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
Split out reusable CUDAFuture from FutureNCCL #48506
Conversation
This commit is part of a stack that reworks FutureNCCL in order to extract a generic CUDA-aware Future subclass. The stack deliberately breaks up this transition into elementary changes, to make it easier to verify that the behavior is preserved (or to highlight how it gets changed). --- FutureNCCL is now a general-purpose type-agnostic multi-device class, so in this commit I extract it from ProcessGroupNCCL to make it available for wider use (notably by the RPC module). We'll call this new class CUDAFuture. We'll keep FutureNCCL as a subclass of CUDAFuture to deal with some NCCL peculiarity, namely the fact that the future becomes complete immediately upon creation. We can clean this up for good once we're done merging Future and Work. I'm not exactly sure of where to put CUDAFuture. It needs to be available to both c10d and RPC (which lives under torch/csrc). If I figured CMake out correctly (and that's a big if) I think c10d can only depend on ATen (I'll maybe add a comment with how I tracked that down). Hence we cannot put CUDAFuture in torch/csrc. On the other hand, RPC currently depends on c10d, because RPC agents use ProcessGroups internally, so it would be "ok" to put CUDAFuture in c10d. However, we want to get rid of ProcessGroups in RPC, and at that point RPC should in principle not depend on c10d. In that case, the only shared dep between the two that I see is ATen itself. While I'm a bit wary of putting it right in ATen, I think it might actually make sense. CUDAFuture is intended to be a general-purpose component that can be reused in all settings and is not particularly tied to c10d or RPC. Moreover, ATen already contains ivalue::Future, and it contains a lot of CUDA helpers, so CUDAFuture definitely belongs to the "closure" of what's already there. Differential Revision: [D25180532](https://our.internmc.facebook.com/intern/diff/D25180532/) [ghstack-poisoned]
This commit is part of a stack that reworks FutureNCCL in order to extract a generic CUDA-aware Future subclass. The stack deliberately breaks up this transition into elementary changes, to make it easier to verify that the behavior is preserved (or to highlight how it gets changed). --- FutureNCCL is now a general-purpose type-agnostic multi-device class, so in this commit I extract it from ProcessGroupNCCL to make it available for wider use (notably by the RPC module). We'll call this new class CUDAFuture. We'll keep FutureNCCL as a subclass of CUDAFuture to deal with some NCCL peculiarity, namely the fact that the future becomes complete immediately upon creation. We can clean this up for good once we're done merging Future and Work. I'm not exactly sure of where to put CUDAFuture. It needs to be available to both c10d and RPC (which lives under torch/csrc). If I figured CMake out correctly (and that's a big if) I think c10d can only depend on ATen (I'll maybe add a comment with how I tracked that down). Hence we cannot put CUDAFuture in torch/csrc. On the other hand, RPC currently depends on c10d, because RPC agents use ProcessGroups internally, so it would be "ok" to put CUDAFuture in c10d. However, we want to get rid of ProcessGroups in RPC, and at that point RPC should in principle not depend on c10d. In that case, the only shared dep between the two that I see is ATen itself. While I'm a bit wary of putting it right in ATen, I think it might actually make sense. CUDAFuture is intended to be a general-purpose component that can be reused in all settings and is not particularly tied to c10d or RPC. Moreover, ATen already contains ivalue::Future, and it contains a lot of CUDA helpers, so CUDAFuture definitely belongs to the "closure" of what's already there. Differential Revision: [D25180532](https://our.internmc.facebook.com/intern/diff/D25180532/) ghstack-source-id: 117391723 Pull Request resolved: #48506
Here we go:
|
💊 CI failures summary and remediationsAs of commit d432680 (more details on the Dr. CI page):
🕵️ 1 new failure recognized by patternsThe following CI failures do not appear to be due to upstream breakages: pytorch_linux_xenial_py3_6_gcc5_4_build (1/1)Step: "Build" (full log | diagnosis details | 🔁 rerun)
|
