ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-05-11 00:40:09 +00:00
Code Issues Packages Projects Releases Wiki Activity

tempsave, fmha_decode

Browse Source
This commit is contained in:
aska-0096
2025-07-08 08:37:20 +00:00
parent 47565f21a5
commit 18686cfe5b
13 changed files with 2562 additions and 71 deletions
Download Patch File Download Diff File Expand all files Collapse all files

19
example/ck_tile/01_fmha/CMakeLists.txt
View File

@@ -1,6 +1,6 @@
# validate user-specified fmha_fwd API list
set(FMHA_FWD_KNOWN_APIS "fwd;fwd_splitkv;fwd_appendkv")
set(FMHA_FWD_ENABLE_APIS "fwd" CACHE STRING
set(FMHA_FWD_KNOWN_APIS "fwd;fwd_splitkv;fwd_appendkv;fwd_decode")
set(FMHA_FWD_ENABLE_APIS "fwd_decode" CACHE STRING
"semicolon-separated list of APIs to generate (${FMHA_FWD_KNOWN_APIS}) & link, or \"all\".")
if(FMHA_FWD_ENABLE_APIS STREQUAL "all")
set(FMHA_FWD_ENABLE_APIS ${FMHA_FWD_KNOWN_APIS})
@@ -13,12 +13,12 @@ foreach(api ${FMHA_FWD_ENABLE_APIS})
endforeach()
# "fwd" is a must-have api for the fmha_fwd example, add it if not specified
if(NOT "fwd" IN_LIST FMHA_FWD_ENABLE_APIS)
list(APPEND FMHA_FWD_ENABLE_APIS "fwd")
endif()
# if(NOT "fwd" IN_LIST FMHA_FWD_ENABLE_APIS)
# list(APPEND FMHA_FWD_ENABLE_APIS "fwd")
# endif()
# Filtering kernel​
# set(KERNEL fmha_fwd_d64_bf16_batch_b128x64x32x64x32x64_r4x1x1_r4x1x1_w32x32x16_w32x32x16_qr_async_vr_psskddv_nlogits_nbias_nmask_nlse_ndropout_nskip_nsquant)
# set(KERNEL fmha_fwd_decode_d64_bf16_batch_b16x64x32x64x32x64_r1x4x1_r1x4x1_w16x16x32_w16x16x32_decode_qr_vr_psskddv_nlogits_nbias_nmask_nlse_ndropout_nskip_nsquant)
string(REPLACE ";" "," FMHA_FWD_APIS "${FMHA_FWD_ENABLE_APIS}")
# generate a list of kernels, but not actually emit files at config sta
@@ -106,6 +106,13 @@ else()
list(APPEND EXAMPLE_FMHA_FWD_COMPILE_OPTIONS -DCK_TILE_FMHA_FWD_APPENDKV_API=0)
endif()
# conditionally enable call to the fwd_appendkv API in fmha_fwd example
if("fwd_decode" IN_LIST FMHA_FWD_ENABLE_APIS)
list(APPEND EXAMPLE_FMHA_FWD_COMPILE_OPTIONS -DCK_TILE_FMHA_FWD_DECODE_API=1)
else()
list(APPEND EXAMPLE_FMHA_FWD_COMPILE_OPTIONS -DCK_TILE_FMHA_FWD_DECODE_API=0)
endif()
# conditionally specify the use of OCP_FP8
if(CK_USE_OCP_FP8)
list(APPEND EXAMPLE_FMHA_FWD_COMPILE_OPTIONS -DCK_TILE_USE_OCP_FP8)

8
example/ck_tile/01_fmha/codegen/cpp_symbol_map.py
View File

@@ -3,7 +3,7 @@
# generate kernel instances to speed up compilation
FWD_DTYPE_MAP = {
# "fp16" : "FmhaFwdFp16",
"fp16" : "FmhaFwdFp16",
"bf16" : "FmhaFwdBf16",
# "fp8" : "FmhaFwdFp8",
# "fp8fp16": "FmhaFwdFp8Fp16",
@@ -22,7 +22,7 @@ MASK_IMPL = {
_MASK_SIMPLIFIED_MAP = {
"s_no" : "ck_tile::SimplifiedGenericAttentionMask<false>",
"s_mask" : "ck_tile::SimplifiedGenericAttentionMask<true>",
# "s_mask" : "ck_tile::SimplifiedGenericAttentionMask<true>",
}
_MASK_MAP = {
@@ -62,8 +62,8 @@ def get_mask_check_map(mask : str):
BIAS_MAP = {
"no" : "ck_tile::BlockAttentionBiasEnum::NO_BIAS",
"bias" : "ck_tile::BlockAttentionBiasEnum::ELEMENTWISE_BIAS",
"alibi" : "ck_tile::BlockAttentionBiasEnum::ALIBI"
# "bias" : "ck_tile::BlockAttentionBiasEnum::ELEMENTWISE_BIAS",
# "alibi" : "ck_tile::BlockAttentionBiasEnum::ALIBI"
}
# TODO: this is ugly

50
example/ck_tile/01_fmha/codegen/ops/fmha_fwd.py
View File

@@ -282,18 +282,19 @@ class FmhaFwdApiPool:
# TODO: do we need to check duplication?
if trait.dtype not in self.pool.keys():
self.pool[trait.dtype] = dict()
if trait.hdim not in self.pool[trait.dtype].keys():
self.pool[trait.dtype][trait.hdim] = list()
hdim = trait.hdim, trait.bn1
if hdim not in self.pool[trait.dtype].keys():
self.pool[trait.dtype][hdim] = list()
self.pool[trait.dtype][trait.hdim].append(copy.copy(trait))
self.pool[trait.dtype][hdim].append(copy.copy(trait))
@property
def api(self) -> str:
per_dtypes=str()
for i, dtype in enumerate(self.pool.keys()):
per_hdim_case=str()
for j, hdim in enumerate(self.pool[dtype].keys()):
traits=self.pool[dtype][hdim]
for j, (hdim, hdim_v) in enumerate(self.pool[dtype].keys()):
traits=self.pool[dtype][(hdim, hdim_v)]
inners=str()
for k, trait in enumerate(traits):
if_k = 'if' if k == 0 else 'else if'
@@ -306,7 +307,7 @@ class FmhaFwdApiPool:
F_bm0=trait.bm0, F_bn0=trait.bn0, F_bk0=trait.bk0, F_bn1=trait.bn1, F_bk1=trait.bk1, F_bk0max=trait.bk0max,
F_hdim=hdim, F_dtype=FWD_DTYPE_MAP[dtype])
if_j = 'if' if j == 0 else 'else if'
per_hdim_case = per_hdim_case + FMHA_FWD_API_PER_HDIM_CASE.format(F_if=if_j, F_hdim=hdim, F_hdim_v=trait.bn1, F_inner_dispatch=inners)
per_hdim_case = per_hdim_case + FMHA_FWD_API_PER_HDIM_CASE.format(F_if=if_j, F_hdim=hdim, F_hdim_v=hdim_v, F_inner_dispatch=inners)
if_i = 'if' if i == 0 else 'else if'
per_dtypes = per_dtypes + FMHA_FWD_API_PER_DTYPE.format(F_if=if_i, F_dtype=dtype, F_hdim_case=per_hdim_case)
if not per_dtypes:
@@ -435,18 +436,20 @@ class FmhaFwdKernel:
def get_fmha_fwd_tile_dict_from_dtype(dtype : str) -> Optional[dict]:
if dtype == 'fp16' or dtype == 'bf16':
return {
# '32' : FmhaFwdTileSize(128, 64, 16, 32, 32, 32, 2, 1, 1, 2, 1, 1, 32, 32, 16, 32, 32, 16, -1),
'64' : FmhaFwdTileSize(16, 64, 32, 64, 32, 64, 1, 4, 1, 1, 4, 1, 16, 16, 16, 16, 16, 16, -1),
