Skip to content

Fix FP8 block scaling with sequence parallel #2637

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
cuichenx wants to merge 7 commits into
base: main
Choose a base branch
from
Open

Fix FP8 block scaling with sequence parallel #2637

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
7 commits
Select commit Hold shift + click to select a range
3ba991b
fix subchannel fp8 + sp
cuichenx Jan 31, 2026
390b2e1
Merge branch 'main' into chcui/fix_subchannel_fp8+sp
timmoon10 Feb 2, 2026
637ba0f
Support sequence-parallel all-gather with small inputs
timmoon10 Feb 2, 2026
9fe572b
Merge branch 'main' into chcui/fix_subchannel_fp8+sp
timmoon10 Feb 3, 2026
19fd927
Fix lint error
ptrendx Feb 10, 2026
adfe33b
Keep the previous behavior with dtype
ptrendx Feb 10, 2026
e6d1559
Merge branch 'main' into chcui/fix_subchannel_fp8+sp
ptrendx Feb 12, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
15 changes: 12 additions & 3 deletions transformer_engine/pytorch/distributed.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1100,6 +1100,9 @@ def _start_all_gather_fp8_blockwise(

# Fall back to high-precision all-gather if FP8 is not supported
if not quantizer.is_quantizable(inp) or quantizer.block_scaling_dim != 1:
warnings.warn("Cannot quantize input tensor. Performing all-gather in high precision.")
if isinstance(inp, QuantizedTensorStorage):
inp = inp.dequantize(dtype=dtype) # Dequantize if needed
out = torch.empty(out_shape, dtype=dtype, device=device)
torch.distributed.all_gather_into_tensor(out, inp, group=process_group, async_op=False)
Copy link
Contributor

@greptile-apps greptile-apps bot Feb 12, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Missing .contiguous() call on inp before all-gather

Other all-gather paths in this file use .contiguous() (lines 1739, 1033). Non-contiguous tensors (from transpose/slicing) can cause runtime errors.

Suggested change
torch.distributed.all_gather_into_tensor(out, inp, group=process_group, async_op=False)
torch.distributed.all_gather_into_tensor(out, inp.contiguous(), group=process_group, async_op=False)

out = quantizer(out)
Comment on lines 1102 to 1108
Copy link
Contributor

@greptile-apps greptile-apps bot Feb 10, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Non-contiguous gather input

In the new high-precision fallback (if not quantizer.is_quantizable(inp) ...), all_gather_into_tensor(out, inp, ...) passes inp directly. Elsewhere in this same module the plain-tensor path uses inp.contiguous() (distributed.py:1737-1742) and the FP8 path uses _data.contiguous() (distributed.py:1031-1035), which strongly suggests the collective expects contiguous inputs. If inp is a non-contiguous view (common after transpose/slicing), this fallback can raise at runtime. This same issue also appears in the NVFP4 and MXFP8 high-precision fallbacks (distributed.py:1353 and :1523).

Expand All @@ -1115,7 +1118,7 @@ def _start_all_gather_fp8_blockwise(
"Input and quantizer do not have matching usages. "
"Dequantizing and requantizing to Float8BlockwiseQTensor."
)
inp = quantizer(inp.dequantize())
inp = quantizer(inp.dequantize(dtype=dtype))

# Construct Float8BlockwiseQTensor output tensor
out = quantizer.make_empty(out_shape, dtype=dtype, device=device)
Expand Down Expand Up @@ -1338,6 +1341,9 @@ def _all_gather_nvfp4(
and quantizer is not None
and not quantizer.is_quantizable(inp)
):
warnings.warn("Cannot quantize input tensor. Performing all-gather in high precision.")
if isinstance(inp, QuantizedTensorStorage):
inp = inp.dequantize(dtype=dtype) # Dequantize if needed
out = torch.empty(
out_shape,
dtype=dtype,
Expand All @@ -1358,7 +1364,7 @@ def _all_gather_nvfp4(
"Input and quantizer do not have matching usages. "
"Dequantizing and requantizing to NVFP4."
)
inp = quantizer(inp.dequantize())
inp = quantizer(inp.dequantize(dtype=dtype))

