Skip to content

GitLab

Unverified Commit 052a7265 authored by carlushuang's avatar carlushuang Committed by GitHub
Browse files

add fp8 as dst (#1830)

parent 1fe2c352
Hide whitespace changes
Inline Side-by-side
Showing with 139 additions and 120 deletions
example/ck_tile/14_moe_smoothquant/instances/moe_smoothquant_fwd_api.cpp
View file @ 052a7265
...@@ -4,7 +4,8 @@ ...@@ -4,7 +4,8 @@
#include <ck_tile/core.hpp> #include <ck_tile/core.hpp>
#include "moe_smoothquant.hpp" #include "moe_smoothquant.hpp"
template <typename DataType_, template <typename InType,
typename OutType,
ck_tile::index_t Repeat_M_, // each thread repeat along M ck_tile::index_t Repeat_M_, // each thread repeat along M
ck_tile::index_t Repeat_N_, // each thread repeat along N ck_tile::index_t Repeat_N_, // each thread repeat along N
ck_tile::index_t ThreadPerBlock_M_, // num threads along M ck_tile::index_t ThreadPerBlock_M_, // num threads along M
...@@ -12,7 +13,8 @@ template <typename DataType_, ...@@ -12,7 +13,8 @@ template <typename DataType_,
ck_tile::index_t Vector_N_, // vector size along N ck_tile::index_t Vector_N_, // vector size along N
bool kPadN_, bool kPadN_,
bool kTwoPass_> bool kTwoPass_>
using trait_ = moe_smoothquant_traits_<DataType_, using trait_ = moe_smoothquant_traits_<InType,
OutType,
Repeat_M_, Repeat_M_,
Repeat_N_, Repeat_N_,
ThreadPerBlock_M_, ThreadPerBlock_M_,
...@@ -21,108 +23,108 @@ using trait_ = moe_smoothquant_traits_<DataType_, ...@@ -21,108 +23,108 @@ using trait_ = moe_smoothquant_traits_<DataType_,
kPadN_, kPadN_,
kTwoPass_>; kTwoPass_>;
template <typename data_type> template <typename in_type, typename out_type>
float moe_smoothquant_dispatch(moe_smoothquant_traits /*t*/, float moe_smoothquant_dispatch(moe_smoothquant_traits /*t*/,
moe_smoothquant_args a, moe_smoothquant_args a,
const ck_tile::stream_config& s) const ck_tile::stream_config& s)
{ {
float r = -1; float r = -1;
// clang-format off // clang-format off
// rm rn tm tn vn pd 2p // rm rn tm tn vn pd 2p
if(a.hidden_size <= 64) { if(a.hidden_size <= 64) {
r = moe_smoothquant_<trait_<data_type, 1, 1, 4, 64, 1, true, false>>(s, a); r = moe_smoothquant_<trait_<in_type, out_type, 1, 1, 4, 64, 1, true, false>>(s, a);
} }
else if(a.hidden_size <= 128) { else if(a.hidden_size <= 128) {
if (a.hidden_size % 2 == 0) if (a.hidden_size % 2 == 0)
r = moe_smoothquant_<trait_<data_type, 1, 1, 4, 64, 2, true, false>>(s, a); r = moe_smoothquant_<trait_<in_type, out_type, 1, 1, 4, 64, 2, true, false>>(s, a);
else else
r = moe_smoothquant_<trait_<data_type, 1, 2, 4, 64, 1, true, false>>(s, a); r = moe_smoothquant_<trait_<in_type, out_type, 1, 2, 4, 64, 1, true, false>>(s, a);
} }
else if(a.hidden_size <= 256) { else if(a.hidden_size <= 256) {
if (a.hidden_size % 4 == 0) if (a.hidden_size % 4 == 0)
r = moe_smoothquant_<trait_<data_type, 1, 1, 4, 64, 4, true, false>>(s, a); r = moe_smoothquant_<trait_<in_type, out_type, 1, 1, 4, 64, 4, true, false>>(s, a);
else if (a.hidden_size % 2 == 0) else if (a.hidden_size % 2 == 0)
r = moe_smoothquant_<trait_<data_type, 1, 2, 4, 64, 2, true, false>>(s, a); r = moe_smoothquant_<trait_<in_type, out_type, 1, 2, 4, 64, 2, true, false>>(s, a);
else else
r = moe_smoothquant_<trait_<data_type, 1, 4, 4, 64, 1, true, false>>(s, a); r = moe_smoothquant_<trait_<in_type, out_type, 1, 4, 4, 64, 1, true, false>>(s, a);
} }
else if(a.hidden_size <= 512) { else if(a.hidden_size <= 512) {
if (a.hidden_size % 8 == 0) if (a.hidden_size % 8 == 0)