### '96' : FmhaFwdTileSize(128, 128, 32, 128, 32, 96, 4, 1, 1, 4, 1, 1, 32, 32, 16, 32, 32, 16, -1),
'128' : FmhaFwdTileSize(16, 64, 64, 128, 64, 128, 1, 4, 1, 1, 4, 1, 16, 16, 16, 16, 16, 16, -1),
# '192' : FmhaFwdTileSize(128, 128, 32, 128, 32, 192, 4, 1, 1, 4, 1, 1, 32, 32, 16, 32, 32, 16, -1),
# '256' : FmhaFwdTileSize(128, 128, 32, 256, 32, 256, 4, 1, 1, 4, 1, 1, 32, 32, 16, 32, 32, 16, -1),
(32, 32) : FmhaFwdTileSize(128, 64, 16, 32, 32, 32, 2, 1, 1, 2, 1, 1, 32, 32, 16, 32, 32, 16, -1),
(64, 64) : FmhaFwdTileSize(128, 64, 32, 64, 32, 64, 4, 1, 1, 4, 1, 1, 32, 32, 16, 32, 32, 16, -1),
### (96, 128) : FmhaFwdTileSize(128, 128, 32, 128, 32, 96, 4, 1, 1, 4, 1, 1, 32, 32, 16, 32, 32, 16, -1),
(128,128) : FmhaFwdTileSize(128, 128, 32, 128, 32, 128, 4, 1, 1, 4, 1, 1, 32, 32, 16, 32, 32, 16, -1),
### (160,160) : FmhaFwdTileSize(128, 128, 32, 160, 32, 160, 4, 1, 1, 4, 1, 1, 32, 32, 16, 32, 32, 16, 1),
(192,128) : FmhaFwdTileSize(128, 128, 32, 128, 32, 192, 4, 1, 1, 4, 1, 1, 32, 32, 16, 32, 32, 16, -1),
### (192,192) : FmhaFwdTileSize(128, 128, 32, 192, 32, 192, 4, 1, 1, 4, 1, 1, 32, 32, 16, 32, 32, 16, 1),
(256,256) : FmhaFwdTileSize(128, 128, 32, 256, 32, 256, 4, 1, 1, 4, 1, 1, 32, 32, 16, 32, 32, 16, -1),
}
elif dtype == 'fp8' or dtype == 'bf8':
return {
'64' : FmhaFwdTileSize(128, 64, 32, 64, 32, 64, 2, 1, 1, 2, 1, 1, 32, 32, 32, 32, 32, 32, -1),
'128' : FmhaFwdTileSize(128, 128, 32, 128, 32, 128, 4, 1, 1, 4, 1, 1, 32, 32, 32, 32, 32, 32, -1),
'256' : FmhaFwdTileSize(128, 128, 32, 256, 32, 256, 4, 1, 1, 4, 1, 1, 32, 32, 32, 32, 32, 32, -1),
(64,64 ) : FmhaFwdTileSize(128, 64, 32, 64, 32, 64, 2, 1, 1, 2, 1, 1, 32, 32, 32, 32, 32, 32, -1),
(128,128) : FmhaFwdTileSize(128, 128, 32, 128, 32, 128, 4, 1, 1, 4, 1, 1, 32, 32, 32, 32, 32, 32, -1),
(256,256) : FmhaFwdTileSize(128, 128, 32, 256, 32, 256, 4, 1, 1, 4, 1, 1, 32, 32, 32, 32, 32, 32, -1),
}
else:
return None
@@ -454,7 +457,7 @@ def get_fmha_fwd_tile_dict_from_dtype(dtype : str) -> Optional[dict]:
def get_fwd_blobs(kernel_filter : Optional[str], receipt, optdim_list, mask_impl) -> Tuple[FmhaFwdApiPool, List[FmhaFwdKernel]]:
# TODO: we don't support tuning yet, so pick up one value for vlayout/pipeline/pad
# support this in future
def get_pipelines(dtype, hdim) -> List[FmhaFwdPipeline]:
def get_pipelines(dtype, hdim, hdim_v) -> List[FmhaFwdPipeline]:
# this function will populate a list possible pipelines
# TODO: the order of List matters! the later in this list will be also be checked later
# TODO: currently for qr pipeline, let 't' padding to appear later!!
@@ -463,7 +466,7 @@ def get_fwd_blobs(kernel_filter : Optional[str], receipt, optdim_list, mask_impl
pipelines = []
if dtype in ['fp16', 'bf16']:
for logits, mask, bias, lse, dropout, skip in itertools.product(["t", "f"], get_mask_map(mask_impl).keys(), BIAS_MAP.keys(), ["t", "f"], ["t", "f"], ["t", "f"]):
if hdim == 256:
if hdim == 256 and hdim_v == 256:
# if True:
pipelines.append(FmhaFwdPipeline('qr', 'row', 'f', 'f', 'f', 'f', logits, bias, lse, dropout, squant, mask, skip))
pipelines.append(FmhaFwdPipeline('qr', 'col', 'f', 'f', 'f', 'f', logits, bias, lse, dropout, squant, mask, skip))
@@ -473,11 +476,6 @@ def get_fwd_blobs(kernel_filter : Optional[str], receipt, optdim_list, mask_impl
pipelines.append(FmhaFwdPipeline('qr', 'row', 't', 't', 't', 't', logits, bias, lse, dropout, squant, mask, skip))
pipelines.append(FmhaFwdPipeline('qr', 'col', 't', 't', 't', 't', logits, bias, lse, dropout, squant, mask, skip))
elif hdim == 64 or hdim == 128:
pipelines.append(FmhaFwdPipeline('decode_qr', 'row', 't', 'f', 't', 't', logits, bias, lse, dropout, squant, mask, skip))
pipelines.append(FmhaFwdPipeline('decode_qr', 'col', 't', 't', 't', 't', logits, bias, lse, dropout, squant, mask, skip))
pipelines.append(FmhaFwdPipeline('decode_qr', 'row', 't', 'f', 't', 't', logits, bias, lse, dropout, squant, mask, skip))
pipelines.append(FmhaFwdPipeline('decode_qr', 'col', 't', 't', 't', 't', logits, bias, lse, dropout, squant, mask, skip))
else:
if bias == "bias":
# TODO: rocm 6.2 compiler problem if using qr_async for bias case
@@ -512,15 +510,13 @@ def get_fwd_blobs(kernel_filter : Optional[str], receipt, optdim_list, mask_impl
if d == None:
continue
#for hdim_str, mode, mask, bias, lse in itertools.product(d.keys(), MODE_MAP.keys(), MASK_MAP.keys(), ["t", "f"], ["t", "f"]):
for hdim_str, mode in itertools.product(d.keys(), MODE_MAP.keys()):
tile = d[hdim_str]
hdim = int(hdim_str)
for pipeline in get_pipelines(dtype, hdim):
for ((hdim, hdim_v), tile), mode in itertools.product(d.items(), MODE_MAP.keys()):
for pipeline in get_pipelines(dtype, hdim, hdim_v):
if mode == "group":
if pipeline.F_spad != 't' or pipeline.F_skpad != 't':
# in group mode, spad/skpad must be true, since we can't predict if seqlen of current batch need pad or not
continue
if hdim == 192 and tile.F_bn1 == 128:
if (hdim, hdim_v) == (192, 128) or hdim == 160:
# NOTE: this is used to speedup deepseek prefill case, we don't gen training
if pipeline.F_bias != 'no' or pipeline.F_dropout == 't':
continue

845
example/ck_tile/01_fmha/codegen/ops/fmha_fwd_decode.py Normal file
View File