This commit is part of a stack that reworks FutureNCCL in order to extract a generic CUDA-aware Future subclass. The stack deliberately breaks up this transition into elementary changes, to make it easier to verify that the behavior is preserved (or to highlight how it gets changed). --- FutureNCCL is now a general-purpose type-agnostic multi-device class, so in this commit I extract it from ProcessGroupNCCL to make it available for wider use (notably by the RPC module). We'll call this new class CUDAFuture. We'll keep FutureNCCL as a subclass of CUDAFuture to deal with some NCCL peculiarity, namely the fact that the future becomes complete immediately upon creation. We can clean this up for good once we're done merging Future and Work. I'm not exactly sure of where to put CUDAFuture. It needs to be available to both c10d and RPC (which lives under torch/csrc). If I figured CMake out correctly (and that's a big if) I think c10d can only depend on ATen (I'll maybe add a comment with how I tracked that down). Hence we cannot put CUDAFuture in torch/csrc. On the other hand, RPC currently depends on c10d, because RPC agents use ProcessGroups internally, so it would be "ok" to put CUDAFuture in c10d. However, we want to get rid of ProcessGroups in RPC, and at that point RPC should in principle not depend on c10d. In that case, the only shared dep between the two that I see is ATen itself. While I'm a bit wary of putting it right in ATen, I think it might actually make sense. CUDAFuture is intended to be a general-purpose component that can be reused in all settings and is not particularly tied to c10d or RPC. Moreover, ATen already contains ivalue::Future, and it contains a lot of CUDA helpers, so CUDAFuture definitely belongs to the "closure" of what's already there. Differential Revision: [D25180532](https://our.internmc.facebook.com/intern/diff/D25180532/) [ghstack-poisoned]
Pull Request resolved: #48506 This commit is part of a stack that reworks FutureNCCL in order to extract a generic CUDA-aware Future subclass. The stack deliberately breaks up this transition into elementary changes, to make it easier to verify that the behavior is preserved (or to highlight how it gets changed). --- FutureNCCL is now a general-purpose type-agnostic multi-device class, so in this commit I extract it from ProcessGroupNCCL to make it available for wider use (notably by the RPC module). We'll call this new class CUDAFuture. We'll keep FutureNCCL as a subclass of CUDAFuture to deal with some NCCL peculiarity, namely the fact that the future becomes complete immediately upon creation. We can clean this up for good once we're done merging Future and Work. I'm not exactly sure of where to put CUDAFuture. It needs to be available to both c10d and RPC (which lives under torch/csrc). If I figured CMake out correctly (and that's a big if) I think c10d can only depend on ATen (I'll maybe add a comment with how I tracked that down). Hence we cannot put CUDAFuture in torch/csrc. On the other hand, RPC currently depends on c10d, because RPC agents use ProcessGroups internally, so it would be "ok" to put CUDAFuture in c10d. However, we want to get rid of ProcessGroups in RPC, and at that point RPC should in principle not depend on c10d. In that case, the only shared dep between the two that I see is ATen itself. While I'm a bit wary of putting it right in ATen, I think it might actually make sense. CUDAFuture is intended to be a general-purpose component that can be reused in all settings and is not particularly tied to c10d or RPC. Moreover, ATen already contains ivalue::Future, and it contains a lot of CUDA helpers, so CUDAFuture definitely belongs to the "closure" of what's already there. ghstack-source-id: 117410075 Differential Revision: [D25180532](https://our.internmc.facebook.com/intern/diff/D25180532/)
} | ||
|
||
void postWaitHook() override { | ||
for (at::cuda::CUDAEvent& cudaEvent : *cudaEvents_) { |
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.
hmm, this means users cannot call wait on CUDAFuture
before it is marked as completed?
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.
Not sure I follow: the postWaitHook
is called by ivalue::Future once a wait has finished, and that only happens when the future is marked complete. Thus one can certainly start waiting before the future is complete, in which case one will block inside ivalue::Future (waiting on a condition variable). Since that only unblocks after the future is complete, we're guaranteed that postWaitHook
will be called after postMarkCompletedHook
. Is this what you were wondering about?
: at::ivalue::Future(c10::ListType::create(c10::TensorType::get())), | ||
cudaEvents_(std::move(cudaEvents)) { | ||
: at::cuda::CUDAFuture(c10::ListType::create(c10::TensorType::get())){ | ||
cudaEvents_ = std::move(cudaEvents); |
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.
Are these events used anywhere? Will the markCompleted
call below this line immediately override these cudaEvents_
?
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.
These events will not be overridden when the future is marked complete, because FutureNCCL provides its own postMarkCompletedHook
which does nothing (and which overrides the one from CUDAFuture, which is where we store the events).