r = moe_smoothquant_<trait_<data_type, 1, 1, 4, 64, 8, true, false>>(s, a); r = moe_smoothquant_<trait_<in_type, out_type, 1, 1, 4, 64, 8, true, false>>(s, a);
else if (a.hidden_size % 4 == 0) else if (a.hidden_size % 4 == 0)
r = moe_smoothquant_<trait_<data_type, 1, 2, 4, 64, 4, true, false>>(s, a); r = moe_smoothquant_<trait_<in_type, out_type, 1, 2, 4, 64, 4, true, false>>(s, a);
else if (a.hidden_size % 2 == 0) else if (a.hidden_size % 2 == 0)
r = moe_smoothquant_<trait_<data_type, 1, 4, 4, 64, 2, true, false>>(s, a); r = moe_smoothquant_<trait_<in_type, out_type, 1, 4, 4, 64, 2, true, false>>(s, a);
else else
r = moe_smoothquant_<trait_<data_type, 1, 8, 4, 64, 1, true, false>>(s, a); r = moe_smoothquant_<trait_<in_type, out_type, 1, 8, 4, 64, 1, true, false>>(s, a);
} }
else if(a.hidden_size <= 768) { else if(a.hidden_size <= 768) {
if (a.hidden_size % 4 == 0) if (a.hidden_size % 4 == 0)
r = moe_smoothquant_<trait_<data_type, 1, 3, 4, 64, 4, true, false>>(s, a); r = moe_smoothquant_<trait_<in_type, out_type, 1, 3, 4, 64, 4, true, false>>(s, a);
else if (a.hidden_size % 2 == 0) else if (a.hidden_size % 2 == 0)
r = moe_smoothquant_<trait_<data_type, 1, 6, 4, 64, 2, true, false>>(s, a); r = moe_smoothquant_<trait_<in_type, out_type, 1, 6, 4, 64, 2, true, false>>(s, a);
else else
r = moe_smoothquant_<trait_<data_type, 1,12, 4, 64, 1, true, false>>(s, a); r = moe_smoothquant_<trait_<in_type, out_type, 1,12, 4, 64, 1, true, false>>(s, a);
} }
else if(a.hidden_size <= 1024) { else if(a.hidden_size <= 1024) {
if (a.hidden_size % 8 == 0) if (a.hidden_size % 8 == 0)
r = moe_smoothquant_<trait_<data_type, 1, 1, 2, 128, 8, true, false>>(s, a); r = moe_smoothquant_<trait_<in_type, out_type, 1, 1, 2, 128, 8, true, false>>(s, a);
else if (a.hidden_size % 4 == 0) else if (a.hidden_size % 4 == 0)
r = moe_smoothquant_<trait_<data_type, 1, 2, 2, 128, 4, true, false>>(s, a); r = moe_smoothquant_<trait_<in_type, out_type, 1, 2, 2, 128, 4, true, false>>(s, a);
else if (a.hidden_size % 2 == 0) else if (a.hidden_size % 2 == 0)
r = moe_smoothquant_<trait_<data_type, 1, 4, 2, 128, 2, true, false>>(s, a); r = moe_smoothquant_<trait_<in_type, out_type, 1, 4, 2, 128, 2, true, false>>(s, a);
else else
r = moe_smoothquant_<trait_<data_type, 1, 4, 1, 256, 1, true, false>>(s, a); r = moe_smoothquant_<trait_<in_type, out_type, 1, 4, 1, 256, 1, true, false>>(s, a);
} }
else if(a.hidden_size <= 1536) { else if(a.hidden_size <= 1536) {
if (a.hidden_size % 8 == 0) if (a.hidden_size % 8 == 0)
r = moe_smoothquant_<trait_<data_type, 1, 3, 4, 64, 8, true, false>>(s, a); r = moe_smoothquant_<trait_<in_type, out_type, 1, 3, 4, 64, 8, true, false>>(s, a);
else if (a.hidden_size % 4 == 0) else if (a.hidden_size % 4 == 0)
r = moe_smoothquant_<trait_<data_type, 1, 3, 2, 128, 4, true, false>>(s, a); r = moe_smoothquant_<trait_<in_type, out_type, 1, 3, 2, 128, 4, true, false>>(s, a);
else if (a.hidden_size % 2 == 0) else if (a.hidden_size % 2 == 0)
r = moe_smoothquant_<trait_<data_type, 1, 3, 1, 256, 2, true, false>>(s, a); r = moe_smoothquant_<trait_<in_type, out_type, 1, 3, 1, 256, 2, true, false>>(s, a);
else else
r = moe_smoothquant_<trait_<data_type, 1, 6, 1, 256, 1, true, false>>(s, a); r = moe_smoothquant_<trait_<in_type, out_type, 1, 6, 1, 256, 1, true, false>>(s, a);
} }
else if(a.hidden_size <= 2048) { else if(a.hidden_size <= 2048) {
if (a.hidden_size % 8 == 0) if (a.hidden_size % 8 == 0)
r = moe_smoothquant_<trait_<data_type, 1, 1, 1, 256, 8, true, false>>(s, a); r = moe_smoothquant_<trait_<in_type, out_type, 1, 1, 1, 256, 8, true, false>>(s, a);