@@ -0,0 +1,845 @@
# SPDX-License-Identifier: MIT
# Copyright (c) 2018-2024, Advanced Micro Devices, Inc. All rights reserved.
# generate kernel instances to speed up compilation
import copy
from dataclasses import dataclass
import fnmatch
import itertools
from pathlib import Path
from typing import List, Optional, Tuple, Union
from codegen.cmake_config import *
from codegen.cpp_symbol_map import *
from codegen.ops.fmha_fwd import (
FmhaFwdTileSize,
FmhaFwdApiTrait,
FMHA_FWD_KERNEL_HEADER,
FMHA_FWD_API_PER_DTYPE,
FMHA_FWD_API_PER_HDIM_CASE,
)
DTYPE_BITS = {
"fp32": 32,
"fp16": 16,
"bf16": 16,
"fp8" : 8,
"bf8" : 8
}
K0_MAX_SUBMAX_MAP = {
32 : 32,
64 : 64,
96 : 128,
128: 128,
256: 256
}
FMHA_FWD_DECODE_PIPELINE_MAP = {
"decode_qr" : "ck_tile::BlockFmhaFwdDecodePipelineQRKSVS",
}
FMHA_FWD_DECODE_KERNEL_BODY="""
using fmha_dtype_{F_idx} = {F_dtype};
using fmha_variant_{F_idx} = ck_tile::ComposedAttention<{F_logits} * ck_tile::LOGITS_SOFT_CAP, CK_TILE_FMHA_FWD_FAST_EXP2>;
using fmha_mask_{F_idx} = {F_mask};
namespace {{
template <bool kHasUnevenSplits, bool kMergeNumHeadGroupsSeqLenQ = false>
struct instance {{
using fmha_block_tile = ck_tile::sequence<{F_bm0}, {F_bn0}, {F_bk0}, {F_bn1}, {F_bk1}, {F_bk0max}>;
using fmha_shape = ck_tile::TileFmhaShape<fmha_block_tile,
ck_tile::sequence<{F_rm0}, {F_rn0}, {F_rk0}>,
ck_tile::sequence<{F_wm0}, {F_wn0}, {F_wk0}>,
ck_tile::sequence<{F_rm1}, {F_rn1}, {F_rk1}>,
ck_tile::sequence<{F_wm1}, {F_wn1}, {F_wk1}>,
{F_vlayout}>;
using fmha_trait = ck_tile::TileFmhaFwdDecodeTraits<{F_spad},
{F_skpad},
{F_dpad},
{F_dvpad},
{F_logits},
{F_bias},
/*kHasBiasGrad=*/false,
{F_lse},
{F_squant},
{F_pagedkv},
kHasUnevenSplits,
kMergeNumHeadGroupsSeqLenQ,
{F_occupancy}>;
using fmha_pipeline_problem = ck_tile::BlockFmhaFwdDecodePipelineProblem<
typename FmhaFwdTypeConfig<fmha_dtype_{F_idx}>::QDataType,
typename FmhaFwdTypeConfig<fmha_dtype_{F_idx}>::KDataType,
typename FmhaFwdTypeConfig<fmha_dtype_{F_idx}>::VDataType,
typename FmhaFwdTypeConfig<fmha_dtype_{F_idx}>::SaccDataType,
typename FmhaFwdTypeConfig<fmha_dtype_{F_idx}>::SMPLComputeDataType,
typename FmhaFwdTypeConfig<fmha_dtype_{F_idx}>::BiasDataType,
typename FmhaFwdTypeConfig<fmha_dtype_{F_idx}>::LSEDataType,
typename FmhaFwdTypeConfig<fmha_dtype_{F_idx}>::PDataType,
typename FmhaFwdTypeConfig<fmha_dtype_{F_idx}>::OaccDataType,
typename FmhaFwdTypeConfig<fmha_dtype_{F_idx}>::ODataType,
fmha_shape,
{F_mode},
fmha_variant_{F_idx},
fmha_mask_{F_idx},
fmha_trait>;
using fmha_pipeline = {F_pipeline}<
fmha_pipeline_problem>;
/// FIXME: use {F_spad}/{F_dvpad} as kPadM/kPadN parameters after solving
/// store_tile_raw() data corruption issue
using fmha_epilogue =
ck_tile::Default2DEpilogue<ck_tile::Default2DEpilogueProblem<typename FmhaFwdTypeConfig<{F_dtype}>::OaccDataType,
typename FmhaFwdTypeConfig<{F_dtype}>::ODataType,
false, false>>;
using fmha_kernel =
ck_tile::FmhaFwdDecodeKernel<fmha_pipeline, fmha_epilogue>;
static void run(const ck_tile::stream_config& s, fmha_fwd_decode_args a)
{{
using k_ = fmha_kernel;
auto [kargs, grids] = fmha_fwd_decode_create_kargs_and_grids<k_>(a);
constexpr dim3 blocks = k_::BlockSize();
constexpr ck_tile::index_t kBlockPerCu = k_::kBlockPerCu;
ck_tile::make_kernel<blocks.x, kBlockPerCu>(k_{{}}, grids, blocks, 0, kargs)(ck_tile::stream_config{{s.stream_id_}});
}}
}};
}}
using trait_{F_idx} = fmha_fwd_decode_traits_<{F_hdim}, {F_dtype}, {F_mode}, {F_bm0}, {F_bn0}, {F_bk0}, {F_bn1}, {F_bk1}, {F_bk0max}, {F_vlayout},
{F_pipeline_enum}, {F_logits}, fmha_mask_{F_idx}, {F_bias}, {F_lse}, {F_squant}, {F_pagedkv}, {F_spad}, {F_skpad}, {F_dpad},
{F_dvpad}>;
#include <iostream>
#pragma clang diagnostic push
#pragma clang diagnostic ignored "-Wtautological-compare"
namespace {{
template <bool kHasUnevenSplits>
void run_instance(const ck_tile::stream_config& s, fmha_fwd_decode_args a) {{
if constexpr ({F_hdim} == 128 && {F_bias} == ck_tile::BlockAttentionBiasEnum::NO_BIAS
&& (std::is_same_v<{F_mask}, ck_tile::SimplifiedGenericAttentionMask<false>>
|| std::is_same_v<{F_mask}, FmhaMasks::NoMask>)) {{
if (a.max_seqlen_q == 1 && a.nhead_k < a.nhead_q) {{
instance<kHasUnevenSplits, /*kMergeNumHeadGroupsSeqLenQ=*/true>::run(s, a);
}} else {{
instance<kHasUnevenSplits>::run(s, a);
}}
}} else {{
instance<kHasUnevenSplits>::run(s, a);
}}
}}
}} // anonymous namespace
#pragma clang diagnostic pop
template<>
void fmha_fwd_decode_oneshot_<trait_{F_idx}>(const ck_tile::stream_config& s, fmha_fwd_decode_args a)
{{
if constexpr({F_mode} == false) {{ // batch mode
// we don't check every seqlen_k values for kvcache
if (a.seqlen_k_ptr != nullptr) {{
run_instance</*kHasUnevenSplits=*/true>(s, a);
// make sure F_bn0 is divisible by F_bk1
}} else if (a.seqlen_k % (a.num_splits * {F_bn0}) == 0) {{
run_instance</*kHasUnevenSplits=*/false>(s, a);
}} else {{
run_instance</*kHasUnevenSplits=*/true>(s, a);
}}
}} else {{
run_instance</*kHasUnevenSplits=*/true>(s, a);
}}
}}
template<>
std::string fmha_fwd_decode_get_name_<trait_{F_idx}>()
{{
using k_ = instance<true>::fmha_kernel; /// FIXME: choose real kernel type
return k_::GetName();
}}
"""
FMHA_FWD_DECODE_COMBINE_KERNEL_BODY="""
using fmha_dtype_{F_idx} = {F_dtype};
namespace {{
template <ck_tile::index_t kLogMaxSplits>
struct instance {{
using fmha_trait = ck_tile::TileFmhaFwdSplitKVCombineTraits<{F_spad},
{F_dvpad},
{F_lse},
{F_squant},
kLogMaxSplits,
{F_occupancy}>;
using fmha_pipeline_problem = ck_tile::BlockFmhaSplitKVCombinePipelineProblem<
typename FmhaFwdTypeConfig<fmha_dtype_{F_idx}>::LSEDataType,
typename FmhaFwdTypeConfig<fmha_dtype_{F_idx}>::OaccDataType,
typename FmhaFwdTypeConfig<fmha_dtype_{F_idx}>::ODataType,
{F_hdim},
{F_mode},
{F_bn1},
fmha_trait>;
using fmha_pipeline = ck_tile::BlockFmhaFwdSplitKVCombinePipeline<
fmha_pipeline_problem>;
/// FIXME: use {F_spad}/{F_dvpad} as kPadM/kPadN parameters after solving
/// store_tile_raw() data corruption issue
using fmha_epilogue =
ck_tile::Default2DEpilogue<ck_tile::Default2DEpilogueProblem<typename FmhaFwdTypeConfig<{F_dtype}>::OaccDataType,
typename FmhaFwdTypeConfig<{F_dtype}>::ODataType,
false, false>>;
using fmha_kernel =
ck_tile::FmhaFwdSplitKVCombineKernel<fmha_pipeline, fmha_epilogue>;
static void run(const ck_tile::stream_config& s, fmha_fwd_decode_args a)
{{
using k_ = fmha_kernel;
auto [kargs, grids] = fmha_fwd_decode_combine_create_kargs_and_grids<k_>(a);
constexpr dim3 blocks = k_::BlockSize();
constexpr ck_tile::index_t kBlockPerCu = k_::kBlockPerCu;
ck_tile::make_kernel<blocks.x, kBlockPerCu>(k_{{}}, grids, blocks, 0, kargs)(ck_tile::stream_config{{s.stream_id_}});
}}
}};
}}
using trait_{F_idx} = fmha_fwd_decode_combine_traits_<{F_hdim}, {F_dtype}, {F_mode}, {F_bn1},
{F_lse}, {F_squant}, {F_spad}, {F_dvpad}>;
#include <iostream>
template<>
void fmha_fwd_decode_combine_oneshot_<trait_{F_idx}>(const ck_tile::stream_config& s, fmha_fwd_decode_args a)
{{
if (a.num_splits <= 8) {{
instance<3>::run(s, a);
}} else if (a.num_splits <= 16) {{
instance<4>::run(s, a);
}} else if (a.num_splits <= 32) {{
instance<5>::run(s, a);
}} else if (a.num_splits <= 64) {{
instance<6>::run(s, a);
}} else if (a.num_splits <= 128) {{
instance<7>::run(s, a);
}}
}}
template<>
std::string fmha_fwd_decode_combine_get_name_<trait_{F_idx}>()
{{
using k_ = instance<6>::fmha_kernel; /// FIXME: choose real kernel type
return k_::GetName();
}}
"""
FMHA_FWD_DECODE_API_FILENAME="fmha_fwd_decode_api.cpp"
FMHA_FWD_DECODE_API="""
#include <iostream>
template<typename fmha_fwd_decode_traits_, typename fmha_fwd_decode_combine_traits_>
float fmha_fwd_decode_(const ck_tile::stream_config& s, fmha_fwd_decode_args a)
{{
if(s.log_level_ > 0)
{{
if (a.num_splits > 1)
{{
std::cout
<< ", " << fmha_fwd_decode_get_name_<fmha_fwd_decode_traits_>()
<< ", " << fmha_fwd_decode_combine_get_name_<fmha_fwd_decode_combine_traits_>()
<< std::flush;
}}
else{{
std::cout
<< ", " << fmha_fwd_decode_get_name_<fmha_fwd_decode_traits_>()
<< std::flush;
}}
}}
// we don't need combine kernel when we don't split the kv.
