Skip to content
GitLab
Menu
Projects
Groups
Snippets
Loading...
Help
Help
Support
Community forum
Keyboard shortcuts
?
Submit feedback
Contribute to GitLab
Sign in / Register
Toggle navigation
Menu
Open sidebar
change
sglang
Commits
79a321af
"vscode:/vscode.git/clone" did not exist on "9add5d0af5f5c770033881177397a041da857d9a"
Unverified
Commit
79a321af
authored
Mar 09, 2025
by
Xiaoyu Zhang
Committed by
GitHub
Mar 08, 2025
Browse files
revert pr 3628 to pass test_mla ci (#4219)
parent
6eec3cdc
Changes
1
Show whitespace changes
Inline
Side-by-side
Showing
1 changed file
with
30 additions
and
46 deletions
+30
-46
sgl-kernel/src/sgl-kernel/csrc/gemm/per_token_group_quant_fp8.cu
...nel/src/sgl-kernel/csrc/gemm/per_token_group_quant_fp8.cu
+30
-46
No files found.
sgl-kernel/src/sgl-kernel/csrc/gemm/per_token_group_quant_fp8.cu
View file @
79a321af
...
...
@@ -2,18 +2,17 @@
#include <c10/util/Float8_e4m3fn.h>
#include <cmath>
#include <flashinfer/vec_dtypes.cuh>
#include "utils.h"
using
FP8_TYPE
=
c10
::
Float8_e4m3fn
;
__device__
__forceinline__
float
GroupReduce
(
float
val
,
const
int
tid
)
{
val
=
fmaxf
(
val
,
__shfl_xor_sync
(
0xffff
,
val
,
8
)
);
val
=
fmaxf
(
val
,
__shfl_xor_sync
(
0xffff
,
val
,
4
)
);
val
=
fmaxf
(
val
,
__shfl_xor_sync
(
0xffff
,
val
,
2
)
);
val
=
fmaxf
(
val
,
__shfl_xor_sync
(
0xffff
,
val
,
1
)
);
return
val
;
__device__
__forceinline__
float
GroupReduce
Max
(
volatile
float
*
smem
,
const
int
tid
)
{
smem
[
tid
]
=
fmaxf
(
smem
[
tid
],
smem
[
tid
+
8
]
);
if
(
tid
<
4
)
smem
[
tid
]
=
fmaxf
(
smem
[
tid
],
smem
[
tid
+
4
]
);
if
(
tid
<
2
)
smem
[
tid
]
=
fmaxf
(
smem
[
tid
],
smem
[
tid
+
2
]
);
if
(
tid
<
1
)
smem
[
tid
]
=
fmaxf
(
smem
[
tid
],
smem
[
tid
+
1
]
);
return
smem
[
0
]
;
}
template
<
typename
T
>
...
...
@@ -27,60 +26,45 @@ __global__ void per_token_group_quant_fp8_kernel(
const
float
fp8_min
,
const
float
fp8_max
)
{
const
int
groups_per_block
=
16
;
const
int
local_group_id
=
threadIdx
.
x
/
16
;
const
int
lane_id
=
threadIdx
.
x
%
16
;
const
int
block_group_id
=
blockIdx
.
x
*
groups_per_block
;
const
int
block_group_offset
=
(
block_group_id
+
local_group_id
)
*
group_size
;
const
int
tid
=
threadIdx
.
x
;
const
int
local_group_id
=
tid
/
16
;
const
int
local_tid
=
tid
%
16
;
__shared__
float
s_absmax
[
16
];
__shared__
float
s_absmax
[
16
]
[
17
]
;
float
local_absmax
=
eps
;
const
T
*
group_input
=
input
+
block_group_offset
;
FP8_TYPE
*
group_output
=
static_cast
<
FP8_TYPE
*>
(
output_q
)
+
block_group_offset
;
if
(
block_group_id
+
local_group_id
<
num_groups
)
{
const
T
*
group_input
=
input
+
(
block_group_id
+
local_group_id
)
*
group_size
;
FP8_TYPE
*
group_output
=
static_cast
<
FP8_TYPE
*>
(
output_q
)
+
(
block_group_id
+
local_group_id
)
*
group_size
;
float
*
scale_output
=
output_s
+
block_group_id
+
local_group_id
;
constexpr
uint32_t
vec_size
=
16
/
sizeof
(
T
);
using
vec_t
=
flashinfer
::
vec_t
<
T
,
vec_size
>
;
const
int32_t
num_vec_elems
=
group_size
/
vec_size
;
for
(
int32_t
i
=
lane_id
;
i
<
num_vec_elems
;
i
+=
16
)
{
vec_t
input_vec
;
input_vec
.
cast_load
(
group_input
+
i
*
vec_size
);
#pragma unroll
for
(
uint32_t
j
=
0
;
j
<
vec_size
;
++
j
)
{
float
val
=
static_cast
<
float
>
(
input_vec
[
j
]);
for
(
int
i
=
local_tid
;
i
<
group_size
;
i
+=
16
)
{
float
val
=
static_cast
<
float
>
(
group_input
[
i
]);
float
abs_val
=
fabsf
(
val
);
local_absmax
=
fmaxf
(
local_absmax
,
abs_val
);
}
}
local_absmax
=
GroupReduce
(
local_absmax
,
lane_id
);
s_absmax
[
local_group_id
][
local_tid
]
=
local_absmax
;
__syncthreads
();
if
(
l
ane_id
==
0
)
{
s_absmax
[
local_group_id
]
=
local_
absmax
;
if
(
l
ocal_tid
<
8
)
{
GroupReduceMax
(
&
s_absmax
[
local_group_id
]
[
0
],
local_
tid
)
;
}
__syncthreads
();
const
float
group_absmax
=
s_absmax
[
local_group_id
];
const
float
group_absmax
=
s_absmax
[
local_group_id
]
[
0
]
;
const
float
y_s
=
group_absmax
/
fp8_max
;
if
(
l
ane_
id
==
0
)
{
if
(
l
ocal_t
id
==
0
)
{
*
scale_output
=
y_s
;
}
for
(
int32_t
i
=
lane_id
;
i
<
num_vec_elems
;
i
+=
16
)
{
vec_t
input_vec
;
input_vec
.
cast_load
(
group_input
+
i
*
vec_size
);
#pragma unroll
for
(
uint32_t
j
=
0
;
j
<
vec_size
;
++
j
)
{
float
val
=
static_cast
<
float
>
(
input_vec
[
j
]);
for
(
int
i
=
local_tid
;
i
<
group_size
;
i
+=
16
)
{
float
val
=
static_cast
<
float
>
(
group_input
[
i
]);
float
q_val
=
fminf
(
fmaxf
(
val
/
y_s
,
fp8_min
),
fp8_max
);
group_output
[
i
*
vec_size
+
j
]
=
FP8_TYPE
(
q_val
);
group_output
[
i
]
=
FP8_TYPE
(
q_val
);
}
}
}
...
...
Write
Preview
Markdown
is supported
0%
Try again
or
attach a new file
.
Attach a file
Cancel
You are about to add
0
people
to the discussion. Proceed with caution.
Finish editing this message first!
Cancel
Please
register
or
sign in
to comment