else if (a.hidden_size % 4 == 0) else if (a.hidden_size % 4 == 0)
r = moe_smoothquant_<trait_<data_type, 1, 2, 1, 256, 4, true, false>>(s, a); r = moe_smoothquant_<trait_<in_type, out_type, 1, 2, 1, 256, 4, true, false>>(s, a);
else if (a.hidden_size % 2 == 0) else if (a.hidden_size % 2 == 0)
r = moe_smoothquant_<trait_<data_type, 1, 4, 1, 256, 2, true, false>>(s, a); r = moe_smoothquant_<trait_<in_type, out_type, 1, 4, 1, 256, 2, true, false>>(s, a);
else else
r = moe_smoothquant_<trait_<data_type, 1, 8, 1, 256, 1, true, false>>(s, a); r = moe_smoothquant_<trait_<in_type, out_type, 1, 8, 1, 256, 1, true, false>>(s, a);
} }
else if(a.hidden_size <= 3072) { else if(a.hidden_size <= 3072) {
if (a.hidden_size % 8 == 0) if (a.hidden_size % 8 == 0)
r = moe_smoothquant_<trait_<data_type, 1, 3, 1, 128, 8, true, false>>(s, a); r = moe_smoothquant_<trait_<in_type, out_type, 1, 3, 1, 128, 8, true, false>>(s, a);
else if (a.hidden_size % 4 == 0) else if (a.hidden_size % 4 == 0)
r = moe_smoothquant_<trait_<data_type, 1, 3, 1, 256, 4, true, false>>(s, a); r = moe_smoothquant_<trait_<in_type, out_type, 1, 3, 1, 256, 4, true, false>>(s, a);
else if (a.hidden_size % 2 == 0) else if (a.hidden_size % 2 == 0)
r = moe_smoothquant_<trait_<data_type, 1, 6, 1, 256, 2, true, false>>(s, a); r = moe_smoothquant_<trait_<in_type, out_type, 1, 6, 1, 256, 2, true, false>>(s, a);
else else
r = moe_smoothquant_<trait_<data_type, 1, 3, 1, 1024, 1, true, false>>(s, a); r = moe_smoothquant_<trait_<in_type, out_type, 1, 3, 1, 1024, 1, true, false>>(s, a);
} }
else if(a.hidden_size <= 4096) { else if(a.hidden_size <= 4096) {
if (a.hidden_size % 8 == 0) if (a.hidden_size % 8 == 0)
r = moe_smoothquant_<trait_<data_type, 1, 2, 1, 256, 8, true, false>>(s, a); r = moe_smoothquant_<trait_<in_type, out_type, 1, 2, 1, 256, 8, true, false>>(s, a);
else if (a.hidden_size % 4 == 0) else if (a.hidden_size % 4 == 0)
r = moe_smoothquant_<trait_<data_type, 1, 4, 1, 256, 4, true, false>>(s, a); r = moe_smoothquant_<trait_<in_type, out_type, 1, 4, 1, 256, 4, true, false>>(s, a);
else if (a.hidden_size % 2 == 0) else if (a.hidden_size % 2 == 0)
r = moe_smoothquant_<trait_<data_type, 1, 2, 1, 1024, 2, true, false>>(s, a); r = moe_smoothquant_<trait_<in_type, out_type, 1, 2, 1, 1024, 2, true, false>>(s, a);
else else
r = moe_smoothquant_<trait_<data_type, 1, 4, 1, 1024, 1, true, false>>(s, a); r = moe_smoothquant_<trait_<in_type, out_type, 1, 4, 1, 1024, 1, true, false>>(s, a);
} }
else if(a.hidden_size > 4096) { else if(a.hidden_size > 4096) {
if (a.hidden_size % 8 == 0) if (a.hidden_size % 8 == 0)
r = moe_smoothquant_<trait_<data_type, 1, 2, 1, 256, 8, true, true>>(s, a); r = moe_smoothquant_<trait_<in_type, out_type, 1, 2, 1, 256, 8, true, true>>(s, a);
else if (a.hidden_size % 4 == 0) else if (a.hidden_size % 4 == 0)
r = moe_smoothquant_<trait_<data_type, 1, 4, 1, 256, 4, true, true>>(s, a); r = moe_smoothquant_<trait_<in_type, out_type, 1, 4, 1, 256, 4, true, true>>(s, a);
else if (a.hidden_size % 2 == 0) else if (a.hidden_size % 2 == 0)
r = moe_smoothquant_<trait_<data_type, 1, 2, 1, 1024, 2, true, true>>(s, a); r = moe_smoothquant_<trait_<in_type, out_type, 1, 2, 1, 1024, 2, true, true>>(s, a);
else else
r = moe_smoothquant_<trait_<data_type, 1, 4, 1, 1024, 1, true, true>>(s, a); r = moe_smoothquant_<trait_<in_type, out_type, 1, 4, 1, 1024, 1, true, true>>(s, a);
} }
return r; return r;
// clang-format on // clang-format on
...@@ -132,13 +134,21 @@ float moe_smoothquant(moe_smoothquant_traits t, ...@@ -132,13 +134,21 @@ float moe_smoothquant(moe_smoothquant_traits t,