# Construct NVFP4 output tensor
out = quantizer.make_empty(out_shape, dtype=dtype, device=device)
Expand Down Expand Up @@ -1505,6 +1511,9 @@ def _all_gather_mxfp8(
and quantizer is not None
and not quantizer.is_quantizable(inp)
):
warnings.warn("Cannot quantize input tensor. Performing all-gather in high precision.")
if isinstance(inp, QuantizedTensorStorage):
inp = inp.dequantize(dtype=dtype) # Dequantize if needed
out = torch.empty(
out_shape,
dtype=dtype,
Expand All @@ -1525,7 +1534,7 @@ def _all_gather_mxfp8(
"Input and quantizer do not have matching usages. "
"Dequantizing and requantizing to MXFP8."
)
inp = quantizer(inp.dequantize())
inp = quantizer(inp.dequantize(dtype=dtype))

# Construct MXFP8 output tensor
out = quantizer.make_empty(out_shape, dtype=dtype, device=device)
Expand Down
2 changes: 0 additions & 2 deletions transformer_engine/pytorch/module/layernorm_linear.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -29,7 +29,6 @@
from ..quantization import FP8GlobalStateManager
from ..utils import (
assert_dim_for_fp8_exec,
assert_dim_for_all_gather,
cast_if_needed,
clear_tensor_data,
divide,
Expand Down Expand Up @@ -158,7 +157,6 @@ def forward(
inputmat = inp
if fp8:
assert_dim_for_fp8_exec(inputmat, weight)
assert_dim_for_all_gather(inputmat, with_input_all_gather, input_quantizer)

# Cast for native AMP
nvtx_range_push(f"{nvtx_label}.norm_input_cast")
Expand Down
2 changes: 0 additions & 2 deletions transformer_engine/pytorch/module/layernorm_mlp.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -40,7 +40,6 @@
init_method_constant,
cast_if_needed,
assert_dim_for_fp8_exec,
assert_dim_for_all_gather,
clear_tensor_data,
requires_grad,
needs_quantized_gemm,
Expand Down Expand Up @@ -334,7 +333,6 @@ def _forward(
inputmat = inp.view((-1, in_features))
if fp8:
assert_dim_for_fp8_exec(inputmat, fc1_weight, fc2_weight)
assert_dim_for_all_gather(inputmat, sequence_parallel, fc1_input_quantizer)

activation_func = _act_func(
activation, FP8GlobalStateManager.get_fp8_recipe() if fp8 else None
Expand Down
2 changes: 0 additions & 2 deletions transformer_engine/pytorch/module/linear.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -34,7 +34,6 @@
requires_grad,
needs_quantized_gemm,
assert_dim_for_fp8_exec,
assert_dim_for_all_gather,
nvtx_range_pop,
nvtx_range_push,
get_nvtx_range_context,
Expand Down Expand Up @@ -175,7 +174,6 @@ def forward(
own_quantized_input = False
if fp8:
assert_dim_for_fp8_exec(inputmat, weight)
assert_dim_for_all_gather(inputmat, with_input_all_gather_nccl, input_quantizer)
if save_original_input:
assert not isinstance(
input_quantizer, Float8Quantizer
Expand Down
11 changes: 0 additions & 11 deletions transformer_engine/pytorch/utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -12,7 +12,6 @@
import numpy as np
import torch

from .quantized_tensor import Quantizer
from .torch_version import torch_version
from ..debug.pytorch.debug_quantization import DebugQuantizedTensor

Expand Down Expand Up @@ -447,16 +446,6 @@ def assert_dim_for_fp8_exec(*tensors: List[torch.Tensor]) -> None:
)


def assert_dim_for_all_gather(
tensor: torch.Tensor, with_all_gather: bool, quantizer: Quantizer
) -> None:
"""Assert that tensor dimensions are supported for all-gather"""
if with_all_gather:
assert quantizer.is_quantizable(tensor), (
"All-gather requires quantizable tensor for quantizer " + quantizer.__class__.__name__
)


def is_bf16_compatible() -> bool:
"""Replaces torch.cuda.is_bf16_compatible() with an explicit
check on device compute capability to enforce sm_80 or higher.
Expand Down
Loading