dlslime 0.0.2.post1

DLSlime Transfer Engine

Roadmap | Slack | WeChat Group | Zhihu

Flexible & Efficient Heterogeneous Transfer Toolkit

Getting Started

DLSlime offers a set of peer-to-peer communication interfaces. For instance, consider the task of batched slice assignment from a remote tensor to a local tensor. You can accomplish this using the following APIs.

.

Here are some examples of DLSlime interface.

P2P Communication

RDMA RC Mode

• RDMA RC Read (Sync / Async mode)

python example/python/p2p_rdma_rc_read.py

• RDMA RC Read (Coroutine mode)

python example/python/p2p_rdma_rc_read_coroutine.py

• RDMA RC Write (Sync / Async mode)

python example/python/p2p_rdma_rc_write.py

• RDMA RC Write with immediate data (Sync / Async mode)

python example/python/p2p_rdma_rc_write_with_imm_data.py

• RDMA RC Send/Recv

python example/python/p2p_rdma_rc_send_recv.py

python example/python/p2p_rdma_rc_send_recv_gdr.py

• DLSlime torch backend

python example/python/p2p_rdma_rc_send_recv_torch.py --rank 0
python example/python/p2p_rdma_rc_send_recv_torch.py --rank 1

NVLink Mode

torchrun --nproc_per_node=2 p2p_nvlink.py

NVShmem Mode

# send
python example/python/p2p_nvshmem_ibgda_sendrecv.py --rank 0 --world-size 2

# recv
python example/python/p2p_nvshmem_ibgda_sendrecv.py --rank 1 --world-size 2

Huawei Ascend Direct Mode

See: Huawei README

[!Caution] DLSlime NVShmem transfer engine and Huawei Ascond Direct mode are in the experimental stage.

Collective Ops

Intra Node

AllGather

torchrun --nnodes 1 --master-addr 10.130.8.143 --node-rank 0 --nproc-per-node 8 --master-port 6007 example/python/all_gather_ll.py --mode intra

Inter Node

AllGather

# Node 0
torchrun --nnodes 2 --master-addr 10.130.8.143 --node-rank 0 --nproc-per-node 8 --master-port 6007 example/python/all_gather_ll.py --mode inter
# Node 1
torchrun --nnodes 2 --master-addr 10.130.8.143 --node-rank 1 --nproc-per-node 8 --master-port 6007 example/python/all_gather_ll.py --mode inter

AllGather Gemm Overlapping

# Node 0
torchrun --nnodes 2 --master-addr 10.130.8.143 --node-rank 0 --nproc-per-node 8 --master-port 6007 example/python/all_gather_gemm_overlap.py
# Node 1
torchrun --nnodes 2 --master-addr 10.130.8.143 --node-rank 1 --nproc-per-node 8 --master-port 6007 example/python/all_gather_gemm_overlap.py

[!Note] The intra- and inter- examples example above enables CUDA Graph by default. --eager-mode falls back to eager mode.

Install

pip install

pip install dlslime==0.0.1.post10

[!Note] The DLSlime pip version is built with default FLAGS (see Build from source for details).

Build from source

Python

git clone https://github.com/deeplink-org/DLSlime.git
FLAG=<ON|OFF> pip install -v --no-build-isolation -e .

CPP

git clone https://github.com/deeplink-org/DLSlime.git
mkdir -p DLSlime/build && cmake -DFLAG=<ON|OFF> ..

Build flags

The FLAG can be

Flag	Description	Platform	default
BUILD_RDMA	Build RDMA Transfer Engine	Hetero	ON
BUILD_PYTHON	Build Python wrapper	Hetero	ON
BUILD_NVLINK	Build NVLINK Transfer Engine	GPGPU	OFF
BUILD_NVSHMEM	Build NVShmem Transfer Engine	NVIDIA	OFF
BUILD_ASCEND_DIRECT	Build Ascend direct transport	ASCEND	OFF
BUILD_TORCH_PLUGIN	Build DLSlime as a torch backend	Hetero	OFF
USE_GLOO_BACKEND	Use GLOO RDMA Send/Recv torch backend	Hetero	OFF
BUILD_INTRA_OPS	Use INTRA Collective OPS	GPGPU	OFF
BUILD_INTER_OPS	Use INTER Collective OPS (NVSHMEM)	NVIDIA	OFF

[!Note] Please enable USE_MECA when using DLSlime as a torch backend in Metax platform.