if (a.num_splits > 1)
{{
return ck_tile::launch_kernel(s,
[=](const ck_tile::stream_config& s_){{ fmha_fwd_decode_oneshot_<fmha_fwd_decode_traits_>(s_, a); }},
[=](const ck_tile::stream_config& s_){{ fmha_fwd_decode_combine_oneshot_<fmha_fwd_decode_combine_traits_>(s_, a); }}
);
}}
else{{
return ck_tile::launch_kernel(s,
[=](const ck_tile::stream_config& s_){{ fmha_fwd_decode_oneshot_<fmha_fwd_decode_traits_>(s_, a); }}
);
}}
}}
float fmha_fwd_decode(fmha_fwd_decode_traits t, fmha_fwd_decode_args a, const ck_tile::stream_config& s){{
float r = -1;
{F_dispatch}
return r;
}}
"""
FMHA_FWD_DECODE_API_INNER_DISPATCH=""" {F_if}((t.is_group_mode == {F_mode}) && (t.is_v_rowmajor == {F_vlayout}) && (t.has_logits_soft_cap == {F_logits}) && ({F_mask_check}) && (t.bias_type == {F_bias_check}) && (t.has_lse == {F_lse}) && (t.do_fp8_static_quant == {F_squant}) &&
((a.block_table_ptr != nullptr) == {F_pagedkv}) && ({F_scheck}) && ({F_skcheck}) && ({F_dcheck}) && ({F_dvcheck})) {{
using traits_ = fmha_fwd_decode_traits_<{F_hdim}, {F_dtype}, {F_mode}, {F_bm0}, {F_bn0}, {F_bk0}, {F_bn1}, {F_bk1}, {F_bk0max}, {F_vlayout}, {F_pipeline_enum}, {F_logits}, {F_mask}, {F_bias}, {F_lse}, {F_squant}, {F_pagedkv}, {F_spad}, {F_skpad}, {F_dpad}, {F_dvpad}>;
// get combine kernel tile sizes
using OaccDataType = typename FmhaFwdTypeConfig<{F_dtype}>::OaccDataType;
constexpr ck_tile::index_t kM0 = ck_tile::BlockFmhaSplitKVCombinePipelineTileSizes<OaccDataType, /*F_bn1=*/32>::kM0;
// make sure we can reuse the padding flags in combine kernels
static_assert({F_bm0} % kM0 == 0);
static_assert({F_bn1} % 32 == 0);
if (t.has_lse) {{
if constexpr (std::is_same_v<{F_dtype}, FmhaFwdFp8>) {{
return -1;
}} else {{
using traits2_ = fmha_fwd_decode_combine_traits_<{F_hdim}, {F_dtype}, {F_mode}, /*F_bn1=*/32, true, {F_squant}, {F_spad}, {F_dvpad}>;
return fmha_fwd_decode_<traits_, traits2_>(s, a);
}}
}} else {{
using traits2_ = fmha_fwd_decode_combine_traits_<{F_hdim}, {F_dtype}, {F_mode}, /*F_bn1=*/32, false, {F_squant}, {F_spad}, {F_dvpad}>;
return fmha_fwd_decode_<traits_, traits2_>(s, a);
}}
}}
"""
@dataclass
class FmhaFwdDecodeApiTrait:
pipeline_tag : str
# sync with fmha_fwd_traits<>, to generate fallback calls
hdim : str
dtype : str # data type
mode : str # value from MODE_MAP
bm0 : int # tile size along q seqlen (block size)
bn0 : int # tile size along qk seqlen
bk0 : int # tile size along qk gemm unroll
bn1 : int # tile size along v head_dim
bk1 : int # tile size along kv gemm unroll
bk0max : int
vlayout : str
mask : str
logits : str
bias : str #
lse : str #
squant : str #
spad : str
skpad : str
dpad : str
dvpad : str
pagedkv : str
@property
def name(self) -> str:
return f'{self.hdim}-{self.dtype}-{self.mode}-{self.bm0}-{self.bn0}-{self.bk0}-{self.bn0}-{self.bk1}-{self.bk0max}-'+\
f'{self.vlayout}-{self.logits}-{self.mask}-{self.bias}-{self.lse}-{self.squant}-{self.spad}-{self.skpad}-{self.dpad}-'+\
f'{self.dvpad}-{self.pagedkv}'
@property
def scheck(self) -> str:
if self.mode == 'group': return 'true/*group mode spad always true*/' # group mode only generate spad/skpad == true
if self.pipeline_tag in ['decode_qr', 'qr_nwarp_sshuffle']:
if self.spad == 't' : return 'true'
else : return 'true'
else: assert False
@property
def skcheck(self) -> str:
if self.mode == 'group': return 'true/*group mode skpad always true*/' # group mode only generate spad/skpad == true
if self.pipeline_tag in ['decode_qr', 'qr_nwarp_sshuffle']:
if self.skpad == 't' : return 'true' # TODO: order of get_pipelines() matters! (ugly)
else : return 'true'
else: assert False
@property
def dcheck(self) -> str:
if self.pipeline_tag in ['decode_qr', 'qr_nwarp_sshuffle']:
bk0submax = K0_MAX_SUBMAX_MAP[self.bk0max]
if self.dpad == 't': return f'true /*a.hdim_q % {bk0submax} != 0*/' # TODO: order of get_pipelines() matters! (ugly)
else : return f'a.hdim_q % {bk0submax} == 0'
else: assert False
@property
def dvcheck(self) -> str:
if self.pipeline_tag in ['decode_qr', 'qr_nwarp_sshuffle']:
bk0submax = K0_MAX_SUBMAX_MAP[self.bk0max]
if self.dvpad == 't': return f'true /*a.hdim_v % {bk0submax} != 0*/' # TODO: order of get_pipelines() matters! (ugly)
else : return f'a.hdim_v % {bk0submax} == 0'
else: assert False
@dataclass
class FmhaFwdDecodePipeline:
tag : str
F_vlayout : str # row/col
F_spad : str # true/false
F_skpad : str #
F_dpad : str #
F_dvpad : str #
F_logits : str # t/f
F_bias : str # true/false
F_lse : str #
F_squant : str #
F_pagedkv : str # t/f
F_mask : str # value from MASK_MAP
@property
def name(self) -> str:
def pad_name() -> str:
n = ''
if self.F_spad == 't': n += 's'
if self.F_skpad == 't' : n += 'sk'
if self.F_dpad == 't' : n += 'd'
if self.F_dvpad == 't' : n += 'dv'
if n != '' : n = 'p' + n
return n
pn = pad_name()
n = f'{self.tag}_v{self.F_vlayout[0]}'
if pn != '' : n += f'_{pn}'
else: n += '_npad'
if self.F_logits == 't' : n += '_logits'
else: n += '_nlogits'
if self.F_bias != 'no' : n += f'_{self.F_bias}'
else: n += '_nbias'
if self.F_mask[0:2] == 's_':
if self.F_mask == 's_mask': n += f'_mask'
else: n += '_nmask'
else:
if self.F_mask != 'no' : n += f'_m{self.F_mask[0]}'
else: n += '_nmask'
if self.F_lse == 't' : n += '_lse'
else: n += '_nlse'
if self.F_squant == 't' : n += '_squant'
else: n += '_nsquant'
if self.F_pagedkv == 't' : n += '_pagedkv'
else: n += '_npagedkv'
return n
@dataclass
class FmhaFwdSplitKVCombinePipeline:
tag : str
F_spad : str # true/false
F_dvpad : str #
F_lse : str #
F_squant : str #
@property
def name(self) -> str:
def pad_name() -> str:
n = ''
if self.F_spad == 't': n += 's'
if self.F_dvpad == 't' : n += 'dv'
if n != '' : n = 'p' + n
return n
pn = pad_name()
n = f'{self.tag}'
if pn != '' : n += f'_{pn}'
else: n += '_npad'
if self.F_lse == 't' : n += '_lse'
else: n += '_nlse'
if self.F_squant == 't' : n += '_squant'
else: n += '_nsquant'
return n
class FmhaFwdDecodeApiPool:
def __init__(self, mask_impl):
self.pool = dict()
self.mask_impl = mask_impl
def register_traits(self, trait : FmhaFwdDecodeApiTrait) -> None:
