pytorch tensorpipe issues

pytorch / tensorpipe

A tensor-aware point-to-point communication primitive for machine learning

Other

249 stars 75 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

Redefinition of 'struct prctl_mm_map'

#462 sanmai-NL opened 1 year ago
1
tensorpipe fails to build on FreeBSD: error: use of undeclared identifier 'getBootIDInternal'

#461 yurivict opened 1 year ago
1
rv < 0: too many open files

#460 hamidralmasi opened 1 year ago
0
pytorch1.8.1/tensorpip build issue

#459 tom00sun closed 2 years ago
3
Whether tensorpipe works for CPU only cluster?

#458 liangan1 closed 2 years ago
0
Question: how to disable IB at runtime?

#457 jasperzhong opened 2 years ago
2
Select ibv device who has active port_state.

#456 SolenoidWGT opened 2 years ago
0
Error: "transport retry counter exceeded" when torch.distributed.rpc.init_rpc between different pods in k8s

#455 SolenoidWGT opened 2 years ago
5
CMake: Prefer CUDAToolkit over CUDA if available

#454 peterbell10 opened 2 years ago
0
Acking the receiver (to read the message) of a completed write operation

#453 LiSu opened 2 years ago
4
Does benchmark_pipe support ibv transport and cuda channel?

#452 baoleai opened 2 years ago
9
undefined reference to `pthread_create'

#451 baoleai opened 2 years ago
1
Set Position independent code in CMake

#450 powderluv opened 2 years ago
0
Need -fPIC for static builds

#449 powderluv opened 2 years ago
0
Any PyTorch RPC code(CPU and CUDA) crashes with Segmentation fault on 4+ GCP a2-megagpu-16g nodes

#448 pbelevich opened 2 years ago
0
Sending empty(numel==0) cuda tensor results in SIGABRT crash

#447 pbelevich opened 2 years ago
1
AWS EFA workaround

#446 kumpera opened 2 years ago
4
Update bundled version in Finduv.cmake

#445 malfet closed 2 years ago
1
Meta device

#444 pbelevich closed 2 years ago
0
Meta device

#443 pbelevich closed 2 years ago
0
PyTorch init_rpc fails with 'Cannot allocate memory' in tensorpipe

#442 cimes-isi opened 2 years ago
0
Resource temporarily unavailable when initializing RPC in multi-node training

#441 gongziyida opened 2 years ago
1
Migrate to new CircleCI GPU executors

#440 lw closed 2 years ago
1
Question about CPU + IB & GPU + IB

#439 eedalong closed 2 years ago
2
Can I compile a modified version of tensorpipe and plug into Pytorch?

#438 eedalong closed 2 years ago
1
Questions regard to IB

#437 eedalong closed 2 years ago
5
tensorpipe/transport/ibv/reactor.cc:132 "Unknown opcode: 136"

#436 eedalong closed 2 years ago
15
CMake: support find_package

#435 luncliff closed 2 years ago
1
Message may be missing if multiple pipes are writing to another same one

#434 Rhett-Ying closed 2 years ago
5
libnop dependency changes

#433 themarpe closed 2 years ago
2
A GREAT project, more people should be aware of it !

#432 eedalong opened 2 years ago
1
How to enable CudaGdrChannel registration in tensorpipeAgent when using pytorch's rpc

#431 eedalong closed 2 years ago
2
Is IB now only supporting CUDA_GDR now ?

#430 eedalong closed 2 years ago
3
Questions regard to source code

#429 eedalong closed 2 years ago
2
Fix test for interleaved zero-length tensors.

#428 beauby closed 2 years ago
2
Enable CPU buffers in CUDA GDR XDTT channel.

#427 beauby closed 2 years ago
2
Handle CPU buffers in CUDA GDR XDTT.

#426 beauby closed 2 years ago
2
Add tests for CUDA GDR XDTT channel.

#425 beauby closed 2 years ago
2
Add IbvNic::registerMemory overload for CPU buffers.

#424 beauby closed 2 years ago
2
Add CMake declaration for cuda_gdr_xdtt.

#423 beauby closed 2 years ago
2
Rename new channel to Cuda Xdtt.

#422 beauby closed 2 years ago
2
Duplicate CUDA GDR channel for XDTT.

#421 beauby closed 2 years ago
2
[Question]How to detect pipe(obtained from ctx->connect()) is writable?

#420 Rhett-Ying opened 2 years ago
6
Is there any plan to integrate DPDK?

#419 eedalong opened 2 years ago
4
Did not find test folder after install with ninja

#418 eedalong closed 2 years ago
3
how to install pytensorpipe

#417 eedalong closed 2 years ago
1
Avoid accessing possibly-undefined InfiniBand field in GDR

#416 lw closed 2 years ago
2
Add a helper to auto-detect a node's address like NCCL does

#415 lw closed 3 years ago
5
Fix send queue being undersized in IBV

#414 lw closed 3 years ago
2
init_rpc fails on the latest build due to tensorpipe

#413 swd543 opened 3 years ago
6