Benchmark

GDRDMA P2P Read/Write

• Platform: NVIDIA ConnectX-7 HHHL Adapter Card; 200GbE (default mode) / NDR200 IB; Dual-port QSFP112; PCIe 5.0 x16 with x16 PCIe extension option; RoCE v2.

#BS=1, #Concurrency=1

torchrun --master-addr 10.130.8.145 --master-port 6006 --nnodes 2 --nproc-per-node 1 --node-rank 1 bench/python/agg_transfer_bench_spmd.py --qp-num 8 --transfer-engine dlslime --batch-size 1 --num-iteration 100 --num-concurrency 1

torchrun --master-addr 10.130.8.145 --master-port 6006 --nnodes 2 --nproc-per-node 1 --node-rank 0 bench/python/agg_transfer_bench_spmd.py --qp-num 8 --transfer-engine dlslime --batch-size 1 --num-iteration 100 --num-concurrency 1

Transfer Engine	#Channels	Message Size (bytes)	Batch Size	Num Concurrency	Avg Latency(ms)	Bandwidth(MB/s)
dlslime	1	2,048	1	1	0.039	52
dlslime	1	4,096	1	1	0.037	111
dlslime	1	8,192	1	1	0.038	216
dlslime	1	16,384	1	1	0.037	442
dlslime	1	32,768	1	1	0.039	836
dlslime	1	65,536	1	1	0.039	1689
dlslime	1	131,072	1	1	0.041	3195
dlslime	1	262,144	1	1	0.043	6059
dlslime	1	524,288	1	1	0.049	10689
dlslime	1	1,048,576	1	1	0.062	17012
dlslime	1	2,097,152	1	1	0.083	25154
dlslime	1	4,194,304	1	1	0.127	33112
dlslime	1	8,388,608	1	1	0.211	39797
dlslime	1	16,777,216	1	1	0.382	43893
dlslime	1	33,554,432	1	1	0.726	46244
dlslime	1	67,108,864	1	1	1.412	47518
dlslime	1	134,217,728	1	1	2.783	48235

#BS=64, #Concurrency=1

torchrun --master-addr 10.130.8.145 --master-port 6006 --nnodes 2 --nproc-per-node 1 --node-rank 1 bench/python/agg_transfer_bench_spmd.py --qp-num 8 --transfer-engine dlslime --batch-size 64 --num-iteration 100 --num-concurrency 1

torchrun --master-addr 10.130.8.145 --master-port 6006 --nnodes 2 --nproc-per-node 1 --node-rank 0 bench/python/agg_transfer_bench_spmd.py --qp-num 8 --transfer-engine dlslime --batch-size 64 --num-iteration 100 --num-concurrency 1

Transfer Engine	#Channels	Message Size (bytes)	Batch Size	Num Concurrency	Avg Latency(ms)	Bandwidth(MB/s)
dlslime	1	2,048	64	1	0.084	1562
dlslime	1	4,096	64	1	0.082	3213
dlslime	1	8,192	64	1	0.086	6095
dlslime	1	16,384	64	1	0.093	11249
dlslime	1	32,768	64	1	0.115	18193
dlslime	1	65,536	64	1	0.158	26542
dlslime	1	131,072	64	1	0.243	34498
dlslime	1	262,144	64	1	0.414	40549
dlslime	1	524,288	64	1	0.758	44248
dlslime	1	1,048,576	64	1	1.443	46510
dlslime	1	2,097,152	64	1	2.809	47782
dlslime	1	4,194,304	64	1	5.555	48327
dlslime	1	8,388,608	64	1	11.041	48624
dlslime	1	16,777,216	64	1	22.003	48798
dlslime	1	33,554,432	64	1	43.941	48872
dlslime	1	67,108,864	64	1	87.809	48912
dlslime	1	134,217,728	64	1	175.512	48942

#BS=64, #Concurrency=8

torchrun --master-addr 10.130.8.145 --master-port 6006 --nnodes 2 --nproc-per-node 1 --node-rank 1 bench/python/agg_transfer_bench_spmd.py --qp-num 8 --transfer-engine dlslime --batch-size 64 --num-iteration 100 --num-concurrency 8

torchrun --master-addr 10.130.8.145 --master-port 6006 --nnodes 2 --nproc-per-node 1 --node-rank 0 bench/python/agg_transfer_bench_spmd.py --qp-num 8 --transfer-engine dlslime --batch-size 64 --num-iteration 100 --num-concurrency 8