moe_smoothquant_args a, moe_smoothquant_args a,
const ck_tile::stream_config& s) const ck_tile::stream_config& s)
{ {
if(t.data_type.compare("fp16") == 0) if(t.in_type.compare("fp16") == 0 && t.out_type == "int8")
{ {
return moe_smoothquant_dispatch<ck_tile::fp16_t>(t, a, s); return moe_smoothquant_dispatch<ck_tile::fp16_t, ck_tile::int8_t>(t, a, s);
} }
else if(t.data_type.compare("bf16") == 0) else if(t.in_type.compare("fp16") == 0 && t.out_type == "fp8")
{ {
return moe_smoothquant_dispatch<ck_tile::bf16_t>(t, a, s); return moe_smoothquant_dispatch<ck_tile::fp16_t, ck_tile::fp8_t>(t, a, s);
}
else if(t.in_type.compare("bf16") == 0 && t.out_type == "int8")
{
return moe_smoothquant_dispatch<ck_tile::bf16_t, ck_tile::int8_t>(t, a, s);
}
else if(t.in_type.compare("bf16") == 0 && t.out_type == "fp8")
{
return moe_smoothquant_dispatch<ck_tile::bf16_t, ck_tile::fp8_t>(t, a, s);
} }
else else
throw std::runtime_error("Without supported instances!"); throw std::runtime_error("Without supported instances!");
......
example/ck_tile/14_moe_smoothquant/instances/moe_smoothquant_instance_common.hpp
View file @ 052a7265
...@@ -11,7 +11,8 @@ ...@@ -11,7 +11,8 @@
using S = ck_tile::stream_config; using S = ck_tile::stream_config;
using A = moe_smoothquant_args; using A = moe_smoothquant_args;
template <typename DataType_, template <typename InputType_,
typename OutputType_,
ck_tile::index_t Repeat_M_, // each thread repeat along M ck_tile::index_t Repeat_M_, // each thread repeat along M
ck_tile::index_t Repeat_N_, // each thread repeat along N ck_tile::index_t Repeat_N_, // each thread repeat along N
ck_tile::index_t ThreadPerBlock_M_, // num threads along M ck_tile::index_t ThreadPerBlock_M_, // num threads along M
...@@ -19,7 +20,8 @@ template <typename DataType_, ...@@ -19,7 +20,8 @@ template <typename DataType_,
ck_tile::index_t Vector_N_, // vector size along N ck_tile::index_t Vector_N_, // vector size along N
bool kPadN_, bool kPadN_,
bool kTwoPass_> bool kTwoPass_>
using trait_ = moe_smoothquant_traits_<DataType_, using trait_ = moe_smoothquant_traits_<InputType_,
OutputType_,
Repeat_M_, Repeat_M_,
Repeat_N_, Repeat_N_,
ThreadPerBlock_M_, ThreadPerBlock_M_,
...@@ -31,14 +33,15 @@ using trait_ = moe_smoothquant_traits_<DataType_, ...@@ -31,14 +33,15 @@ using trait_ = moe_smoothquant_traits_<DataType_,
template <typename Traits_> template <typename Traits_>
float moe_smoothquant_(const S& s, A a) float moe_smoothquant_(const S& s, A a)
{ {
using DataType = typename Traits_::DataType; using InputType = typename Traits_::InputType;
using OutputType = typename Traits_::OutputType;
using PipelineProblem = ck_tile::SmoothquantPipelineProblem< using PipelineProblem = ck_tile::SmoothquantPipelineProblem<
typename MoeSmoothquantTypeConfig<DataType>::XDataType, typename MoeSmoothquantTypeConfig<InputType, OutputType>::XDataType,
typename MoeSmoothquantTypeConfig<DataType>::SmoothScaleDataType, typename MoeSmoothquantTypeConfig<InputType, OutputType>::SmoothScaleDataType,
typename MoeSmoothquantTypeConfig<DataType>::ComputeDataType, typename MoeSmoothquantTypeConfig<InputType, OutputType>::ComputeDataType,
typename MoeSmoothquantTypeConfig<DataType>::YScaleDataType, typename MoeSmoothquantTypeConfig<InputType, OutputType>::YScaleDataType,
typename MoeSmoothquantTypeConfig<DataType>::QYDataType, typename MoeSmoothquantTypeConfig<InputType, OutputType>::QYDataType,
typename Traits_::Shape, typename Traits_::Shape,
Traits_::kPadN, Traits_::kPadN,
Traits_::kTwoPass>; Traits_::kTwoPass>;
......
