Project

General

Profile

Feature #2890

Feature #2816: Device-side update&constraits, buffer ops and multi-gpu comms

GPU Halo Exchange

Added by Alan Gray 6 months ago. Updated about 2 months ago.

Status:
New
Priority:
Normal
Assignee:
-
Category:
-
Target version:
Difficulty:
uncategorized
Close

Description

When utilizing multiple GPUs with PP domain decomposition, halo data must be exchanged between PP tasks. Currently, this is routed through the host CPUs: D2H transfer; buffer packing on CPU; host-side MPI; and same in reverse. This is true for both the position and force buffers.

Instead, we can transfer data directly between GPU memory spaces using GPU peer-to-peer communication. Modern MPI implementations are CUDA-aware and support this, or D2D cudaMemcopies can be used directly when in threadMPI mode. The complication is that we need to pack buffers directly on the GPU. This is done by first preparing (on the CPU) a mapping of array indices involved in the exchange, which can then be used in GPU buffer packing/unpacking kernels. This mapping array only needs to populated and transferred during neighbor search steps.

Limitation: still only supports 1D data decomposition. Higher numbers of dimensions should be relatively straightforward to implement by extending the existing developments, but this still requires design and testing.

TODO: implement support for higher numbers of dimensions.
TODO: integrate call to force buffer halo exchange, when force buffer ops
patches accepted.


Subtasks

Feature #3089: relax dlb scaling limit when that would suit GPU halo exchangeNewBerk Hess
Task #3092: implement better receiver ready / notify in halo exchange New
Task #3093: rework GPU direct halo-ecxhange related force reduction complexities New

Related issues

Related to GROMACS - Feature #3052: GPU virial reduction/calculationNew
Related to GROMACS - Feature #3087: enable GPU peer to peer accessNew

Associated revisions

Revision 8f42be1e (diff)
Added by Alan Gray about 1 month ago

GPU halo exchange

Activate with GMX_GPU_DD_COMMS environment variable.

Class to initialize and apply halo exchange functionality directly on
GPU memory space.

Fully operational for position buffer. Functionality also present for
force buffer, but not yet called (awaiting acceptance of force buffer
ops patches).

Data transfer for halo exchange is wrapped and has 2 implementations:
cuda-aware MPI (default with "real" MPI) and direct cuda memcpy
(default with thread MPI). With the latter, the P2P path will be taken
if the hardware support it, otherwise D2H,H2D.

Limitation: still only supports 1D data decomposition

TODO: implement support for higher numbers of dimensions.
TODO: integrate call to force buffer halo exchange, when force buffer ops
patches accepted.

Implements part of #2890
Associated with #2915

Change-Id: I8e6473481ad4d943df78d7019681bfa821bd5798

Revision 44f607d7 (diff)
Added by Alan Gray 7 days ago

GPU halo exchange

Activate with GMX_GPU_DD_COMMS and GMX_USE_GPU_BUFFER_OPS environment
variable.

Class to initialize and apply coordinate buffer halo exchange
functionality directly on GPU memory space.

Currently only supports direct cuda memcpy, and relies on thread MPI
being in use.

Updated gpucomm testing matrices to cover non-GPU case.

Limitation: still only supports thread MPI, 1D data decomposition and
only coordinate halo exchange

Implements part of #2890
Associated with #2915

Change-Id: I8e6473481ad4d943df78d7019681bfa821bd5798

Revision 54c24729 (diff)
Added by Alan Gray 5 days ago

GPU Force Halo Exchange

Activate with GMX_GPU_DD_COMMS environment variable.

Extends GPU Halo exchange feature to provide GPU Force halo exchange
functionality. Does not yet support virial steps, which require an
extra shift force reduction - these are currently performed on the
non-buffer ops / non direct-comm path. Also has same limitations as
coordinate halo exchange.

Performs part of #2890. Future work to improve synchronization towards
a more one-sided scheme (#3092) and to make depenencies more
explicit (#3093)

Change-Id: Ifc23cc8db2655f7258e68b34e7cdc7b71994e1e8

History

#1 Updated by Alan Gray 6 months ago

  • Target version set to 2020

#2 Updated by Gerrit Code Review Bot 6 months ago

Gerrit received a related patchset '8' for Issue #2890.
Uploader: Alan Gray ()
Change-Id: gromacs~master~I8e6473481ad4d943df78d7019681bfa821bd5798
Gerrit URL: https://gerrit.gromacs.org/9225

#3 Updated by Alan Gray 6 months ago

  • Description updated (diff)

#4 Updated by Alan Gray 6 months ago

  • Description updated (diff)

#5 Updated by Alan Gray 6 months ago

Position halo exchange patch awaiting review.

#6 Updated by Alan Gray 4 months ago

  • Description updated (diff)

#7 Updated by Alan Gray 4 months ago

  • Description updated (diff)

#8 Updated by Szilárd Páll about 2 months ago

  • Related to Feature #3052: GPU virial reduction/calculation added

#9 Updated by Szilárd Páll about 2 months ago

Note that the f exchange needs to consider #3052. Short-term solution could be to have the direct communication disabled on virial steps, but a final solution should ideally avoid such complexities in control code.

#10 Updated by Szilárd Páll 11 days ago

Also available in: Atom PDF