# TODO: do we need to check duplication?
if trait.dtype not in self.pool.keys():
self.pool[trait.dtype] = dict()
if trait.hdim not in self.pool[trait.dtype].keys():
self.pool[trait.dtype][trait.hdim] = list()
self.pool[trait.dtype][trait.hdim].append(copy.copy(trait))
@property
def api(self) -> str:
per_dtypes=str()
for i, dtype in enumerate(self.pool.keys()):
per_hdim_case=str()
for j, hdim in enumerate(self.pool[dtype].keys()):
traits=self.pool[dtype][hdim]
inners=str()
for k, trait in enumerate(traits):
if_k = 'if' if k == 0 else 'else if'
inners = inners + FMHA_FWD_DECODE_API_INNER_DISPATCH.format(F_if=if_k, F_mode=MODE_MAP[trait.mode], F_vlayout=LAYOUT_MAP[trait.vlayout],
F_pipeline_enum=PIPELINE_ENUM_MAP[trait.pipeline_tag], F_logits=BOOL_MAP[trait.logits], F_mask=get_mask_map(self.mask_impl)[trait.mask],
F_mask_check=get_mask_check_map(self.mask_impl)[trait.mask], F_bias_check=BIAS_CHECK_MAP[trait.bias], F_bias=BIAS_MAP[trait.bias],
F_lse=BOOL_MAP[trait.lse], F_squant=BOOL_MAP[trait.squant], F_pagedkv=BOOL_MAP[trait.pagedkv],
F_scheck=trait.scheck, F_skcheck=trait.skcheck, F_dcheck=trait.dcheck, F_dvcheck=trait.dvcheck,
F_spad=BOOL_MAP[trait.spad], F_skpad=BOOL_MAP[trait.skpad], F_dpad=BOOL_MAP[trait.dpad], F_dvpad=BOOL_MAP[trait.dvpad],
F_bm0=trait.bm0, F_bn0=trait.bn0, F_bk0=trait.bk0, F_bn1=trait.bn1, F_bk1=trait.bk1, F_bk0max=trait.bk0max,
F_hdim=hdim, F_dtype=FWD_DTYPE_MAP[dtype])
if_j = 'if' if j == 0 else 'else if'
per_hdim_case = per_hdim_case + FMHA_FWD_API_PER_HDIM_CASE.format(F_if=if_j, F_hdim=hdim, F_hdim_v=hdim, F_inner_dispatch=inners)
if_i = 'if' if i == 0 else 'else if'
per_dtypes = per_dtypes + FMHA_FWD_API_PER_DTYPE.format(F_if=if_i, F_dtype=dtype, F_hdim_case=per_hdim_case)
if not per_dtypes:
# empty string we add some ignore to suppress warning in api
per_dtypes += ' (void)t ; (void)s ; (void)a;'
return FMHA_FWD_KERNEL_HEADER + FMHA_FWD_DECODE_API.format(F_dispatch = per_dtypes)
@dataclass
class FmhaFwdSplitKVCombineTileSize:
F_bn1 : int # tile size along v head_dim
F_occupancy : int # occupancy, -1 will let pipeline decide the occupancy, other value will overwrite occupancy
@property
def name(self) -> str:
return f"b{self.F_bn1}" +\
("" if self.F_occupancy == -1 else f"_o{self.F_occupancy}")
@dataclass
class FmhaFwdDecodeKernel:
F_idx : int # this is not a tunable, but a counter to differentiate symbol
F_hdim : int # hdim
F_dtype : str # data type
F_mode : str # value from MODE_MAP
F_tile : FmhaFwdTileSize
F_pipeline : FmhaFwdDecodePipeline
mask_impl : str
@property
def template(self) -> str:
kernel_body = str()
return FMHA_FWD_KERNEL_HEADER + \
FMHA_FWD_DECODE_KERNEL_BODY.format(
F_idx = self.F_idx,
F_hdim = self.F_hdim,
F_dtype = FWD_DTYPE_MAP[self.F_dtype],
F_bm0 = self.F_tile.F_bm0,
F_bn0 = self.F_tile.F_bn0,
F_bk0 = self.F_tile.F_bk0,
F_bn1 = self.F_tile.F_bn1,
F_bk1 = self.F_tile.F_bk1,
F_bk0max = self.F_tile.F_bk0max,
F_rm0 = self.F_tile.F_rm0,
F_rn0 = self.F_tile.F_rn0,
F_rk0 = self.F_tile.F_rk0,
F_rm1 = self.F_tile.F_rm1,
F_rn1 = self.F_tile.F_rn1,
F_rk1 = self.F_tile.F_rk1,
F_wm0 = self.F_tile.F_wm0,
F_wn0 = self.F_tile.F_wn0,
F_wk0 = self.F_tile.F_wk0,
F_wm1 = self.F_tile.F_wm1,
F_wn1 = self.F_tile.F_wn1,
F_wk1 = self.F_tile.F_wk1,
F_vlayout = LAYOUT_MAP[self.F_pipeline.F_vlayout],
F_spad = BOOL_MAP[self.F_pipeline.F_spad],
F_skpad = BOOL_MAP[self.F_pipeline.F_skpad],
F_dpad = BOOL_MAP[self.F_pipeline.F_dpad],
F_dvpad = BOOL_MAP[self.F_pipeline.F_dvpad],
F_logits = BOOL_MAP[self.F_pipeline.F_logits],
F_bias = BIAS_MAP[self.F_pipeline.F_bias],
F_lse = BOOL_MAP[self.F_pipeline.F_lse],
F_squant = BOOL_MAP[self.F_pipeline.F_squant],
F_pagedkv = BOOL_MAP[self.F_pipeline.F_pagedkv],
F_occupancy = self.F_tile.F_occupancy,
F_pipeline_enum = PIPELINE_ENUM_MAP[self.F_pipeline.tag],
F_mask = get_mask_map(self.mask_impl)[self.F_pipeline.F_mask],
F_mode = MODE_MAP[self.F_mode],
F_pipeline = FMHA_FWD_DECODE_PIPELINE_MAP[self.F_pipeline.tag])
@property
def name(self) -> str:
# TODO: we don't encode idx here
return f"fmha_fwd_decode_d{self.F_hdim}_{self.F_dtype}_{self.F_mode}_" + \
self.F_tile.name + '_' + self.F_pipeline.name
@property
def filename(self) -> str:
return self.name + ".cpp"
def api_trait(self) -> FmhaFwdDecodeApiTrait:
return FmhaFwdDecodeApiTrait(
pipeline_tag=self.F_pipeline.tag,
hdim=str(self.F_hdim),
dtype=self.F_dtype,
mode=self.F_mode,
bm0=self.F_tile.F_bm0,
bn0=self.F_tile.F_bn0,
bk0=self.F_tile.F_bk0,
bn1=self.F_tile.F_bn1,
bk1=self.F_tile.F_bk1,
bk0max=self.F_tile.F_bk0max,
vlayout=self.F_pipeline.F_vlayout,
logits=self.F_pipeline.F_logits,
mask=self.F_pipeline.F_mask,
bias=self.F_pipeline.F_bias,
lse=self.F_pipeline.F_lse,
squant=self.F_pipeline.F_squant,
pagedkv=self.F_pipeline.F_pagedkv,
spad=self.F_pipeline.F_spad,
skpad=self.F_pipeline.F_skpad,
dpad=self.F_pipeline.F_dpad,
dvpad=self.F_pipeline.F_dvpad)
@dataclass
class FmhaFwdSplitKVCombineKernel:
F_idx : int # this is not a tunable, but a counter to differentiate symbol
F_hdim : int # hdim
F_dtype : str # data type
F_mode : str # value from MODE_MAP
F_tile : FmhaFwdSplitKVCombineTileSize
F_pipeline : FmhaFwdSplitKVCombinePipeline
@property
def template(self) -> str:
kernel_body = str()
return FMHA_FWD_KERNEL_HEADER + \
FMHA_FWD_DECODE_COMBINE_KERNEL_BODY.format(
F_idx = self.F_idx,
F_hdim = self.F_hdim,
F_dtype = FWD_DTYPE_MAP[self.F_dtype],
F_bn1 = self.F_tile.F_bn1,
F_spad = BOOL_MAP[self.F_pipeline.F_spad],