However, these events will still be used by the postWaitHook, by the wrapCallback, etc.
It's somewhat hacky, but it works, and it's a temporary solution only until we finish merging Future and Work (at which point ProcessGroupNCCL should be able to use CUDAFuture unmodified).
This commit is part of a stack that reworks FutureNCCL in order to extract a generic CUDA-aware Future subclass. The stack deliberately breaks up this transition into elementary changes, to make it easier to verify that the behavior is preserved (or to highlight how it gets changed). --- FutureNCCL is now a general-purpose type-agnostic multi-device class, so in this commit I extract it from ProcessGroupNCCL to make it available for wider use (notably by the RPC module). We'll call this new class CUDAFuture. We'll keep FutureNCCL as a subclass of CUDAFuture to deal with some NCCL peculiarity, namely the fact that the future becomes complete immediately upon creation. We can clean this up for good once we're done merging Future and Work. I'm not exactly sure of where to put CUDAFuture. It needs to be available to both c10d and RPC (which lives under torch/csrc). If I figured CMake out correctly (and that's a big if) I think c10d can only depend on ATen (I'll maybe add a comment with how I tracked that down). Hence we cannot put CUDAFuture in torch/csrc. On the other hand, RPC currently depends on c10d, because RPC agents use ProcessGroups internally, so it would be "ok" to put CUDAFuture in c10d. However, we want to get rid of ProcessGroups in RPC, and at that point RPC should in principle not depend on c10d. In that case, the only shared dep between the two that I see is ATen itself. While I'm a bit wary of putting it right in ATen, I think it might actually make sense. CUDAFuture is intended to be a general-purpose component that can be reused in all settings and is not particularly tied to c10d or RPC. Moreover, ATen already contains ivalue::Future, and it contains a lot of CUDA helpers, so CUDAFuture definitely belongs to the "closure" of what's already there. Differential Revision: [D25180532](https://our.internmc.facebook.com/intern/diff/D25180532/) [ghstack-poisoned]
Pull Request resolved: #48506 This commit is part of a stack that reworks FutureNCCL in order to extract a generic CUDA-aware Future subclass. The stack deliberately breaks up this transition into elementary changes, to make it easier to verify that the behavior is preserved (or to highlight how it gets changed). --- FutureNCCL is now a general-purpose type-agnostic multi-device class, so in this commit I extract it from ProcessGroupNCCL to make it available for wider use (notably by the RPC module). We'll call this new class CUDAFuture. We'll keep FutureNCCL as a subclass of CUDAFuture to deal with some NCCL peculiarity, namely the fact that the future becomes complete immediately upon creation. We can clean this up for good once we're done merging Future and Work. I'm not exactly sure of where to put CUDAFuture. It needs to be available to both c10d and RPC (which lives under torch/csrc). If I figured CMake out correctly (and that's a big if) I think c10d can only depend on ATen (I'll maybe add a comment with how I tracked that down). Hence we cannot put CUDAFuture in torch/csrc. On the other hand, RPC currently depends on c10d, because RPC agents use ProcessGroups internally, so it would be "ok" to put CUDAFuture in c10d. However, we want to get rid of ProcessGroups in RPC, and at that point RPC should in principle not depend on c10d. In that case, the only shared dep between the two that I see is ATen itself. While I'm a bit wary of putting it right in ATen, I think it might actually make sense. CUDAFuture is intended to be a general-purpose component that can be reused in all settings and is not particularly tied to c10d or RPC. Moreover, ATen already contains ivalue::Future, and it contains a lot of CUDA helpers, so CUDAFuture definitely belongs to the "closure" of what's already there. ghstack-source-id: 117420192 Differential Revision: [D25180532](https://our.internmc.facebook.com/intern/diff/D25180532/)
