Skip to content

Commit

Permalink
support max3 in smoothquant and add+ rmsnorm + rdquant (#1654)
Browse files Browse the repository at this point in the history 
* Fix cmake example build

* Support max3 in smoothquant one pass

* support max3 in two pass

* support max3 in add_rmsnorm_rdquant
  • Loading branch information
@rocking5566
rocking5566 authored Nov 26, 2024
1 parent bfe983a commit abae2af
Show file tree
Hide file tree
Showing 5 changed files with 94 additions and 19 deletions.
4 changes: 2 additions & 2 deletions example/ck_tile/12_smoothquant/CMakeLists.txt
View file
Original file line number Diff line number Diff line change
Expand Up @@ -18,7 +18,7 @@ function (add_smoothquant_example TARGET_NAME MAIN_SRC)
target_compile_options(${TARGET_NAME} PRIVATE ${COMPILE_OPTIONS})
endfunction(add_smoothquant_example TARGET_NAME MAIN_SRC)

file(GLOB INSTANCE_SRCS instances/*.cpp)

add_smoothquant_example(tile_smoothquant smoothquant.cpp ${INSTANCE_SRCS})
add_smoothquant_example(tile_example_smoothquant example_smoothquant.cpp)
file(GLOB INSTANCE_SRCS instances/*.cpp)
add_smoothquant_example(tile_smoothquant smoothquant.cpp ${INSTANCE_SRCS})
37 changes: 30 additions & 7 deletions ...k_tile/ops/add_rmsnorm2d_rdquant/pipeline/add_rmsnorm2d_rdquant_fwd_pipeline_one_pass.hpp
View file
Original file line number Diff line number Diff line change
Expand Up @@ -28,8 +28,9 @@ struct AddRmsnorm2dRdquantFwdPipelineOnePass
static constexpr bool kSaveX = Problem::kSaveX;

static constexpr bool kNeedCrossWarpSync = Problem::kNeedCrossWarpSync;
static constexpr bool kPadM = false; // TODO - BlockAddRmsnorm2dRdquantFwdProblem::kPadM
static constexpr bool kPadN = Problem::kPadN;
static constexpr bool kPadM = false; // TODO - BlockAddRmsnorm2dRdquantFwdProblem::kPadM
static constexpr bool kPadN = Problem::kPadN;
static constexpr bool UseMax3 = true; // TODO - Move to trait

static constexpr const char* name = []() {
if constexpr(kNeedCrossWarpSync)
Expand Down Expand Up @@ -69,9 +70,16 @@ struct AddRmsnorm2dRdquantFwdPipelineOnePass
auto reduce_square_sum_func = ReduceOp::SquareAdd{};
auto reduce_sum_func = ReduceOp::Add{};
auto reduce_absmax_func = ReduceOp::AbsMax{};
auto reduce_max_func = ReduceOp::Max{};
auto block_reduce2d = Policy::template GetBlockReduce2d<Problem>();
auto block_reduce2d_sync = Policy::template GetBlockReduce2dSync<Problem>();
auto reduce_absmax3_func = [](auto acc_, auto v_0_, auto v_1_) {
float rtn;
asm volatile("v_max3_f32 %0, %1, abs(%2), abs(%3)"
: "=v"(rtn)
: "v"(acc_), "v"(v_0_), "v"(v_1_));
return rtn;
};
auto reduce_max_func = ReduceOp::Max{};
auto block_reduce2d = Policy::template GetBlockReduce2d<Problem>();
auto block_reduce2d_sync = Policy::template GetBlockReduce2dSync<Problem>();
auto block_reduce2d_cross_warp_sync =
Policy::template GetBlockReduce2dCrossWarpSync<Problem>();

Expand Down Expand Up @@ -116,8 +124,23 @@ struct AddRmsnorm2dRdquantFwdPipelineOnePass
});