F_dvpad = BOOL_MAP[self.F_pipeline.F_dvpad],
F_lse = BOOL_MAP[self.F_pipeline.F_lse],
F_squant = BOOL_MAP[self.F_pipeline.F_squant],
F_occupancy = self.F_tile.F_occupancy,
F_mode = MODE_MAP[self.F_mode])
@property
def name(self) -> str:
# TODO: we don't encode idx here
return f"fmha_fwd_decode_combine_d{self.F_hdim}_{self.F_dtype}_{self.F_mode}_" + \
self.F_tile.name + '_' + self.F_pipeline.name
@property
def filename(self) -> str:
return self.name + ".cpp"
# TODO: design a more practical way to do it
# this is current supported tile size per hdim
def get_fmha_fwd_tile_dict_from_dtype(dtype : str) -> Optional[dict]:
if dtype == 'fp16' or dtype == 'bf16':
return {
# '32' : FmhaFwdTileSize(32, 64, 16, 32, 32, 32, 2, 1, 1, 2, 1, 1, 16, 16, 32, 16, 16, 32, -1),
'64' : FmhaFwdTileSize(16, 64, 64, 64, 64, 64, 1, 4, 1, 1, 4, 1, 16, 16, 32, 16, 16, 32, -1),
# '64' : FmhaFwdTileSize(32, 64, 64, 64, 64, 64, 1, 4, 1, 1, 4, 1, 16, 16, 32, 16, 16, 32, -1),
# '64' : FmhaFwdTileSize(64, 64, 64, 64, 64, 64, 1, 4, 1, 1, 4, 1, 16, 16, 32, 16, 16, 32, -1),
# '64' : FmhaFwdTileSize(128, 64, 64, 64, 64, 64, 1, 4, 1, 1, 4, 1, 16, 16, 32, 16, 16, 32, -1),
# '64' : FmhaFwdTileSize(256, 64, 64, 64, 64, 64, 1, 4, 1, 1, 4, 1, 16, 16, 32, 16, 16, 32, -1),
### '96' : FmhaFwdTileSize(64, 128, 32, 128, 32, 96, 4, 1, 1, 4, 1, 1, 16, 16, 16, 16, 16, 16, -1),
'128' : FmhaFwdTileSize(16, 64, 64, 128, 64, 128, 1, 4, 1, 1, 4, 1, 16, 16, 32, 16, 16, 32, -1),
# '128' : FmhaFwdTileSize(32, 64, 64, 128, 64, 128, 1, 4, 1, 1, 4, 1, 16, 16, 32, 16, 16, 32, -1),
# '128' : FmhaFwdTileSize(64, 64, 64, 128, 64, 128, 1, 4, 1, 1, 4, 1, 16, 16, 32, 16, 16, 32, -1),
# '128' : FmhaFwdTileSize(128, 64, 64, 128, 64, 128, 1, 4, 1, 1, 4, 1, 16, 16, 32, 16, 16, 32, -1),
# '128' : FmhaFwdTileSize(256, 64, 64, 128, 64, 128, 1, 4, 1, 1, 4, 1, 16, 16, 32, 16, 16, 32, -1),
# '256' : FmhaFwdTileSize(64, 128, 32, 256, 32, 256, 4, 1, 1, 4, 1, 1, 16, 16, 32, 16, 16, 32, -1),
}
else:
return None
def get_fmha_fwd_decode_combine_tile_dict_from_dtype(dtype : str) -> Optional[dict]:
if dtype == 'fp16' or dtype == 'bf16':
return {
# '32' : FmhaFwdSplitKVCombineTileSize(32, -1),
'64' : FmhaFwdSplitKVCombineTileSize(32, -1),
### '96' : FmhaFwdSplitKVCombineTileSize(32, -1),
'128' : FmhaFwdSplitKVCombineTileSize(32, -1),
# '256' : FmhaFwdSplitKVCombineTileSize(32, -1),
}
else:
return None
def get_fwd_decode_blobs(kernel_filter : Optional[str], receipt, mask_impl) -> Tuple[FmhaFwdDecodeApiPool, List[FmhaFwdDecodeKernel]]:
Pipeline = FmhaFwdDecodePipeline
Kernel = FmhaFwdDecodeKernel
# TODO: we don't support tuning yet, so pick up one value for vlayout/pipeline/pad
# support this in future
def get_pipelines(dtype, hdim) -> List[FmhaFwdDecodePipeline]:
# this function will populate a list possible pipelines
# TODO: the order of List matters! the later in this list will be also be checked later
# TODO: currently for qr pipeline, let 't' padding to appear later!!
# TODO: how to design this more generic?
squant = 't' if dtype == 'fp8' else 'f'
pipelines = []
if dtype in ['fp16', 'bf16']:
for logits, mask, bias, pagedkv in itertools.product(["f"], get_mask_map(mask_impl).keys(), BIAS_MAP.keys(), ["f"]):
for lse in ['t', 'f']:
if hdim in [64, 128]: ### [32, 64, 96, 128]:
pipelines.append(Pipeline('decode_qr', 'row', 'f', 'f', 'f', 'f', logits, bias, lse, squant, pagedkv, mask))
else:
assert False
else:
assert False
return pipelines
gen = list()
api_pool = FmhaFwdDecodeApiPool(mask_impl)
for dtype in FWD_DTYPE_MAP.keys():
d = get_fmha_fwd_tile_dict_from_dtype(dtype)
if d == None:
continue
#for hdim_str, mode, mask, bias, lse in itertools.product(d.keys(), MODE_MAP.keys(), MASK_MAP.keys(), ["t", "f"], ["t", "f"]):
for hdim_str, mode in itertools.product(d.keys(), MODE_MAP.keys()):
tile = d[hdim_str]
hdim = int(hdim_str)
for pipeline in get_pipelines(dtype, hdim):
if mode == "group":
if pipeline.F_spad != 't' or pipeline.F_skpad != 't':
# in group mode, spad/skpad must be true, since we can't predict if seqlen of current batch need pad or not
continue
# logits_soft_cap is only allowed if no bias
if not ((pipeline.F_logits == 't' and pipeline.F_bias == 'no') or pipeline.F_logits == 'f'):
continue
k = Kernel(F_idx=0,
F_hdim=hdim,
F_dtype=dtype,
F_mode=mode,
F_tile=tile,
F_pipeline=pipeline,
mask_impl=mask_impl)
if kernel_filter != '':
if not fnmatch.fnmatch(k.name, kernel_filter):
continue
# Flash attention integration
if receipt == 2:
cond = dtype in ['fp16', 'bf16']
cond &= pipeline.F_vlayout == 'row'
cond &= pipeline.F_bias in ['no', 'alibi']
cond &= pipeline.F_squant == 'f'
if not cond:
continue
# Aiter(mha_varlen_fwd) integration
elif receipt == 200:
cond = dtype in ['fp16', 'bf16']
cond &= mode == "group"
cond &= pipeline.F_vlayout == 'row'
cond &= pipeline.F_squant == 'f'
if not cond:
continue
# aiter::mha_fwd_splikv C++ api integration
elif receipt == 600:
cond = dtype in ['fp16', 'bf16']
cond &= pipeline.F_vlayout == 'row'
cond &= pipeline.F_squant == 'f'
if not cond:
continue
api_pool.register_traits(k.api_trait())
gen.append(k)
return (api_pool, gen)
def get_fwd_decode_combine_blobs(kernel_filter : Optional[str], receipt) -> List[FmhaFwdSplitKVCombineKernel]:
Pipeline = FmhaFwdSplitKVCombinePipeline
Kernel = FmhaFwdSplitKVCombineKernel
# TODO: we don't support tuning yet, so pick up one value for vlayout/pipeline/pad
# support this in future
def get_pipelines(dtype, hdim) -> List[FmhaFwdSplitKVCombinePipeline]:
# this function will populate a list possible pipelines
# TODO: the order of List matters! the later in this list will be also be checked later
# TODO: currently for qr pipeline, let 't' padding to appear later!!
# TODO: how to design this more generic?
squant = 't' if dtype == 'fp8' else 'f'
pipelines = []
if dtype in ['fp16', 'bf16']:
# for spad, dvpad, lse in itertools.product(["t", "f"], ["t", "f"], ["t", "f"]):
for spad, dvpad, lse in itertools.product(["f"], ["f"], ["t", "f"]):
pipelines.append(Pipeline('unused', spad, dvpad, lse, squant))
elif dtype in ['fp8', 'bf8']:
# no need lse kernels
pipelines.append(Pipeline('unused', 'f', 'f', 'f', squant))
else:
assert False
return pipelines
gen = list()
for dtype in FWD_DTYPE_MAP.keys():
d = get_fmha_fwd_decode_combine_tile_dict_from_dtype(dtype)
if d == None:
continue
#for hdim_str, mode, mask, bias, lse in itertools.product(d.keys(), MODE_MAP.keys(), MASK_MAP.keys(), ["t", "f"], ["t", "f"]):
for hdim_str, mode in itertools.product(d.keys(), MODE_MAP.keys()):
tile = d[hdim_str]
hdim = int(hdim_str)
for pipeline in get_pipelines(dtype, hdim):
if mode == "group":
if pipeline.F_spad != 't':
# in group mode, spad/skpad must be true, since we can't predict if seqlen of current batch need pad or not
continue
k = Kernel(F_idx=0,
F_hdim=hdim,
F_dtype=dtype,
F_mode=mode,
F_tile=tile,
F_pipeline=pipeline)
if kernel_filter != '':
if not fnmatch.fnmatch(k.name, kernel_filter):
continue
# Aiter(mha_varlen_fwd) integration
if receipt == 200:
cond = dtype in ['fp16', 'bf16']
cond &= mode == "group"
if not cond:
continue
# aiter::mha_fwd_splikv C++ api integration
elif receipt == 600:
cond = dtype in ['fp16', 'bf16']
if not cond:
continue
gen.append(k)
return gen
def write_single_kernel(kernel: Union[FmhaFwdDecodeKernel, FmhaFwdSplitKVCombineKernel], autogen_dir: Path) -> None:
(autogen_dir / kernel.filename).write_text(kernel.template)
def write_fwd_decode_api(api_pool : FmhaFwdDecodeApiPool, autogen_dir: Path) -> None:
file_path = autogen_dir / FMHA_FWD_DECODE_API_FILENAME
file_path.write_text(api_pool.api)
def write_blobs(output_dir : Path, filter_list : str, receipt, optdim_list, mask_impl) -> None:
filter_list = filter_list.split('@')
filter_list.extend([''] * (2 - len(filter_list)))
assert optdim_list == [-1]
kernels = get_fwd_decode_combine_blobs(filter_list[0], receipt)
for kernel in kernels:
write_single_kernel(kernel, output_dir)
api_pool, kernels = get_fwd_decode_blobs(filter_list[1], receipt, mask_impl)
for kernel in kernels:
write_single_kernel(kernel, output_dir)
write_fwd_decode_api(api_pool, output_dir)
def list_blobs(file_path : Path, filter_list : str, receipt, optdim_list, mask_impl) -> None:
filter_list = filter_list.split('@')
filter_list.extend([''] * (2 - len(filter_list)))
assert optdim_list == [-1]
with file_path.open('a') as f:
kernels = get_fwd_decode_combine_blobs(filter_list[0], receipt)
for kernel in kernels:
f.write(str(file_path.parent / GEN_DIR / kernel.filename) + "\n")
_, kernels = get_fwd_decode_blobs(filter_list[1], receipt, mask_impl)
for kernel in kernels:
f.write(str(file_path.parent / GEN_DIR / kernel.filename) + "\n")
f.write(str(file_path.parent / GEN_DIR / FMHA_FWD_DECODE_API_FILENAME) + "\n")

15
example/ck_tile/01_fmha/fmha_fwd.cpp
View File

@@ -823,7 +823,7 @@ bool run(const ck_tile::ArgParser& arg_parser)
<< (is_rotary_interleaved ? "inter" : "half") << ")";
}
#endif
#if CK_TILE_FMHA_FWD_SPLITKV_API
#if CK_TILE_FMHA_FWD_SPLITKV_API || CK_TILE_FMHA_FWD_DECODE_API
if(1 < num_splits)
{
std::cout << ", num_splits:" << num_splits;
@@ -1048,7 +1048,7 @@ bool run(const ck_tile::ArgParser& arg_parser)
args.drop_seed_offset = std::make_pair(drop_seed, drop_offset);
}
}
else if constexpr(std::is_same_v<fmha_fwd_splitkv_args, std::decay_t<decltype(args)>>)
else if constexpr(std::is_same_v<fmha_fwd_splitkv_args, std::decay_t<decltype(args)>> || std::is_same_v<fmha_fwd_decode_args, std::decay_t<decltype(args)>>)
{
args.lse_acc_ptr = lse_acc_buf.GetDeviceBuffer();
args.o_acc_ptr = o_acc_buf.GetDeviceBuffer();
@@ -1103,7 +1103,15 @@ bool run(const ck_tile::ArgParser& arg_parser)
return fmha_fwd_splitkv(fmha_splitkv_traits, fmha_splitkv_args, stream_config);
}
#endif
#elif CK_TILE_FMHA_FWD_DECODE_API
fmha_fwd_decode_traits fmha_decode_traits;
init_traits(fmha_decode_traits);
fmha_fwd_decode_args fmha_decode_args;
init_args(fmha_decode_args);
return fmha_fwd_decode(fmha_decode_traits, fmha_decode_args, stream_config);
#else
fmha_fwd_traits fmha_traits;
init_traits(fmha_traits);
@@ -1111,6 +1119,7 @@ bool run(const ck_tile::ArgParser& arg_parser)
init_args(fmha_args);
return fmha_fwd(fmha_traits, fmha_args, stream_config);
#endif
}();
if(appendkv_ave_time < 0.0f || fwd_ave_time < 0.0f)

348
example/ck_tile/01_fmha/fmha_fwd.hpp
View File

@@ -392,6 +392,95 @@ struct fmha_batch_prefill_args
drop_seed_offset;
};
struct fmha_fwd_decode_args
{
const void* q_ptr;
const void* k_ptr;
const void* v_ptr;
const void* bias_ptr; // bias or alibi_slope pointer
void* lse_acc_ptr;
void* o_acc_ptr;
void* lse_ptr;
void* o_ptr;
void* block_table_ptr;
ck_tile::index_t batch_stride_block_table; // only used if 'block_table_ptr' is not nullptr
ck_tile::index_t page_block_size; // only used if 'block_table_ptr' is not nullptr
bool is_gappy; // differentiate seqstart_k_ptr usage. only used if 'block_table_ptr' is not
// nullptr.
const void* cache_batch_idx;
// the real seqlen_q & seqlen_k are decided by following:
// batch mode: seqlen_q = kargs.seqlen_q
// seqlen_k = kargs.seqlen_k
// group mode: seqlen_q = kargs.seqstart_q_ptr[b + 1] - kargs.seqstart_q_ptr[b]
// seqlen_k = kargs.seqstart_k_ptr[b + 1] - kargs.seqstart_k_ptr[b]
// or kargs.seqlen_k_ptr[b]
//
// batch mode (kvcache):
// seqlen_q = kargs.seqlen_q
// seqlen_k = kargs.seqlen_k_ptr[b]
// group mode (kvcache):
// seqlen_q = kargs.seqstart_q_ptr[b + 1] - kargs.seqstart_q_ptr[b]
//
// when is_gappy=true:
// seqlen_k = kargs.seqlen_k_ptr[b]
// seqstart_k_ptr[b] now store local offset of each batch
//
// when is_gappy=false:
// seqlen_k = kargs.seqstart_k_ptr[b + 1] - kargs.seqstart_k_ptr[b]
// or kargs.seqlen_k_ptr[b]
const void* seqstart_q_ptr;
const void* seqstart_k_ptr;
const void* seqlen_k_ptr;
ck_tile::index_t seqlen_q;
ck_tile::index_t seqlen_k;
ck_tile::index_t batch;
ck_tile::index_t max_seqlen_q;
ck_tile::index_t hdim_q;
ck_tile::index_t hdim_v;
ck_tile::index_t nhead_q;
ck_tile::index_t nhead_k;
ck_tile::index_t num_splits;
float scale_s;
float scale_p;
float scale_o;
float logits_soft_cap;
ck_tile::index_t stride_q;
ck_tile::index_t stride_k;
ck_tile::index_t stride_v;
ck_tile::index_t stride_bias; // if alibi, b*h need set this to h, 1*h need set this to 0
ck_tile::index_t stride_o_acc;
ck_tile::index_t stride_o;
ck_tile::index_t nhead_stride_q;
ck_tile::index_t nhead_stride_k;
ck_tile::index_t nhead_stride_v;
ck_tile::index_t nhead_stride_bias;
ck_tile::index_t nhead_stride_lse;
ck_tile::index_t nhead_stride_lse_acc;
ck_tile::index_t nhead_stride_o_acc;
ck_tile::index_t nhead_stride_o;
ck_tile::index_t batch_stride_q;
ck_tile::index_t batch_stride_k;
ck_tile::index_t batch_stride_v;
ck_tile::index_t batch_stride_bias;
ck_tile::index_t batch_stride_lse;
ck_tile::index_t batch_stride_lse_acc;
ck_tile::index_t batch_stride_o_acc;
ck_tile::index_t batch_stride_o;
ck_tile::index_t split_stride_lse_acc;
ck_tile::index_t split_stride_o_acc;