example/ck_tile/14_moe_smoothquant/moe_smoothquant.cpp
View file @ 052a7265
...@@ -63,7 +63,8 @@ auto create_args(int argc, char* argv[]) ...@@ -63,7 +63,8 @@ auto create_args(int argc, char* argv[])
.insert("stride", "-1", "stride per row, if -1 then equal to hidden_size") .insert("stride", "-1", "stride per row, if -1 then equal to hidden_size")
.insert("v", "1", "cpu validation or not") .insert("v", "1", "cpu validation or not")
.insert("kname", "1", "print kernel name or not") .insert("kname", "1", "print kernel name or not")
.insert("prec", "fp16", "precision") .insert("prec_i", "fp16", "input precision, fp16/bf16")
.insert("prec_o", "int8", "precision, int8/fp8")
.insert("warmup", "5", "cold iter") .insert("warmup", "5", "cold iter")
.insert("repeat", "20", "hot iter"); .insert("repeat", "20", "hot iter");
...@@ -71,7 +72,7 @@ auto create_args(int argc, char* argv[]) ...@@ -71,7 +72,7 @@ auto create_args(int argc, char* argv[])
return std::make_tuple(result, arg_parser); return std::make_tuple(result, arg_parser);
} }
template <typename DataType> template <typename InputType, typename OutputType>
bool run(const ck_tile::ArgParser& arg_parser) bool run(const ck_tile::ArgParser& arg_parser)
{ {
ck_tile::index_t tokens = arg_parser.get_int("t"); ck_tile::index_t tokens = arg_parser.get_int("t");
...@@ -81,7 +82,8 @@ bool run(const ck_tile::ArgParser& arg_parser) ...@@ -81,7 +82,8 @@ bool run(const ck_tile::ArgParser& arg_parser)
stride = hidden_size; stride = hidden_size;
ck_tile::index_t experts = arg_parser.get_int("e"); ck_tile::index_t experts = arg_parser.get_int("e");
ck_tile::index_t topk = arg_parser.get_int("k"); ck_tile::index_t topk = arg_parser.get_int("k");
std::string data_type = arg_parser.get_str("prec"); std::string prec_i = arg_parser.get_str("prec_i");
std::string prec_o = arg_parser.get_str("prec_o");
int kname = arg_parser.get_int("kname"); int kname = arg_parser.get_int("kname");
int do_validation = arg_parser.get_int("v"); int do_validation = arg_parser.get_int("v");
int warmup = arg_parser.get_int("warmup"); int warmup = arg_parser.get_int("warmup");
...@@ -89,7 +91,7 @@ bool run(const ck_tile::ArgParser& arg_parser) ...@@ -89,7 +91,7 @@ bool run(const ck_tile::ArgParser& arg_parser)
assert(stride >= hidden_size); assert(stride >= hidden_size);
using TypeConfig = MoeSmoothquantTypeConfig<DataType>; using TypeConfig = MoeSmoothquantTypeConfig<InputType, OutputType>;
using XDataType = typename TypeConfig::XDataType; using XDataType = typename TypeConfig::XDataType;
using SmoothScaleDataType = typename TypeConfig::SmoothScaleDataType; using SmoothScaleDataType = typename TypeConfig::SmoothScaleDataType;
...@@ -122,11 +124,11 @@ bool run(const ck_tile::ArgParser& arg_parser) ...@@ -122,11 +124,11 @@ bool run(const ck_tile::ArgParser& arg_parser)
smscale_buf.ToDevice(smscale_host.data()); smscale_buf.ToDevice(smscale_host.data());
topk_ids_buf.ToDevice(topk_ids_host.data()); topk_ids_buf.ToDevice(topk_ids_host.data());
std::cout << "[" << data_type << "]" std::cout << "[" << prec_i << "-" << prec_o << "]"
<< " tokens:" << tokens << ", hidden_size:" << hidden_size << ", stride:" << stride << " tokens:" << tokens << ", hidden_size:" << hidden_size << ", stride:" << stride
<< ", experts:" << experts << ", topk:" << topk << std::flush; << ", experts:" << experts << ", topk:" << topk << std::flush;
moe_smoothquant_traits traits{data_type}; moe_smoothquant_traits traits{prec_i, prec_o};
moe_smoothquant_args args{x_buf.GetDeviceBuffer(), moe_smoothquant_args args{x_buf.GetDeviceBuffer(),
smscale_buf.GetDeviceBuffer(), smscale_buf.GetDeviceBuffer(),
...@@ -251,14 +253,23 @@ int main(int argc, char* argv[]) ...@@ -251,14 +253,23 @@ int main(int argc, char* argv[])
if(!result) if(!result)
return -1; return -1;
const std::string data_type = arg_parser.get_str("prec"); const std::string prec_i = arg_parser.get_str("prec_i");
if(data_type == "fp16") const std::string prec_o = arg_parser.get_str("prec_o");
if(prec_i == "fp16" && prec_o == "int8")
{
return run<ck_tile::half_t, ck_tile::int8_t>(arg_parser) ? 0 : -2;
}
else if(prec_i == "fp16" && prec_o == "fp8")
{
return run<ck_tile::half_t, ck_tile::fp8_t>(arg_parser) ? 0 : -2;
}
else if(prec_i == "bf16" && prec_o == "int8")
{ {
return run<ck_tile::half_t>(arg_parser) ? 0 : -2; return run<ck_tile::bf16_t, ck_tile::int8_t>(arg_parser) ? 0 : -2;
} }
else if(data_type == "bf16") else if(prec_i == "bf16" && prec_o == "fp8")
{ {
return run<ck_tile::bf16_t>(arg_parser) ? 0 : -2; return run<ck_tile::bf16_t, ck_tile::fp8_t>(arg_parser) ? 0 : -2;
} }
return -3; return -3;
......