This commit is part of a stack that reworks FutureNCCL in order to extract a generic CUDA-aware Future subclass. The stack deliberately breaks up this transition into elementary changes, to make it easier to verify that the behavior is preserved (or to highlight how it gets changed). --- FutureNCCL is now a general-purpose type-agnostic multi-device class, so in this commit I extract it from ProcessGroupNCCL to make it available for wider use (notably by the RPC module). We'll call this new class CUDAFuture. We'll keep FutureNCCL as a subclass of CUDAFuture to deal with some NCCL peculiarity, namely the fact that the future becomes complete immediately upon creation. We can clean this up for good once we're done merging Future and Work. I'm not exactly sure of where to put CUDAFuture. It needs to be available to both c10d and RPC (which lives under torch/csrc). If I figured CMake out correctly (and that's a big if) I think c10d can only depend on ATen (I'll maybe add a comment with how I tracked that down). Hence we cannot put CUDAFuture in torch/csrc. On the other hand, RPC currently depends on c10d, because RPC agents use ProcessGroups internally, so it would be "ok" to put CUDAFuture in c10d. However, we want to get rid of ProcessGroups in RPC, and at that point RPC should in principle not depend on c10d. In that case, the only shared dep between the two that I see is ATen itself. While I'm a bit wary of putting it right in ATen, I think it might actually make sense. CUDAFuture is intended to be a general-purpose component that can be reused in all settings and is not particularly tied to c10d or RPC. Moreover, ATen already contains ivalue::Future, and it contains a lot of CUDA helpers, so CUDAFuture definitely belongs to the "closure" of what's already there. Differential Revision: [D25180532](https://our.internmc.facebook.com/intern/diff/D25180532/) [ghstack-poisoned]
Pull Request resolved: #48506 This commit is part of a stack that reworks FutureNCCL in order to extract a generic CUDA-aware Future subclass. The stack deliberately breaks up this transition into elementary changes, to make it easier to verify that the behavior is preserved (or to highlight how it gets changed). --- FutureNCCL is now a general-purpose type-agnostic multi-device class, so in this commit I extract it from ProcessGroupNCCL to make it available for wider use (notably by the RPC module). We'll call this new class CUDAFuture. We'll keep FutureNCCL as a subclass of CUDAFuture to deal with some NCCL peculiarity, namely the fact that the future becomes complete immediately upon creation. We can clean this up for good once we're done merging Future and Work. I'm not exactly sure of where to put CUDAFuture. It needs to be available to both c10d and RPC (which lives under torch/csrc). If I figured CMake out correctly (and that's a big if) I think c10d can only depend on ATen (I'll maybe add a comment with how I tracked that down). Hence we cannot put CUDAFuture in torch/csrc. On the other hand, RPC currently depends on c10d, because RPC agents use ProcessGroups internally, so it would be "ok" to put CUDAFuture in c10d. However, we want to get rid of ProcessGroups in RPC, and at that point RPC should in principle not depend on c10d. In that case, the only shared dep between the two that I see is ATen itself. While I'm a bit wary of putting it right in ATen, I think it might actually make sense. CUDAFuture is intended to be a general-purpose component that can be reused in all settings and is not particularly tied to c10d or RPC. Moreover, ATen already contains ivalue::Future, and it contains a lot of CUDA helpers, so CUDAFuture definitely belongs to the "closure" of what's already there. ghstack-source-id: 117437508 Differential Revision: [D25180532](https://our.internmc.facebook.com/intern/diff/D25180532/)
This commit is part of a stack that reworks FutureNCCL in order to extract a generic CUDA-aware Future subclass. The stack deliberately breaks up this transition into elementary changes, to make it easier to verify that the behavior is preserved (or to highlight how it gets changed). --- FutureNCCL is now a general-purpose type-agnostic multi-device class, so in this commit I extract it from ProcessGroupNCCL to make it available for wider use (notably by the RPC module). We'll call this new class CUDAFuture. We'll keep FutureNCCL as a subclass of CUDAFuture to deal with some NCCL peculiarity, namely the fact that the future becomes complete immediately upon creation. We can clean this up for good once we're done merging Future and Work. I'm not exactly sure of where to put CUDAFuture. It needs to be available to both c10d and RPC (which lives under torch/csrc). If I figured CMake out correctly (and that's a big if) I think c10d can only depend on ATen (I'll maybe add a comment with how I tracked that down). Hence we cannot put CUDAFuture in torch/csrc. On the other hand, RPC currently depends on c10d, because RPC agents use ProcessGroups internally, so it would be "ok" to put CUDAFuture in c10d. However, we want to get rid of ProcessGroups in RPC, and at that point RPC should in principle not depend on c10d. In that case, the only shared dep between the two that I see is ATen itself. While I'm a bit wary of putting it right in ATen, I think it might actually make sense. CUDAFuture is intended to be a general-purpose component that can be reused in all settings and is not particularly tied to c10d or RPC. Moreover, ATen already contains ivalue::Future, and it contains a lot of CUDA helpers, so CUDAFuture definitely belongs to the "closure" of what's already there. Differential Revision: [D25180532](https://our.internmc.facebook.com/intern/diff/D25180532/) [ghstack-poisoned]