ck_tile::index_t window_size_left;
ck_tile::index_t window_size_right;
ck_tile::index_t mask_type;
};
template <typename FmhaKernel>
auto fmha_fwd_create_kargs_and_grids(fmha_fwd_args args)
{
@@ -662,6 +751,168 @@ auto fmha_fwd_splitkv_combine_create_kargs_and_grids(fmha_fwd_splitkv_args args)
return ck_tile::make_tuple(kargs, grids);
}
template <typename Kernel>
auto fmha_fwd_decode_create_kargs_and_grids(fmha_fwd_decode_args args)
{
assert(args.nhead_q % args.nhead_k == 0);
auto kargs = [&] {
// create group mode kernel arguments
if constexpr(Kernel::kIsGroupMode)
{
return Kernel::MakeKargs(args.q_ptr,
args.k_ptr,
args.v_ptr,
args.bias_ptr,
args.lse_acc_ptr,
args.o_acc_ptr,
args.batch,
args.seqstart_q_ptr,
args.seqstart_k_ptr,
args.seqlen_k_ptr,
args.hdim_q,
args.hdim_v,
args.nhead_q,
args.nhead_q / args.nhead_k,
args.num_splits,
args.block_table_ptr,
args.batch_stride_block_table,
args.page_block_size,
args.is_gappy,
args.scale_s,
args.scale_p,
args.logits_soft_cap,
args.stride_q,
args.stride_k,
args.stride_v,
args.stride_bias,
args.stride_o_acc,
args.nhead_stride_q,
args.nhead_stride_k,
args.nhead_stride_v,
args.nhead_stride_bias,
args.nhead_stride_lse_acc,
args.nhead_stride_o_acc,
args.batch_stride_k, // only used for paged-kvcache
args.batch_stride_v, // only used for paged-kvcache
args.split_stride_lse_acc,
args.split_stride_o_acc,
args.window_size_left,
args.window_size_right,
args.mask_type);
}
else
{ // create batch mode kernel arguments
return Kernel::MakeKargs(args.q_ptr,
args.k_ptr,
args.v_ptr,
args.bias_ptr,
args.lse_acc_ptr,
// args.o_acc_ptr,
args.o_ptr, // hardcoding
args.batch,
args.seqlen_q,
args.seqlen_k,
args.seqlen_k_ptr,
args.hdim_q,
args.hdim_v,
args.nhead_q,
args.nhead_q / args.nhead_k,
args.num_splits,
args.block_table_ptr,
args.batch_stride_block_table,
args.page_block_size,
args.cache_batch_idx,
args.scale_s,
args.scale_p,
args.logits_soft_cap,
args.stride_q,
args.stride_k,
args.stride_v,
args.stride_bias,
args.stride_o_acc,
args.nhead_stride_q,
args.nhead_stride_k,
args.nhead_stride_v,
args.nhead_stride_bias,
args.nhead_stride_lse_acc,
args.nhead_stride_o_acc,
args.batch_stride_q,
args.batch_stride_k,
args.batch_stride_v,
args.batch_stride_bias,
args.batch_stride_lse_acc,
args.batch_stride_o_acc,
args.split_stride_lse_acc,
args.split_stride_o_acc,
args.window_size_left,
args.window_size_right,
args.mask_type);
}
}();
dim3 grids = Kernel::GridSize(
args.batch, args.nhead_q, args.nhead_k, args.max_seqlen_q, args.hdim_v, args.num_splits);
return ck_tile::make_tuple(kargs, grids);
}
template <typename Kernel>
auto fmha_fwd_decode_combine_create_kargs_and_grids(fmha_fwd_decode_args args)
{
assert(args.nhead_q % args.nhead_k == 0);
auto kargs = [&] {
// create group mode kernel argumentszs
if constexpr(Kernel::kIsGroupMode)
{
return Kernel::MakeKargs(args.lse_acc_ptr,
args.o_acc_ptr,
args.lse_ptr,
args.o_ptr,
args.batch,
args.seqstart_q_ptr,
args.hdim_v,
args.num_splits,
args.scale_o,
args.stride_o_acc,
args.stride_o,
args.nhead_stride_lse_acc,
args.nhead_stride_o_acc,
args.nhead_stride_lse,
args.nhead_stride_o,
args.split_stride_lse_acc,
args.split_stride_o_acc);
}
else
{ // create batch mode kernel arguments
return Kernel::MakeKargs(args.lse_acc_ptr,
args.o_acc_ptr,
args.lse_ptr,
args.o_ptr,
args.batch,
args.seqlen_q,
args.hdim_v,
args.num_splits,
args.scale_o,
args.stride_o_acc,
args.stride_o,
args.nhead_stride_lse_acc,
args.nhead_stride_o_acc,
args.nhead_stride_lse,
args.nhead_stride_o,
args.batch_stride_lse_acc,
args.batch_stride_o_acc,
args.batch_stride_lse,
args.batch_stride_o,
args.split_stride_lse_acc,
args.split_stride_o_acc);
}
}();
dim3 grids = Kernel::GridSize(args.batch, args.nhead_q, args.max_seqlen_q, args.hdim_v);
return ck_tile::make_tuple(kargs, grids);
}
template <typename Kernel>
auto fmha_fwd_appendkv_create_kargs_and_grids(fmha_fwd_appendkv_args args)
{
@@ -948,6 +1199,84 @@ void fmha_fwd_splitkv_combine_oneshot_(const ck_tile::stream_config&, fmha_fwd_s
template <typename Traits_>
std::string fmha_fwd_splitkv_combine_get_name_();
template <ck_tile::index_t HDim_,
typename DataType_,
bool kIsGroupMode_,
ck_tile::index_t kM0_,
ck_tile::index_t kN0_,
ck_tile::index_t kK0_,
ck_tile::index_t kN1_,
ck_tile::index_t kK1_,
ck_tile::index_t kK0BlockLength_,
bool kIsVLayoutRowMajor_,
ck_tile::BlockFmhaPipelineEnum FmhaPipelineEnum_,
bool kHasLogitsSoftCap_,
typename FmhaMask_,
ck_tile::BlockAttentionBiasEnum BiasEnum_,
bool kStoreLse_,
bool kDoFp8StaticQuant_,
bool kIsPagedKV_,
bool kPadS_,
bool kPadSK_,
bool kPadD_,
bool kPadDv_>
struct fmha_fwd_decode_traits_
{
static constexpr ck_tile::index_t HDim = HDim_;
using DataType = ck_tile::remove_cvref_t<DataType_>;
static constexpr bool kIsGroupMode = kIsGroupMode_;
static constexpr ck_tile::index_t kM0 = kM0_;
static constexpr ck_tile::index_t kN0 = kN0_;
static constexpr ck_tile::index_t kK0 = kK0_;
static constexpr ck_tile::index_t kN1 = kN1_;
static constexpr ck_tile::index_t kK1 = kK1_;
static constexpr ck_tile::index_t kK0BlockLength = kK0BlockLength_;
static constexpr bool kIsVLayoutRowMajor = kIsVLayoutRowMajor_;
static constexpr auto FmhaPipelineEnum = FmhaPipelineEnum_;
static constexpr bool kHasLogitsSoftCap = kHasLogitsSoftCap_;
using FmhaMask = ck_tile::remove_cvref_t<FmhaMask_>;
static constexpr auto BiasEnum = BiasEnum_;
static constexpr bool kStoreLse = kStoreLse_;
static constexpr bool kDoFp8StaticQuant = kDoFp8StaticQuant_;
static constexpr bool kPadS = kPadS_;
static constexpr bool kPadSK = kPadSK_;
static constexpr bool kPadD = kPadD_;
static constexpr bool kPadDv = kPadDv_;
static constexpr bool kIsPagedKV = kIsPagedKV_;
};
template <typename Traits_>
void fmha_fwd_decode_oneshot_(const ck_tile::stream_config&, fmha_fwd_decode_args);
template <typename Traits_>
std::string fmha_fwd_decode_get_name_();
template <ck_tile::index_t HDim_,
typename DataType_,
bool kIsGroupMode_,
ck_tile::index_t kN1_,
bool kStoreLse_,
bool kDoFp8StaticQuant_,
bool kPadS_,
bool kPadDv_>
struct fmha_fwd_decode_combine_traits_
{
static constexpr ck_tile::index_t HDim = HDim_;
using DataType = ck_tile::remove_cvref_t<DataType_>;
static constexpr bool kIsGroupMode = kIsGroupMode_;
static constexpr ck_tile::index_t kN1 = kN1_;
static constexpr bool kStoreLse = kStoreLse_;
static constexpr bool kDoFp8StaticQuant = kDoFp8StaticQuant_;
static constexpr bool kPadS = kPadS_;
static constexpr bool kPadDv = kPadDv_;
};
template <typename Traits_>
void fmha_fwd_decode_combine_oneshot_(const ck_tile::stream_config&, fmha_fwd_decode_args);
template <typename Traits_>
std::string fmha_fwd_decode_combine_get_name_();
// this is used to pattern-match internl kernel implementation, not to instantiate kernel
template <ck_tile::index_t HDim_,
typename DataType_,
@@ -1022,6 +1351,25 @@ float fmha_fwd_splitkv(fmha_fwd_splitkv_traits,
fmha_fwd_splitkv_args,
const ck_tile::stream_config&);
struct fmha_fwd_decode_traits
{
int hdim_q;
int hdim_v;
std::string data_type;
bool is_group_mode;
bool is_v_rowmajor;
bool has_logits_soft_cap;
mask_enum mask_type;
bias_enum bias_type; // 0:no bias, 1:elementwise bias, 2:alibi. sync with BlockAttentionBiasEnum
bool has_lse;
bool do_fp8_static_quant;
// TODO: padding check is inside this api
};
float fmha_fwd_decode(fmha_fwd_decode_traits,
fmha_fwd_decode_args,
const ck_tile::stream_config&);
struct fmha_fwd_appendkv_traits
{
int hdim_q;