-
Notifications
You must be signed in to change notification settings - Fork 4.1k
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.
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
Adding the new feature of FPDT #6462
base: master
Are you sure you want to change the base?
Conversation
deepspeed/sequence/fpdt_layer.py
Outdated
return out, lse | ||
|
||
|
||
def single_all_to_all(input_, scatter_idx, gather_idx, group): |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Hi @YJHMITWEB is this single_all_to_all
the same as the one in
DeepSpeed/deepspeed/sequence/layer.py
Line 41 in 89c4d9f
def single_all_to_all(input, scatter_idx, gather_idx, batch_dim_idx, group, async_op=False, handle=None, type=None): |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
We have not tested the non-blocking All2all with our FPDT design, therefore, we use the original non-blocking version. If this is preferred, we can test this.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This is fixed. We will now use the single_all_to_all
from layer.py
deepspeed/sequence/fpdt_layer.py
Outdated
|
||
compute_stream.wait_stream(offload_stream) | ||
compute_stream.synchronize() | ||
with torch.cuda.stream(offload_stream): |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
torch.cuda.stream(stream)
should be replaced by get_accelerator().stream(stream)
, same of other occurance in this file.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Got it!
deepspeed/sequence/fpdt_layer.py
Outdated
grad_qkv_linear_bias = torch.zeros(qkv_linear_bias.shape, device=qkv_linear_weight.device, dtype=torch.float) | ||
|
||
grad_global_attn_output_chunk = single_all_to_all(grad_output[:, :chunk_size].contiguous(), scatter_idx, gather_idx, spg) | ||
torch.cuda.synchronize() |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Here is should be get_accelerator().synchronize()
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Got it, thanks.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This one is solved.
deepspeed/sequence/layer.py
Outdated
@@ -168,7 +196,8 @@ def __init__( | |||
self.gather_idx = gather_idx | |||
self.sp_overlap_comm = False | |||
self.overlap_handles = None | |||
self.sp_stream = sp_stream | |||
self.sp_stream = None | |||
# self.sp_stream = sp_stream |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Why comment out this line?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This one is solved.
fix format and add unit test for fpdt
[FPDT](https://arxiv.org/abs/2408.16978) can only be used with [this version](microsoft/Megatron-DeepSpeed#441) of Megatron-DeepSpeed. --------- Co-authored-by: Jinghan Yao <[email protected]> Co-authored-by: Sam Ade Jacobs <[email protected]> Co-authored-by: Jinghan Yao <[email protected]> Co-authored-by: Logan Adams <[email protected]> Co-authored-by: Olatunji Ruwase <[email protected]> Co-authored-by: Jinghan Yao <[email protected]> Co-authored-by: Logan Adams <[email protected]> Co-authored-by: Masahiro Tanaka <[email protected]> Co-authored-by: Masahiro Tanaka <[email protected]>
[FPDT](https://arxiv.org/abs/2408.16978) can only be used with [this version](microsoft/Megatron-DeepSpeed#441) of Megatron-DeepSpeed. --------- Co-authored-by: Jinghan Yao <[email protected]> Co-authored-by: Sam Ade Jacobs <[email protected]> Co-authored-by: Jinghan Yao <[email protected]> Co-authored-by: Logan Adams <[email protected]> Co-authored-by: Olatunji Ruwase <[email protected]> Co-authored-by: Jinghan Yao <[email protected]> Co-authored-by: Logan Adams <[email protected]> Co-authored-by: Masahiro Tanaka <[email protected]> Co-authored-by: Masahiro Tanaka <[email protected]>
FPDT can only be used with this version of Megatron-DeepSpeed.