:-cpu -compute +//DISABLE_TEST(compute):COMPARE_COMPUTE_EX:-slang -compute +//TEST(compute):COMPARE_COMPUTE_EX:-slang -compute -dx12 -use-dxil -profile cs_6_0 +//TEST(compute, vulkan):COMPARE_COMPUTE_EX:-vk -compute +//TEST(compute):COMPARE_COMPUTE_EX:-cuda -compute + +//TEST_INPUT:ubuffer(data=[0 0 0 0 0 0 0 0], stride=4):out,name outputBuffer +RWStructuredBuffer outputBuffer; + +[numthreads(8, 1, 1)] +void computeMain(uint3 dispatchThreadID : SV_DispatchThreadID) +{ + int idx = int(dispatchThreadID.x); + int val = WavePrefixSum(1 << idx); + outputBuffer[idx] = val; +} \ No newline at end of file diff --git a/tests/hlsl-intrinsic/wave-prefix-sum.slang.expected.txt b/tests/hlsl-intrinsic/wave-prefix-sum.slang.expected.txt new file mode 100644 index 000000000..6ec6deeea --- /dev/null +++ b/tests/hlsl-intrinsic/wave-prefix-sum.slang.expected.txt @@ -0,0 +1,8 @@ +0 +1 +3 +7 +F +1F +3F +7F -- cgit v1.2.3