// compute absmax, each-thread->cross-lane->cross-warp
auto absmax = block_reduce2d(
y, reduce_absmax_func.GetIdentityValue<ComputeDataType>(), reduce_absmax_func);
auto absmax = [&]() {
constexpr auto x_size_per_row =
x.get_tile_distribution().get_ys_to_d_descriptor().get_lengths().at(number<1>{});
if constexpr(UseMax3 && std::is_same_v<ComputeDataType, float> &&
x_size_per_row % 2 == 0)
{
return block_reduce2d(y,
reduce_absmax_func.GetIdentityValue<ComputeDataType>(),
reduce_absmax3_func,
sequence<1, 2>{});
}
else
{
return block_reduce2d(
y, reduce_absmax_func.GetIdentityValue<ComputeDataType>(), reduce_absmax_func);
}
}();
block_reduce2d_sync(absmax, reduce_max_func);
block_reduce2d_cross_warp_sync(absmax, smem, reduce_max_func);

Expand Down
26 changes: 20 additions & 6 deletions ...tile/ops/add_rmsnorm2d_rdquant/pipeline/add_rmsnorm2d_rdquant_fwd_pipeline_three_pass.hpp
View file
Original file line number Diff line number Diff line change
Expand Up @@ -28,8 +28,9 @@ struct AddRmsnorm2dRdquantFwdPipelineThreePass
static constexpr bool kSaveX = Problem::kSaveX;

static constexpr bool kNeedCrossWarpSync = Problem::kNeedCrossWarpSync;
static constexpr bool kPadM = false; // TODO - BlockAddRmsnorm2dRdquantFwdProblem::kPadM
static constexpr bool kPadN = Problem::kPadN;
static constexpr bool kPadM = false; // TODO - BlockAddRmsnorm2dRdquantFwdProblem::kPadM
static constexpr bool kPadN = Problem::kPadN;
static constexpr bool UseMax3 = true; // TODO - Move to trait

static constexpr const char* name = []() {
if constexpr(kNeedCrossWarpSync)
Expand Down Expand Up @@ -76,9 +77,16 @@ struct AddRmsnorm2dRdquantFwdPipelineThreePass
auto reduce_square_sum_func = ReduceOp::SquareAdd{};
auto reduce_sum_func = ReduceOp::Add{};
auto reduce_absmax_func = ReduceOp::AbsMax{};
auto reduce_max_func = ReduceOp::Max{};
auto block_reduce2d = Policy::template GetBlockReduce2d<Problem>();
auto block_reduce2d_sync = Policy::template GetBlockReduce2dSync<Problem>();
auto reduce_absmax3_func = [](auto acc_, auto v_0_, auto v_1_) {
float rtn;
asm volatile("v_max3_f32 %0, %1, abs(%2), abs(%3)"
: "=v"(rtn)
: "v"(acc_), "v"(v_0_), "v"(v_1_));
return rtn;
};
auto reduce_max_func = ReduceOp::Max{};
auto block_reduce2d = Policy::template GetBlockReduce2d<Problem>();
auto block_reduce2d_sync = Policy::template GetBlockReduce2dSync<Problem>();
auto block_reduce2d_cross_warp_sync =
Policy::template GetBlockReduce2dCrossWarpSync<Problem>();

Expand Down Expand Up @@ -177,7 +185,13 @@ struct AddRmsnorm2dRdquantFwdPipelineThreePass
y(idx) = type_convert<ComputeDataType>(y_);
});

block_reduce2d(y, absmax, reduce_absmax_func);
constexpr auto x_size_per_row =
x.get_tile_distribution().get_ys_to_d_descriptor().get_lengths().at(number<1>{});
if constexpr(UseMax3 && std::is_same_v<ComputeDataType, float> &&
x_size_per_row % 2 == 0)
block_reduce2d(y, absmax, reduce_absmax3_func, sequence<1, 2>{});
else
block_reduce2d(y, absmax, reduce_absmax_func);

if constexpr(kSaveX)
move_tile_window(x_window, {0, -Block_N});
Expand Down
30 changes: 27 additions & 3 deletions include/ck_tile/ops/smoothquant/pipeline/smoothquant_pipeline_one_pass.hpp
View file
Original file line number Diff line number Diff line change
Expand Up @@ -25,6 +25,7 @@ struct SmoothquantPipelineOnePass
static constexpr bool kNeedCrossWarpSync = Problem::kNeedCrossWarpSync;
static constexpr bool kPadM = false; // TODO - BlockSmoothquantProblem::kPadM
static constexpr bool kPadN = Problem::kPadN;
static constexpr bool UseMax3 = true; // TODO - Move to trait