Pull Request resolved: #48506 This commit is part of a stack that reworks FutureNCCL in order to extract a generic CUDA-aware Future subclass. The stack deliberately breaks up this transition into elementary changes, to make it easier to verify that the behavior is preserved (or to highlight how it gets changed). --- FutureNCCL is now a general-purpose type-agnostic multi-device class, so in this commit I extract it from ProcessGroupNCCL to make it available for wider use (notably by the RPC module). We'll call this new class CUDAFuture. We'll keep FutureNCCL as a subclass of CUDAFuture to deal with some NCCL peculiarity, namely the fact that the future becomes complete immediately upon creation. We can clean this up for good once we're done merging Future and Work. I'm not exactly sure of where to put CUDAFuture. It needs to be available to both c10d and RPC (which lives under torch/csrc). If I figured CMake out correctly (and that's a big if) I think c10d can only depend on ATen (I'll maybe add a comment with how I tracked that down). Hence we cannot put CUDAFuture in torch/csrc. On the other hand, RPC currently depends on c10d, because RPC agents use ProcessGroups internally, so it would be "ok" to put CUDAFuture in c10d. However, we want to get rid of ProcessGroups in RPC, and at that point RPC should in principle not depend on c10d. In that case, the only shared dep between the two that I see is ATen itself. While I'm a bit wary of putting it right in ATen, I think it might actually make sense. CUDAFuture is intended to be a general-purpose component that can be reused in all settings and is not particularly tied to c10d or RPC. Moreover, ATen already contains ivalue::Future, and it contains a lot of CUDA helpers, so CUDAFuture definitely belongs to the "closure" of what's already there. ghstack-source-id: 117439444 Differential Revision: [D25180532](https://our.internmc.facebook.com/intern/diff/D25180532/)
This commit is part of a stack that reworks FutureNCCL in order to extract a generic CUDA-aware Future subclass. The stack deliberately breaks up this transition into elementary changes, to make it easier to verify that the behavior is preserved (or to highlight how it gets changed). --- FutureNCCL is now a general-purpose type-agnostic multi-device class, so in this commit I extract it from ProcessGroupNCCL to make it available for wider use (notably by the RPC module). We'll call this new class CUDAFuture. We'll keep FutureNCCL as a subclass of CUDAFuture to deal with some NCCL peculiarity, namely the fact that the future becomes complete immediately upon creation. We can clean this up for good once we're done merging Future and Work. I'm not exactly sure of where to put CUDAFuture. It needs to be available to both c10d and RPC (which lives under torch/csrc). If I figured CMake out correctly (and that's a big if) I think c10d can only depend on ATen (I'll maybe add a comment with how I tracked that down). Hence we cannot put CUDAFuture in torch/csrc. On the other hand, RPC currently depends on c10d, because RPC agents use ProcessGroups internally, so it would be "ok" to put CUDAFuture in c10d. However, we want to get rid of ProcessGroups in RPC, and at that point RPC should in principle not depend on c10d. In that case, the only shared dep between the two that I see is ATen itself. While I'm a bit wary of putting it right in ATen, I think it might actually make sense. CUDAFuture is intended to be a general-purpose component that can be reused in all settings and is not particularly tied to c10d or RPC. Moreover, ATen already contains ivalue::Future, and it contains a lot of CUDA helpers, so CUDAFuture definitely belongs to the "closure" of what's already there. Differential Revision: [D25180532](https://our.internmc.facebook.com/intern/diff/D25180532/) [ghstack-poisoned]
