Discussion on performance
I have been working on performance optimizations on Qwen3-TTS at high concurrency on H100. I noticed through profiling that there was a lot of cpu bound work that appeared to be tied to serialization/deserialization when using the multiproc (mp) executor which compounds at higher concurrencies which appears to be eliminated by switching to the uniproc (uni) executor.
I just ran benchmark_tts_serve.py (Base cloning task) using the stock bs16 config and then the same config only switching the distributed_executor_backend to uni and saw across the board especially at higher concurrency (16) with an approximately 130% reduction in RTF.
Should we be defaulting to this for Qwen3-TTS in single-gpu setups?
Your current environment (if you think it is necessary)
==============================
System Info
==============================
OS : Debian GNU/Linux 12 (bookworm) (x86_64)
GCC version : (Debian 12.2.0-14+deb12u1) 12.2.0
Clang version : Could not collect
CMake version : Could not collect
Libc version : glibc-2.36
==============================
PyTorch Info
==============================
PyTorch version : 2.10.0+cu128
Is debug build : False
CUDA used to build PyTorch : 12.8
ROCM used to build PyTorch : N/A
==============================
Python Environment
==============================
Python version : 3.12.13 (main, Mar 24 2026, 22:49:22) [Clang 22.1.1 ] (64-bit runtime)
Python platform : Linux-5.15.0-156-generic-x86_64-with-glibc2.36
==============================
CUDA / GPU Info
==============================
Is CUDA available : True
CUDA runtime version : Could not collect
CUDA_MODULE_LOADING set to :
GPU models and configuration : GPU 0: NVIDIA H100 80GB HBM3
Nvidia driver version : 570.148.08
cuDNN version : Could not collect
HIP runtime version : N/A
MIOpen runtime version : N/A
Is XNNPACK available : True
==============================
CPU Info
==============================
Architecture: x86_64
CPU op-mode(s): 32-bit, 64-bit
Address sizes: 48 bits physical, 57 bits virtual
Byte Order: Little Endian
CPU(s): 128
On-line CPU(s) list: 0-127
Vendor ID: GenuineIntel
Model name: Intel(R) Xeon(R) Platinum 8468
CPU family: 6
Model: 143
Thread(s) per core: 2
Core(s) per socket: 32
Socket(s): 2
Stepping: 8
BogoMIPS: 4200.00
Flags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts mmx fxsr sse sse2 ss ht syscall nx pdpe1gb rdtscp lm constant_tsc arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc cpuid tsc_known_freq pni pclmulqdq dtes64 ssse3 fma cx16 pdcm pcid sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand hypervisor lahf_lm abm 3dnowprefetch cpuid_fault invpcid_single ssbd ibrs ibpb stibp ibrs_enhanced fsgsbase tsc_adjust bmi1 avx2 smep bmi2 erms invpcid avx512f avx512dq rdseed adx smap avx512ifma clflushopt clwb avx512cd sha_ni avx512bw avx512vl xsaveopt xsavec xgetbv1 xsaves avx_vnni avx512_bf16 wbnoinvd arat avx512vbmi umip pku ospke waitpkg avx512_vbmi2 gfni vaes vpclmulqdq avx512_vnni avx512_bitalg avx512_vpopcntdq la57 rdpid bus_lock_detect cldemote movdiri movdir64b fsrm md_clear serialize tsxldtrk amx_bf16 avx512_fp16 amx_tile amx_int8 arch_capabilities
Hypervisor vendor: KVM
Virtualization type: full
L1d cache: 4 MiB (128 instances)
L1i cache: 4 MiB (128 instances)
L2 cache: 256 MiB (64 instances)
L3 cache: 32 MiB (2 instances)
NUMA node(s): 2
NUMA node0 CPU(s): 0-63
NUMA node1 CPU(s): 64-127
Vulnerability Gather data sampling: Not affected
Vulnerability Indirect target selection: Mitigation; Aligned branch/return thunks
Vulnerability Itlb multihit: Not affected
Vulnerability L1tf: Not affected
Vulnerability Mds: Not affected
Vulnerability Meltdown: Not affected
Vulnerability Mmio stale data: Unknown: No mitigations
Vulnerability Reg file data sampling: Not affected
Vulnerability Retbleed: Not affected
Vulnerability Spec rstack overflow: Not affected
Vulnerability Spec store bypass: Mitigation; Speculative Store Bypass disabled via prctl and seccomp
Vulnerability Spectre v1: Mitigation; usercopy/swapgs barriers and __user pointer sanitization
Vulnerability Spectre v2: Mitigation; Enhanced / Automatic IBRS; IBPB conditional; PBRSB-eIBRS SW sequence; BHI SW loop, KVM SW loop