static constexpr const char* name = []() {
if constexpr(kNeedCrossWarpSync)
Expand Down Expand Up @@ -52,7 +53,15 @@ struct SmoothquantPipelineOnePass
xscale_window_, Policy::template MakeXScaleBlockTileDistribution<Problem>());

auto reduce_absmax_func = ReduceOp::AbsMax{};
auto reduce_max_func = ReduceOp::Max{};
auto reduce_absmax3_func = [](auto acc_, auto v_0_, auto v_1_) {
float rtn;
asm volatile("v_max3_f32 %0, %1, abs(%2), abs(%3)"
: "=v"(rtn)
: "v"(acc_), "v"(v_0_), "v"(v_1_));
return rtn;
};
auto reduce_max_func = ReduceOp::Max{};

auto block_reduce2d = Policy::template GetBlockReduce2d<Problem>();
auto block_reduce2d_sync = Policy::template GetBlockReduce2dSync<Problem>();
auto block_reduce2d_cross_warp_sync =
Expand All @@ -68,8 +77,23 @@ struct SmoothquantPipelineOnePass
xscale);

// compute absmax, cross-lane->cross-warp
auto absmax = block_reduce2d(
y, reduce_absmax_func.GetIdentityValue<ComputeDataType>(), reduce_absmax_func);
auto absmax = [&]() {
constexpr auto x_size_per_row =
x.get_tile_distribution().get_ys_to_d_descriptor().get_lengths().at(number<1>{});
if constexpr(UseMax3 && std::is_same_v<ComputeDataType, float> &&
x_size_per_row % 2 == 0)
{
return block_reduce2d(y,
reduce_absmax_func.GetIdentityValue<ComputeDataType>(),
reduce_absmax3_func,
sequence<1, 2>{});
}
else
{
return block_reduce2d(
y, reduce_absmax_func.GetIdentityValue<ComputeDataType>(), reduce_absmax_func);
}
}();
block_reduce2d_sync(absmax, reduce_max_func);
block_reduce2d_cross_warp_sync(absmax, smem, reduce_max_func);

Expand Down
16 changes: 15 additions & 1 deletion include/ck_tile/ops/smoothquant/pipeline/smoothquant_pipeline_two_pass.hpp
View file
Original file line number Diff line number Diff line change
Expand Up @@ -25,6 +25,7 @@ struct SmoothquantPipelineTwoPass
static constexpr bool kNeedCrossWarpSync = Problem::kNeedCrossWarpSync;
static constexpr bool kPadM = false; // TODO - BlockSmoothquantProblem::kPadM
static constexpr bool kPadN = Problem::kPadN;
static constexpr bool UseMax3 = true; // TODO - Move to trait

static constexpr const char* name = []() {
if constexpr(kNeedCrossWarpSync)
Expand Down Expand Up @@ -56,6 +57,13 @@ struct SmoothquantPipelineTwoPass
__builtin_amdgcn_readfirstlane(integer_divide_ceil(row_size, Block_N));

auto reduce_absmax_func = ReduceOp::AbsMax{};
auto reduce_absmax3_func = [](auto acc_, auto v_0_, auto v_1_) {
float rtn;
asm volatile("v_max3_f32 %0, %1, abs(%2), abs(%3)"
: "=v"(rtn)
: "v"(acc_), "v"(v_0_), "v"(v_1_));
return rtn;
};
auto reduce_max_func = ReduceOp::Max{};
auto block_reduce2d = Policy::template GetBlockReduce2d<Problem>();
auto block_reduce2d_sync = Policy::template GetBlockReduce2dSync<Problem>();
Expand All @@ -77,7 +85,13 @@ struct SmoothquantPipelineTwoPass
x,
xscale);

block_reduce2d(y, absmax, reduce_absmax_func);
constexpr auto x_size_per_row =
x.get_tile_distribution().get_ys_to_d_descriptor().get_lengths().at(number<1>{});
if constexpr(UseMax3 && std::is_same_v<ComputeDataType, float> &&
x_size_per_row % 2 == 0)
block_reduce2d(y, absmax, reduce_absmax3_func, sequence<1, 2>{});
else
block_reduce2d(y, absmax, reduce_absmax_func);

move_tile_window(x_window, {0, Block_N});
move_tile_window(xscale_window, {Block_N});
Expand Down

0 comments on commit abae2af

Please sign in to comment.