Pull Request resolved: #48506 This commit is part of a stack that reworks FutureNCCL in order to extract a generic CUDA-aware Future subclass. The stack deliberately breaks up this transition into elementary changes, to make it easier to verify that the behavior is preserved (or to highlight how it gets changed). --- FutureNCCL is now a general-purpose type-agnostic multi-device class, so in this commit I extract it from ProcessGroupNCCL to make it available for wider use (notably by the RPC module). We'll call this new class CUDAFuture. We'll keep FutureNCCL as a subclass of CUDAFuture to deal with some NCCL peculiarity, namely the fact that the future becomes complete immediately upon creation. We can clean this up for good once we're done merging Future and Work. I'm not exactly sure of where to put CUDAFuture. It needs to be available to both c10d and RPC (which lives under torch/csrc). If I figured CMake out correctly (and that's a big if) I think c10d can only depend on ATen (I'll maybe add a comment with how I tracked that down). Hence we cannot put CUDAFuture in torch/csrc. On the other hand, RPC currently depends on c10d, because RPC agents use ProcessGroups internally, so it would be "ok" to put CUDAFuture in c10d. However, we want to get rid of ProcessGroups in RPC, and at that point RPC should in principle not depend on c10d. In that case, the only shared dep between the two that I see is ATen itself. While I'm a bit wary of putting it right in ATen, I think it might actually make sense. CUDAFuture is intended to be a general-purpose component that can be reused in all settings and is not particularly tied to c10d or RPC. Moreover, ATen already contains ivalue::Future, and it contains a lot of CUDA helpers, so CUDAFuture definitely belongs to the "closure" of what's already there. ghstack-source-id: 117453067 Differential Revision: [D25180532](https://our.internmc.facebook.com/intern/diff/D25180532/)
This commit is part of a stack that reworks FutureNCCL in order to extract a generic CUDA-aware Future subclass. The stack deliberately breaks up this transition into elementary changes, to make it easier to verify that the behavior is preserved (or to highlight how it gets changed). --- FutureNCCL is now a general-purpose type-agnostic multi-device class, so in this commit I extract it from ProcessGroupNCCL to make it available for wider use (notably by the RPC module). We'll call this new class CUDAFuture. We'll keep FutureNCCL as a subclass of CUDAFuture to deal with some NCCL peculiarity, namely the fact that the future becomes complete immediately upon creation. We can clean this up for good once we're done merging Future and Work. I'm not exactly sure of where to put CUDAFuture. It needs to be available to both c10d and RPC (which lives under torch/csrc). If I figured CMake out correctly (and that's a big if) I think c10d can only depend on ATen (I'll maybe add a comment with how I tracked that down). Hence we cannot put CUDAFuture in torch/csrc. On the other hand, RPC currently depends on c10d, because RPC agents use ProcessGroups internally, so it would be "ok" to put CUDAFuture in c10d. However, we want to get rid of ProcessGroups in RPC, and at that point RPC should in principle not depend on c10d. In that case, the only shared dep between the two that I see is ATen itself. While I'm a bit wary of putting it right in ATen, I think it might actually make sense. CUDAFuture is intended to be a general-purpose component that can be reused in all settings and is not particularly tied to c10d or RPC. Moreover, ATen already contains ivalue::Future, and it contains a lot of CUDA helpers, so CUDAFuture definitely belongs to the "closure" of what's already there. Differential Revision: [D25180532](https://our.internmc.facebook.com/intern/diff/D25180532/) [ghstack-poisoned]
Pull Request resolved: #48506 This commit is part of a stack that reworks FutureNCCL in order to extract a generic CUDA-aware Future subclass. The stack deliberately breaks up this transition into elementary changes, to make it easier to verify that the behavior is preserved (or to highlight how it gets changed). --- FutureNCCL is now a general-purpose type-agnostic multi-device class, so in this commit I extract it from ProcessGroupNCCL to make it available for wider use (notably by the RPC module). We'll call this new class CUDAFuture. We'll keep FutureNCCL as a subclass of CUDAFuture to deal with some NCCL peculiarity, namely the fact that the future becomes complete immediately upon creation. We can clean this up for good once we're done merging Future and Work. I'm not exactly sure of where to put CUDAFuture. It needs to be available to both c10d and RPC (which lives under torch/csrc). If I figured CMake out correctly (and that's a big if) I think c10d can only depend on ATen (I'll maybe add a comment with how I tracked that down). Hence we cannot put CUDAFuture in torch/csrc. On the other hand, RPC currently depends on c10d, because RPC agents use ProcessGroups internally, so it would be "ok" to put CUDAFuture in c10d. However, we want to get rid of ProcessGroups in RPC, and at that point RPC should in principle not depend on c10d. In that case, the only shared dep between the two that I see is ATen itself. While I'm a bit wary of putting it right in ATen, I think it might actually make sense. CUDAFuture is intended to be a general-purpose component that can be reused in all settings and is not particularly tied to c10d or RPC. Moreover, ATen already contains ivalue::Future, and it contains a lot of CUDA helpers, so CUDAFuture definitely belongs to the "closure" of what's already there. ghstack-source-id: 117750948 Differential Revision: [D25180532](https://our.internmc.facebook.com/intern/diff/D25180532/)
