huangxialu 6881c19458 [main] convert the format of gmm to nz (#2474) ...

### What this PR does / why we need it?
convert the format of gmm to nz

### Does this PR introduce _any_ user-facing change?
not involved

### How was this patch tested?
ut: test_fused_ops.py and e2e: test_fused_moe.py

**performance**:
(qwen3 30B, 2k->20k)

base:
Total Token throughput (tok/s):          719.93

gmm nz:
Total Token throughput (tok/s):          728.52


- vLLM version: v0.10.1.1
- vLLM main:
bfc1edc9f5

Signed-off-by: huangxialu <huangxialu1@huawei.com>

2025-08-27 11:25:02 +08:00

..

expert_map.json

Add unit test local cpu guide and enable base testcase (#1566)

2025-07-06 10:42:27 +08:00

test_activation.py

[1/N][CustomOp] Register activation customop instead of overwrite forward_oot (#1841)

2025-07-18 23:07:14 +08:00

test_expert_load_balancer.py

Add unit test local cpu guide and enable base testcase (#1566)

2025-07-06 10:42:27 +08:00

test_fused_ops.py

[main] convert the format of gmm to nz (#2474)

2025-08-27 11:25:02 +08:00

test_layernorm.py

[CustomOp] Register RMSNorm instead of overwrite forward_oot (#2284)

2025-08-14 17:18:30 +08:00

test_linear.py

add mlp tp optimze (#2120)

2025-08-21 09:22:07 +08:00

test_rotary_embedding.py

[CustomOp] Register RotaryEmbedding instead of overwrite forward (#2385)

2025-08-25 09:32:35 +08:00

test_token_dispatcher.py

[main] refactor alltoallv in fused_moe (#2487)

2025-08-23 20:38:17 +08:00

test_vocab_parallel_embedding.py

add ut for vocab_parallel_embedding (#2067)

2025-07-30 14:35:45 +08:00