example/ck_tile/14_moe_smoothquant/moe_smoothquant.hpp
View file @ 052a7265
...@@ -8,26 +8,13 @@ ...@@ -8,26 +8,13 @@
#include "ck_tile/ops/smoothquant.hpp" #include "ck_tile/ops/smoothquant.hpp"
#include <string> #include <string>
template <typename DataType> template <typename InputType, typename OutputType>
struct MoeSmoothquantTypeConfig; struct MoeSmoothquantTypeConfig
template <>
struct MoeSmoothquantTypeConfig<ck_tile::half_t>
{
using XDataType = ck_tile::half_t;
using SmoothScaleDataType = float;
using YScaleDataType = float;
using QYDataType = ck_tile::int8_t;
using ComputeDataType = float;
};
template <>
struct MoeSmoothquantTypeConfig<ck_tile::bf16_t>
{ {
using XDataType = ck_tile::bf16_t; using XDataType = InputType;
using SmoothScaleDataType = float; using SmoothScaleDataType = float;
using YScaleDataType = float; using YScaleDataType = float;
using QYDataType = ck_tile::int8_t; using QYDataType = OutputType;
using ComputeDataType = float; using ComputeDataType = float;
}; };
...@@ -37,7 +24,8 @@ struct moe_smoothquant_args : public ck_tile::MoeSmoothquantHostArgs ...@@ -37,7 +24,8 @@ struct moe_smoothquant_args : public ck_tile::MoeSmoothquantHostArgs
}; };
// this is used to pattern-match internl kernel implementation, not to instantiate kernel // this is used to pattern-match internl kernel implementation, not to instantiate kernel
template <typename DataType_, template <typename InputType_,
typename OutputType_,
ck_tile::index_t Repeat_M_, // each thread repeat along M ck_tile::index_t Repeat_M_, // each thread repeat along M
ck_tile::index_t Repeat_N_, // each thread repeat along N ck_tile::index_t Repeat_N_, // each thread repeat along N
ck_tile::index_t ThreadPerBlock_M_, // num threads along M ck_tile::index_t ThreadPerBlock_M_, // num threads along M
...@@ -47,7 +35,8 @@ template <typename DataType_, ...@@ -47,7 +35,8 @@ template <typename DataType_,
bool kTwoPass_> bool kTwoPass_>
struct moe_smoothquant_traits_ struct moe_smoothquant_traits_
{ {
using DataType = ck_tile::remove_cvref_t<DataType_>; using InputType = ck_tile::remove_cvref_t<InputType_>;
using OutputType = ck_tile::remove_cvref_t<OutputType_>;
static constexpr bool is_warp_per_row = ThreadPerBlock_N_ <= warpSize; static constexpr bool is_warp_per_row = ThreadPerBlock_N_ <= warpSize;
static_assert((ThreadPerBlock_M_ * ThreadPerBlock_N_) % warpSize == 0); static_assert((ThreadPerBlock_M_ * ThreadPerBlock_N_) % warpSize == 0);
...@@ -108,7 +97,8 @@ float moe_smoothquant_(const ck_tile::stream_config& s, moe_smoothquant_args a); ...@@ -108,7 +97,8 @@ float moe_smoothquant_(const ck_tile::stream_config& s, moe_smoothquant_args a);
// This is the public API, will be generated by script // This is the public API, will be generated by script
struct moe_smoothquant_traits struct moe_smoothquant_traits
{ {
std::string data_type; std::string in_type; // input type
std::string out_type; // output type
}; };
float moe_smoothquant(moe_smoothquant_traits, moe_smoothquant_args, const ck_tile::stream_config&); float moe_smoothquant(moe_smoothquant_traits, moe_smoothquant_args, const ck_tile::stream_config&);
example/ck_tile/14_moe_smoothquant/script/smoke_test.sh
View file @ 052a7265
...@@ -2,29 +2,31 @@ ...@@ -2,29 +2,31 @@
EXE=build/bin/tile_example_moe_smoothquant EXE=build/bin/tile_example_moe_smoothquant