This commit is part of a stack that reworks FutureNCCL in order to extract a generic CUDA-aware Future subclass. The stack deliberately breaks up this transition into elementary changes, to make it easier to verify that the behavior is preserved (or to highlight how it gets changed). --- FutureNCCL is now a general-purpose type-agnostic multi-device class, so in this commit I extract it from ProcessGroupNCCL to make it available for wider use (notably by the RPC module). We'll call this new class CUDAFuture. We'll keep FutureNCCL as a subclass of CUDAFuture to deal with some NCCL peculiarity, namely the fact that the future becomes complete immediately upon creation. We can clean this up for good once we're done merging Future and Work. I'm not exactly sure of where to put CUDAFuture. It needs to be available to both c10d and RPC (which lives under torch/csrc). If I figured CMake out correctly (and that's a big if) I think c10d can only depend on ATen (I'll maybe add a comment with how I tracked that down). Hence we cannot put CUDAFuture in torch/csrc. On the other hand, RPC currently depends on c10d, because RPC agents use ProcessGroups internally, so it would be "ok" to put CUDAFuture in c10d. However, we want to get rid of ProcessGroups in RPC, and at that point RPC should in principle not depend on c10d. In that case, the only shared dep between the two that I see is ATen itself. While I'm a bit wary of putting it right in ATen, I think it might actually make sense. CUDAFuture is intended to be a general-purpose component that can be reused in all settings and is not particularly tied to c10d or RPC. Moreover, ATen already contains ivalue::Future, and it contains a lot of CUDA helpers, so CUDAFuture definitely belongs to the "closure" of what's already there. Differential Revision: [D25180532](https://our.internmc.facebook.com/intern/diff/D25180532/) [ghstack-poisoned]
This commit is part of a stack that reworks FutureNCCL in order to extract a generic CUDA-aware Future subclass. The stack deliberately breaks up this transition into elementary changes, to make it easier to verify that the behavior is preserved (or to highlight how it gets changed). --- FutureNCCL is now a general-purpose type-agnostic multi-device class, so in this commit I extract it from ProcessGroupNCCL to make it available for wider use (notably by the RPC module). We'll call this new class CUDAFuture. We'll keep FutureNCCL as a subclass of CUDAFuture to deal with some NCCL peculiarity, namely the fact that the future becomes complete immediately upon creation. We can clean this up for good once we're done merging Future and Work. I'm not exactly sure of where to put CUDAFuture. It needs to be available to both c10d and RPC (which lives under torch/csrc). If I figured CMake out correctly (and that's a big if) I think c10d can only depend on ATen (I'll maybe add a comment with how I tracked that down). Hence we cannot put CUDAFuture in torch/csrc. On the other hand, RPC currently depends on c10d, because RPC agents use ProcessGroups internally, so it would be "ok" to put CUDAFuture in c10d. However, we want to get rid of ProcessGroups in RPC, and at that point RPC should in principle not depend on c10d. In that case, the only shared dep between the two that I see is ATen itself. While I'm a bit wary of putting it right in ATen, I think it might actually make sense. CUDAFuture is intended to be a general-purpose component that can be reused in all settings and is not particularly tied to c10d or RPC. Moreover, ATen already contains ivalue::Future, and it contains a lot of CUDA helpers, so CUDAFuture definitely belongs to the "closure" of what's already there. Differential Revision: [D25180532](https://our.internmc.facebook.com/intern/diff/D25180532/) [ghstack-poisoned]
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.
looks good, left some minor questions.
} | ||
|
||
protected: | ||
c10::intrusive_ptr<Future> createInstance(at::TypePtr type) override { |
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.
if this is the only way we want to create an instance, can we delete the default constructor? also, where's the PR that have the createInstance
for ivalue::Future?