Transfer Engine	#Channels	Message Size (bytes)	Batch Size	Num Concurrency	Avg Latency(ms)	Bandwidth(MB/s)
dlslime	1	2,048	64	8	0.037	3519
dlslime	1	4,096	64	8	0.038	6948
dlslime	1	8,192	64	8	0.038	13758
dlslime	1	16,384	64	8	0.04	26416
dlslime	1	32,768	64	8	0.057	36997
dlslime	1	65,536	64	8	0.098	42618
dlslime	1	131,072	64	8	0.184	45602
dlslime	1	262,144	64	8	0.356	47148
dlslime	1	524,288	64	8	0.699	47975
dlslime	1	1,048,576	64	8	1.384	48478
dlslime	1	2,097,152	64	8	2.755	48709
dlslime	1	4,194,304	64	8	5.498	48823
dlslime	1	8,388,608	64	8	10.982	48884
dlslime	1	16,777,216	64	8	21.954	48908
dlslime	1	33,554,432	64	8	43.895	48923
dlslime	1	67,108,864	64	8	87.766	48936
dlslime	1	134,217,728	64	8	175.517	48940

GDRDMA Aggregated Bandwidth

#BS=1, #Concurrency=1

torchrun --master-addr 10.130.8.145 --master-port 6006 --nnodes 2 --nproc-per-node 8 --node-rank 1 bench/python/agg_transfer_bench_spmd.py --qp-num 8 --transfer-engine dlslime --batch-size 1 --num-iteration 100 --num-concurrency 1

torchrun --master-addr 10.130.8.145 --master-port 6006 --nnodes 2 --nproc-per-node 8 --node-rank 0 bench/python/agg_transfer_bench_spmd.py --qp-num 8 --transfer-engine dlslime --batch-size 1 --num-iteration 100 --num-concurrency 1

Transfer Engine	#Channels	Message Size (bytes)	Batch Size	Num Concurrency	Avg Latency(ms)	Bandwidth(MB/s)
dlslime	8	2,048	1	1	0.051	157
dlslime	8	4,096	1	1	0.042	768
dlslime	8	8,192	1	1	0.04	1576
dlslime	8	16,384	1	1	0.054	2929
dlslime	8	32,768	1	1	0.051	5713
dlslime	8	65,536	1	1	0.052	11547
dlslime	8	131,072	1	1	0.055	22039
dlslime	8	262,144	1	1	0.058	42313
dlslime	8	524,288	1	1	0.064	74753
dlslime	8	1,048,576	1	1	0.072	127489
dlslime	8	2,097,152	1	1	0.101	184823
dlslime	8	4,194,304	1	1	0.149	246861
dlslime	8	8,388,608	1	1	0.237	299510
dlslime	8	16,777,216	1	1	0.403	340252
dlslime	8	33,554,432	1	1	0.743	364918
dlslime	8	67,108,864	1	1	1.423	378620
dlslime	8	134,217,728	1	1	2.79	384630

#BS=64, #Concurrency=1

torchrun --master-addr 10.130.8.145 --master-port 6006 --nnodes 2 --nproc-per-node 8 --node-rank 1 bench/python/agg_transfer_bench_spmd.py --qp-num 8 --transfer-engine dlslime --batch-size 64 --num-iteration 100 --num-concurrency 1

torchrun --master-addr 10.130.8.145 --master-port 6006 --nnodes 2 --nproc-per-node 8 --node-rank 0 bench/python/agg_transfer_bench_spmd.py --qp-num 8 --transfer-engine dlslime --batch-size 64 --num-iteration 100 --num-concurrency 1

Transfer Engine	#Channels	Message Size (bytes)	Batch Size	Num Concurrency	Avg Latency(ms)	Bandwidth(MB/s)
dlslime	8	2,048	64	1	0.091	11690
dlslime	8	4,096	64	1	0.081	24403
dlslime	8	8,192	64	1	0.091	45926
dlslime	8	16,384	64	1	0.098	84092
dlslime	8	32,768	64	1	0.117	138696
dlslime	8	65,536	64	1	0.16	206866
dlslime	8	131,072	64	1	0.241	273976
dlslime	8	262,144	64	1	0.415	320008
dlslime	8	524,288	64	1	0.757	353714
dlslime	8	1,048,576	64	1	1.439	372217
dlslime	8	2,097,152	64	1	2.819	381397
dlslime	8	4,194,304	64	1	5.555	386489
dlslime	8	8,388,608	64	1	11.044	388927
dlslime	8	16,777,216	64	1	22.009	390278
dlslime	8	33,554,432	64	1	43.951	390978
dlslime	8	67,108,864	64	1	87.804	391370
dlslime	8	134,217,728	64	1	175.508	391588

#BS=64, #Concurrency=8

torchrun --master-addr 10.130.8.145 --master-port 6006 --nnodes 2 --nproc-per-node 8 --node-rank 1 bench/python/agg_transfer_bench_spmd.py --qp-num 8 --transfer-engine dlslime --batch-size 64 --num-iteration 100 --num-concurrency 8

torchrun --master-addr 10.130.8.145 --master-port 6006 --nnodes 2 --nproc-per-node 8 --node-rank 0 bench/python/agg_transfer_bench_spmd.py --qp-num 8 --transfer-engine dlslime --batch-size 64 --num-iteration 100 --num-concurrency 8

Transfer Engine	#Channels	Message Size (bytes)	Batch Size	Num Concurrency	Avg Latency(ms)	Bandwidth(MB/s)
dlslime	8	2,048	64	8	0.036	28494
dlslime	8	4,096	64	8	0.038	50860
dlslime	8	8,192	64	8	0.048	104545
dlslime	8	16,384	64	8	0.041	207051
dlslime	8	32,768	64	8	0.056	297354
dlslime	8	65,536	64	8	0.099	337571
dlslime	8	131,072	64	8	0.185	363003
dlslime	8	262,144	64	8	0.356	376743
dlslime	8	524,288	64	8	0.701	383701
dlslime	8	1,048,576	64	8	1.386	387629
dlslime	8	2,097,152	64	8	2.757	389493
dlslime	8	4,194,304	64	8	5.5	390523
dlslime	8	8,388,608	64	8	10.984	391043
dlslime	8	16,777,216	64	8	21.955	391291
dlslime	8	33,554,432	64	8	43.891	391407
dlslime	8	67,108,864	64	8	87.771	391480
dlslime	8	134,217,728	64	8	175.518	391530

GDRDMA P2P Send/Recv

SLIME_QP_NUM=2 python bench/python/dlslime_torch_dist_sendrecv_bench.py --mode send --use-gpu --iterations 100

SLIME_QP_NUM=2 python bench/python/dlslime_torch_dist_sendrecv_bench.py --mode recv --use-gpu --iterations 100

Message Size (bytes)	Avg Latency	Bandwidth	Device
1,024	0.027 ms	37.65 MB/s	GPU
2,048	0.028 ms	72.17 MB/s	GPU
4,096	0.028 ms	144.81 MB/s	GPU
8,192	0.028 ms	295.98 MB/s	GPU
16,384	0.029 ms	564.15 MB/s	GPU
32,768	0.031 ms	1069.90 MB/s	GPU
65,536	0.031 ms	2083.20 MB/s	GPU
131,072	0.032 ms	4038.17 MB/s	GPU
262,144	0.036 ms	7299.42 MB/s	GPU
524,288	0.042 ms	12495.87 MB/s	GPU
1,048,576	0.053 ms	19961.18 MB/s	GPU
2,097,152	0.075 ms	27924.99 MB/s	GPU
4,194,304	0.117 ms	35716.55 MB/s	GPU
8,388,608	0.212 ms	39637.66 MB/s	GPU
16,777,216	0.387 ms	43386.08 MB/s	GPU
33,554,432	0.871 ms	38532.98 MB/s	GPU
67,108,864	1.665 ms	40298.91 MB/s	GPU
134,217,728	3.159 ms	42487.69 MB/s	GPU
268,435,456	5.643 ms	47572.53 MB/s	GPU
536,870,912	11.137 ms	48204.20 MB/s	GPU

Heterogeneous Interconnection​

• hardware configs

Device	NIC Model	Bandwidth	PCIe Version	PCIe Lanes
A	Mellanox ConnectX-7 Lx (MT4129)	400 Gbps	PCIe 5.0	x16
B	Mellanox ConnectX-7 Lx (MT4129)	400 Gbps	PCIe 5.0	x8
C	Mellanox ConnectX-7 Lx (MT4129)	200 Gbps	PCIe 5.0	x16
D	Mellanox ConnectX-7 Lx (MT4129)	400 Gbps	PCIe 5.0	x16

• experiments configs

  • Message Size = 128 MB
  • RDMA RC Read(single NIC)
  • Under affinity scenario
  • RDMA with GPU Direct

• Interconnect bandwidth matrix：(MB/s, demonstrates attainment of the theoretical bound).

Throughput (MB/s)	A	B	C	D
A	48967.45	28686.29	24524.29	27676.57
B	28915.72	28275.85	23472.29	27234.60
C	24496.14	24496.51	24513.57	24493.89
D	29317.66	28683.25	24515.30	27491.33

detailed results: bench