for pr_i in "fp16" "bf16" ; do for pr_i in "fp16" "bf16" ; do
$EXE -prec=$pr_i -t=99 -h=13 for pr_o in "int8" "fp8" ; do
$EXE -prec=$pr_i -t=17 -h=16 $EXE -prec_i=$pr_i -prec_o=$pr_o -t=99 -h=13
$EXE -prec=$pr_i -t=1 -h=100 $EXE -prec_i=$pr_i -prec_o=$pr_o -t=17 -h=16
$EXE -prec=$pr_i -t=4 -h=128 $EXE -prec_i=$pr_i -prec_o=$pr_o -t=1 -h=100
$EXE -prec=$pr_i -t=80 -h=127 $EXE -prec_i=$pr_i -prec_o=$pr_o -t=4 -h=128
$EXE -prec=$pr_i -t=22 -h=255 -stride=256 $EXE -prec_i=$pr_i -prec_o=$pr_o -t=80 -h=127
$EXE -prec=$pr_i -t=7 -h=599 $EXE -prec_i=$pr_i -prec_o=$pr_o -t=22 -h=255 -stride=256
$EXE -prec=$pr_i -t=19 -h=512 $EXE -prec_i=$pr_i -prec_o=$pr_o -t=7 -h=599
$EXE -prec=$pr_i -t=33 -h=313 -stride=1000 $EXE -prec_i=$pr_i -prec_o=$pr_o -t=19 -h=512
$EXE -prec=$pr_i -t=11 -h=510 $EXE -prec_i=$pr_i -prec_o=$pr_o -t=33 -h=313 -stride=1000
$EXE -prec=$pr_i -t=171 -h=676 -stride=818 $EXE -prec_i=$pr_i -prec_o=$pr_o -t=11 -h=510
$EXE -prec=$pr_i -t=91 -h=636 $EXE -prec_i=$pr_i -prec_o=$pr_o -t=171 -h=676 -stride=818
$EXE -prec=$pr_i -t=12 -h=768 -stride=800 $EXE -prec_i=$pr_i -prec_o=$pr_o -t=91 -h=636
$EXE -prec=$pr_i -t=100 -h=766 -stride=812 $EXE -prec_i=$pr_i -prec_o=$pr_o -t=12 -h=768 -stride=800
$EXE -prec=$pr_i -t=31 -h=1024 $EXE -prec_i=$pr_i -prec_o=$pr_o -t=100 -h=766 -stride=812
$EXE -prec=$pr_i -t=64 -h=1000 -stride=1004 $EXE -prec_i=$pr_i -prec_o=$pr_o -t=31 -h=1024
$EXE -prec=$pr_i -t=8 -h=1501 $EXE -prec_i=$pr_i -prec_o=$pr_o -t=64 -h=1000 -stride=1004
$EXE -prec=$pr_i -t=3 -h=1826 $EXE -prec_i=$pr_i -prec_o=$pr_o -t=8 -h=1501
$EXE -prec=$pr_i -t=5 -h=2040 $EXE -prec_i=$pr_i -prec_o=$pr_o -t=3 -h=1826
$EXE -prec=$pr_i -t=7 -h=2734 $EXE -prec_i=$pr_i -prec_o=$pr_o -t=5 -h=2040
$EXE -prec=$pr_i -t=1 -h=3182 $EXE -prec_i=$pr_i -prec_o=$pr_o -t=7 -h=2734
$EXE -prec=$pr_i -t=9 -h=4096 $EXE -prec_i=$pr_i -prec_o=$pr_o -t=1 -h=3182
$EXE -prec=$pr_i -t=3 -h=8192 $EXE -prec_i=$pr_i -prec_o=$pr_o -t=9 -h=4096
$EXE -prec=$pr_i -t=1 -h=10547 $EXE -prec_i=$pr_i -prec_o=$pr_o -t=3 -h=8192
$EXE -prec=$pr_i -t=3 -h=17134 $EXE -prec_i=$pr_i -prec_o=$pr_o -t=1 -h=10547
$EXE -prec_i=$pr_i -prec_o=$pr_o -t=3 -h=17134
done
done done
include/ck_tile/core/utility/unary_element_function.hpp
View file @ 052a7265
...@@ -51,16 +51,18 @@ struct composes<F> ...@@ -51,16 +51,18 @@ struct composes<F>
template <typename... Ts> template <typename... Ts>
__host__ __device__ composes(Ts&&...)->composes<remove_cvref_t<Ts>...>; __host__ __device__ composes(Ts&&...)->composes<remove_cvref_t<Ts>...>;
template <typename To> template <typename SaturateType>
struct saturates struct saturates
{ {
template <typename From> // NOTE: this function does not return SaturateType value
CK_TILE_HOST_DEVICE constexpr auto operator()(const From& from) const // it is user's responsiblity to do further cast or not
-> std::enable_if_t<std::is_arithmetic_v<From>, From> template <typename AccType>
CK_TILE_HOST_DEVICE constexpr auto operator()(const AccType& a_) const
-> std::enable_if_t<std::is_arithmetic_v<AccType>, AccType>
{ {
return clamp(from, return clamp(a_,
type_convert<From>(numeric<To>::lowest()), type_convert<AccType>(numeric<SaturateType>::lowest()),
type_convert<From>(numeric<To>::max())); type_convert<AccType>(numeric<SaturateType>::max()));
} }
}; };
......