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 the only way I found to have at::ivalue::Future::then()
be able to create an instance of at::cuda::CUDAFuture
. In short, the a method of the superclass needs to create an instance of the subclass. It's not supposed to be used by external users, hence it's protected
.
The superclass implementation of createInstance
is here: https://github.com/pytorch/pytorch/pull/48505/files#diff-2f833078d12338d0ac920ab654b5c791cb1219729d6e7c97a08393b32a46d173R491
|
||
std::function<void(void)> wrapCallback( | ||
std::function<void(void)> callback) override { | ||
return [this, callback{std::move(callback)}]() { |
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.
what's the difference of this between capture by reference? are the callbacks gonna destructed sometime but we still want them alive in the lambda?
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.
std::function
AFAIK is a RAII wrapper so it might hold some state which gets destroyed once the function goes out of scope. If we captured a reference to the std::function
that would thus risk being a dangling reference.
This commit is part of a stack that reworks FutureNCCL in order to extract a generic CUDA-aware Future subclass. The stack deliberately breaks up this transition into elementary changes, to make it easier to verify that the behavior is preserved (or to highlight how it gets changed). --- FutureNCCL is now a general-purpose type-agnostic multi-device class, so in this commit I extract it from ProcessGroupNCCL to make it available for wider use (notably by the RPC module). We'll call this new class CUDAFuture. We'll keep FutureNCCL as a subclass of CUDAFuture to deal with some NCCL peculiarity, namely the fact that the future becomes complete immediately upon creation. We can clean this up for good once we're done merging Future and Work. I'm not exactly sure of where to put CUDAFuture. It needs to be available to both c10d and RPC (which lives under torch/csrc). If I figured CMake out correctly (and that's a big if) I think c10d can only depend on ATen (I'll maybe add a comment with how I tracked that down). Hence we cannot put CUDAFuture in torch/csrc. On the other hand, RPC currently depends on c10d, because RPC agents use ProcessGroups internally, so it would be "ok" to put CUDAFuture in c10d. However, we want to get rid of ProcessGroups in RPC, and at that point RPC should in principle not depend on c10d. In that case, the only shared dep between the two that I see is ATen itself. While I'm a bit wary of putting it right in ATen, I think it might actually make sense. CUDAFuture is intended to be a general-purpose component that can be reused in all settings and is not particularly tied to c10d or RPC. Moreover, ATen already contains ivalue::Future, and it contains a lot of CUDA helpers, so CUDAFuture definitely belongs to the "closure" of what's already there. Differential Revision: [D25180532](https://our.internmc.facebook.com/intern/diff/D25180532/) [ghstack-poisoned]
This pull request has been merged in 030fa6c. |
Stack from ghstack:
This commit is part of a stack that reworks FutureNCCL in order to extract a generic CUDA-aware Future subclass. The stack deliberately breaks up this transition into elementary changes, to make it easier to verify that the behavior is preserved (or to highlight how it gets changed).
FutureNCCL is now a general-purpose type-agnostic multi-device class, so in this commit I extract it from ProcessGroupNCCL to make it available for wider use (notably by the RPC module). We'll call this new class CUDAFuture. We'll keep FutureNCCL as a subclass of CUDAFuture to deal with some NCCL peculiarity, namely the fact that the future becomes complete immediately upon creation. We can clean this up for good once we're done merging Future and Work.
I'm not exactly sure of where to put CUDAFuture. It needs to be available to both c10d and RPC (which lives under torch/csrc). If I figured CMake out correctly (and that's a big if) I think c10d can only depend on ATen (I'll maybe add a comment with how I tracked that down). Hence we cannot put CUDAFuture in torch/csrc. On the other hand, RPC currently depends on c10d, because RPC agents use ProcessGroups internally, so it would be "ok" to put CUDAFuture in c10d. However, we want to get rid of ProcessGroups in RPC, and at that point RPC should in principle not depend on c10d. In that case, the only shared dep between the two that I see is ATen itself.
While I'm a bit wary of putting it right in ATen, I think it might actually make sense. CUDAFuture is intended to be a general-purpose component that can be reused in all settings and is not particularly tied to c10d or RPC. Moreover, ATen already contains ivalue::Future, and it contains a lot of CUDA helpers, so CUDAFuture definitely belongs to the "closure" of what's already there.
Differential Revision: D25180532