Vulnerability Srbds: Not affected
Vulnerability Tsa: Not affected
Vulnerability Tsx async abort: Mitigation; TSX disabled
==============================
Versions of relevant libraries
==============================
[pip3] ema-pytorch==0.7.9
[pip3] flashinfer-python==0.6.6
[pip3] numpy==2.2.6
[pip3] nvidia-cublas==13.1.0.3
[pip3] nvidia-cublas-cu12==12.8.4.1
[pip3] nvidia-cuda-cupti==13.0.85
[pip3] nvidia-cuda-cupti-cu12==12.8.90
[pip3] nvidia-cuda-nvrtc==13.0.88
[pip3] nvidia-cuda-nvrtc-cu12==12.8.93
[pip3] nvidia-cuda-runtime==13.0.96
[pip3] nvidia-cuda-runtime-cu12==12.8.90
[pip3] nvidia-cudnn-cu12==9.10.2.21
[pip3] nvidia-cudnn-cu13==9.19.0.56
[pip3] nvidia-cudnn-frontend==1.18.0
[pip3] nvidia-cufft==12.0.0.61
[pip3] nvidia-cufft-cu12==11.3.3.83
[pip3] nvidia-cufile==1.15.1.6
[pip3] nvidia-cufile-cu12==1.13.1.3
[pip3] nvidia-curand==10.4.0.35
[pip3] nvidia-curand-cu12==10.3.9.90
[pip3] nvidia-cusolver==12.0.4.66
[pip3] nvidia-cusolver-cu12==11.7.3.90
[pip3] nvidia-cusparse==12.6.3.3
[pip3] nvidia-cusparse-cu12==12.5.8.93
[pip3] nvidia-cusparselt-cu12==0.7.1
[pip3] nvidia-cusparselt-cu13==0.8.0
[pip3] nvidia-cutlass-dsl==4.4.2
[pip3] nvidia-cutlass-dsl-libs-base==4.4.2
[pip3] nvidia-ml-py==13.595.45
[pip3] nvidia-nccl-cu12==2.27.5
[pip3] nvidia-nccl-cu13==2.28.9
[pip3] nvidia-nvjitlink==13.0.88
[pip3] nvidia-nvjitlink-cu12==12.8.93
[pip3] nvidia-nvshmem-cu12==3.4.5
[pip3] nvidia-nvshmem-cu13==3.4.5
[pip3] nvidia-nvtx==13.0.85
[pip3] nvidia-nvtx-cu12==12.8.90
[pip3] onnxruntime==1.24.4
[pip3] pyzmq==27.1.0
[pip3] torch==2.10.0
[pip3] torch-c-dlpack-ext==0.1.5
[pip3] torchaudio==2.10.0
[pip3] torchsde==0.2.6
[pip3] torchvision==0.25.0
[pip3] transformers==4.57.6
[pip3] triton==3.6.0
[pip3] x-transformers==2.17.9
[conda] Could not collect
==============================
vLLM Info
==============================
ROCM Version : Could not collect
vLLM Version : 0.19.0
vLLM-Omni Version : 0.19.0rc2.dev47+gfcda835f4 (git sha: fcda835f4)
vLLM Build Flags:
CUDA Archs: Not Set; ROCm: Disabled
GPU Topology:
GPU0 NIC0 NIC1 NIC2 NIC3 NIC4 NIC5 NIC6 NIC7 CPU Affinity NUMA Affinity GPU NUMA ID
GPU0 X PHB PHB PHB PIX SYS SYS SYS SYS 64-127 1 N/A
NIC0 PHB X PHB PHB PHB SYS SYS SYS SYS
NIC1 PHB PHB X PHB PHB SYS SYS SYS SYS
NIC2 PHB PHB PHB X PHB SYS SYS SYS SYS
NIC3 PIX PHB PHB PHB X SYS SYS SYS SYS
NIC4 SYS SYS SYS SYS SYS X PHB PHB PHB
NIC5 SYS SYS SYS SYS SYS PHB X PHB PHB
NIC6 SYS SYS SYS SYS SYS PHB PHB X PHB
NIC7 SYS SYS SYS SYS SYS PHB PHB PHB X
Legend:
X = Self
SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI)
NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node
PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU)
PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge)
PIX = Connection traversing at most a single PCIe bridge
NV# = Connection traversing a bonded set of # NVLinks
NIC Legend:
NIC0: mlx5_0
NIC1: mlx5_1
NIC2: mlx5_2
NIC3: mlx5_3
NIC4: mlx5_4
NIC5: mlx5_5
NIC6: mlx5_6
NIC7: mlx5_7
==============================
Environment Variables
==============================
NVIDIA_VISIBLE_DEVICES=GPU-ec02b276-8098-65ab-b8c3-ef414ac052be
LD_LIBRARY_PATH=/root/.venv/lib/python3.12/site-packages/cv2/../../lib64:/usr/local/lib/python3.10/site-packages/nvidia/cudnn/lib:$LD_LIBRARY_PATH
PYTORCH_NVML_BASED_CUDA_CHECK=1
TORCHINDUCTOR_COMPILE_THREADS=1
TORCHINDUCTOR_CACHE_DIR=/tmp/torchinductor_root
Before submitting a new issue...
Discussion on performance
I have been working on performance optimizations on Qwen3-TTS at high concurrency on H100. I noticed through profiling that there was a lot of cpu bound work that appeared to be tied to serialization/deserialization when using the multiproc (mp) executor which compounds at higher concurrencies which appears to be eliminated by switching to the uniproc (uni) executor.
I just ran
benchmark_tts_serve.py(Base cloning task) using the stock bs16 config and then the same config only switching thedistributed_executor_backendtouniand saw across the board especially at higher concurrency (16) with an approximately 130% reduction in RTF.Should we be defaulting to this for Qwen3-TTS in single-gpu setups?
Your current environment (if you think it is necessary)
Before submitting a new issue...