include/ck_tile/host/reference/reference_rowwise_quantization2d.hpp
View file @ 052a7265
...@@ -22,7 +22,7 @@ CK_TILE_HOST void reference_rowwise_quantization2d(const HostTensor<XDataType>& ...@@ -22,7 +22,7 @@ CK_TILE_HOST void reference_rowwise_quantization2d(const HostTensor<XDataType>&
// scale = amax / 127 for int8 // scale = amax / 127 for int8
auto v_scale = type_convert<XDataType>(scale_m(m)); auto v_scale = type_convert<XDataType>(scale_m(m));
auto v_qx = v_x / v_scale; auto v_qx = v_x / v_scale;
qx_m_n(m, n) = saturates<QXDataType>{}(v_qx); qx_m_n(m, n) = type_convert<QXDataType>(saturates<QXDataType>{}(v_qx));
} }
}; };
......
include/ck_tile/ops/smoothquant/kernel/moe_smoothquant_kernel.hpp
View file @ 052a7265
...@@ -101,6 +101,7 @@ struct MoeSmoothquant ...@@ -101,6 +101,7 @@ struct MoeSmoothquant
template <> struct t2s<ck_tile::bf16_t> { static constexpr const char * name = "bf16"; }; template <> struct t2s<ck_tile::bf16_t> { static constexpr const char * name = "bf16"; };
template <> struct t2s<ck_tile::fp8_t> { static constexpr const char * name = "fp8"; }; template <> struct t2s<ck_tile::fp8_t> { static constexpr const char * name = "fp8"; };
template <> struct t2s<ck_tile::bf8_t> { static constexpr const char * name = "bf8"; }; template <> struct t2s<ck_tile::bf8_t> { static constexpr const char * name = "bf8"; };
template <> struct t2s<ck_tile::int8_t> { static constexpr const char * name = "i8"; };
// clang-format on // clang-format on
// in byte // in byte
...@@ -118,7 +119,7 @@ struct MoeSmoothquant ...@@ -118,7 +119,7 @@ struct MoeSmoothquant
#define _SS_ std::string #define _SS_ std::string
#define _TS_ std::to_string #define _TS_ std::to_string
return _SS_("moe_smoothquant_") + _SS_(t2s<XDataType>::name) + "_" + return _SS_("moe_smoothquant_") + _SS_(t2s<XDataType>::name) + "_" + _SS_(t2s<QYDataType>::name) + "_" +
_TS_(S_::Block_M) + "x" + _TS_(S_::Block_N) + "_" + _TS_(S_::WarpPerBlock_M) + "x" + _TS_(S_::WarpPerBlock_N) + "_" + _TS_(S_::Block_M) + "x" + _TS_(S_::Block_N) + "_" + _TS_(S_::WarpPerBlock_M) + "x" + _TS_(S_::WarpPerBlock_N) + "_" +
_TS_(S_::Warp_M) + "x" + _TS_(S_::Warp_N) + "_" + _TS_(S_::Vector_M) + "x" + _TS_(S_::Vector_N) + "_" + _TS_(S_::Warp_M) + "x" + _TS_(S_::Warp_N) + "_" + _TS_(S_::Vector_M) + "x" + _TS_(S_::Vector_N) + "_" +
_SS_(Pipeline::name) + surfix; _SS_(Pipeline::name) + surfix;
......
include/ck_tile/ops/smoothquant/pipeline/smoothquant_pipeline_one_pass.hpp
View file @ 052a7265
...@@ -113,7 +113,7 @@ struct SmoothquantPipelineOnePass ...@@ -113,7 +113,7 @@ struct SmoothquantPipelineOnePass
sweep_tile(qy, [&](auto idx) { sweep_tile(qy, [&](auto idx) {
constexpr auto i_idx = make_tuple(idx[number<0>{}]); constexpr auto i_idx = make_tuple(idx[number<0>{}]);
auto qy_ = y[idx] / yscale[i_idx]; auto qy_ = y[idx] / yscale[i_idx];
qy(idx) = saturates<QYDataType>{}(qy_); qy(idx) = type_convert<QYDataType>(saturates<QYDataType>{}(qy_));
}); });
store_tile(qy_window, qy); store_tile(qy_window, qy);
} }
......
include/ck_tile/ops/smoothquant/pipeline/smoothquant_pipeline_two_pass.hpp
View file @ 052a7265
...@@ -136,7 +136,7 @@ struct SmoothquantPipelineTwoPass ...@@ -136,7 +136,7 @@ struct SmoothquantPipelineTwoPass
sweep_tile(qy, [&](auto idx) { sweep_tile(qy, [&](auto idx) {
constexpr auto i_idx = make_tuple(idx[number<0>{}]); constexpr auto i_idx = make_tuple(idx[number<0>{}]);
auto qy_ = y[idx] / yscale[i_idx]; auto qy_ = y[idx] / yscale[i_idx];
qy(idx) = saturates<QYDataType>{}(qy_); qy(idx) = type_convert<QYDataType>(saturates<QYDataType>{}(qy_));
}); });
store_tile(qy_window, qy); store_tile(qy_window, qy);
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment