[Bug]: TorchSDPAMetadata is out of date

[Reichenbachian]

Your current environment

alexr@roxy:~$ python collect_env.py Collecting environment information... PyTorch version: 2.0.0+cu117 Is debug build: False CUDA used to build PyTorch: 11.7 ROCM used to build PyTorch: N/A

OS: Ubuntu 22.04.1 LTS (x86_64) GCC version: (Ubuntu 11.4.0-1ubuntu1~22.04) 11.4.0 Clang version: Could not collect CMake version: version 3.26.3 Libc version: glibc-2.35

Python version: 3.10.12 (main, Nov 20 2023, 15:14:05) [GCC 11.4.0] (64-bit runtime) Python platform: Linux-5.15.0-105-generic-x86_64-with-glibc2.35 Is CUDA available: True CUDA runtime version: 11.5.119 CUDA_MODULE_LOADING set to: LAZY GPU models and configuration: GPU 0: NVIDIA GeForce RTX 3070 Nvidia driver version: 535.161.08 cuDNN version: Could not collect HIP runtime version: N/A MIOpen runtime version: N/A Is XNNPACK available: True

CPU: Architecture: x86_64 CPU op-mode(s): 32-bit, 64-bit Address sizes: 39 bits physical, 48 bits virtual Byte Order: Little Endian CPU(s): 16 On-line CPU(s) list: 0-15 Vendor ID: GenuineIntel Model name: Intel(R) Core(TM) i7-10700K CPU @ 3.80GHz CPU family: 6 Model: 165 Thread(s) per core: 2 Core(s) per socket: 8 Socket(s): 1 Stepping: 5 CPU max MHz: 5100.0000 CPU min MHz: 800.0000 BogoMIPS: 7599.80 Flags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc art arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc cpuid aperfmperf pni pclmulqdq dtes64 monitor ds_cpl smx est tm2 ssse3 sdbg fma cx16 xtpr pdcm pcid sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm abm 3dnowprefetch cpuid_fault invpcid_single ssbd ibrs ibpb stibp ibrs_enhanced fsgsbase tsc_adjust bmi1 avx2 smep bmi2 erms invpcid mpx rdseed adx smap clflushopt intel_pt xsaveopt xsavec xgetbv1 xsaves dtherm ida arat pln pts pku ospke md_clear flush_l1d arch_capabilities L1d cache: 256 KiB (8 instances) L1i cache: 256 KiB (8 instances) L2 cache: 2 MiB (8 instances) L3 cache: 16 MiB (1 instance) NUMA node(s): 1 NUMA node0 CPU(s): 0-15 Vulnerability Gather data sampling: Mitigation; Microcode Vulnerability Itlb multihit: KVM: Mitigation: VMX unsupported Vulnerability L1tf: Not affected Vulnerability Mds: Not affected Vulnerability Meltdown: Not affected Vulnerability Mmio stale data: Mitigation; Clear CPU buffers; SMT vulnerable Vulnerability Retbleed: Mitigation; Enhanced IBRS 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 IBRS, IBPB conditional, RSB filling, PBRSB-eIBRS SW sequence Vulnerability Srbds: Mitigation; Microcode Vulnerability Tsx async abort: Not affected

Versions of relevant libraries: [pip3] flake8==6.0.0 [pip3] mypy-extensions==1.0.0 [pip3] numpy==1.24.2 [pip3] nvidia-nccl-cu11==2.14.3 [pip3] sentence-transformers==2.2.2 [pip3] torch==2.0.0 [pip3] torchvision==0.15.1 [pip3] transformers==4.26.0 [pip3] triton==2.0.0 [conda] Could not collect ROCM Version: Could not collect Neuron SDK Version: N/A vLLM Version: N/A vLLM Build Flags: CUDA Archs: Not Set; ROCm: Disabled; Neuron: Disabled GPU Topology: GPU0 CPU Affinity NUMA Affinity GPU NUMA ID GPU0 X 0-15 0 N/A

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

🐛 Describe the bug

I believe the TorchSDPA backend for attention is no longer compatible. I received a crash message that said as much and did some debugging. If you set export VLLM_ATTENTION_BACKEND=TORCH_SDPA, then the metadata is set with

                attn_metadata = self.attn_backend.make_metadata(
                    num_prefills=0,
                    num_prefill_tokens=0,
                    num_decode_tokens=batch_size,
                    slot_mapping=slot_mapping[:batch_size],
                    seq_lens=None,
                    seq_lens_tensor=seq_lens[:batch_size],
                    max_query_len=None,
                    max_prefill_seq_len=0,
                    max_decode_seq_len=self.max_seq_len_to_capture,
                    query_start_loc=None,
                    seq_start_loc=None,
                    context_lens_tensor=None,
                    block_tables=block_tables[:batch_size],
                    use_cuda_graph=True,
                )

which is sent to

    @staticmethod
    def make_metadata(*args, **kwargs) -> "TorchSDPAMetadata":
        return TorchSDPAMetadata(*args, **kwargs)

and neither TorchSDPAMetadata (nor its superclasses) have most of those values

@dataclass
class TorchSDPAMetadata(AttentionMetadata, PagedAttentionMetadata):
    """Metadata for TorchSDPABackend.
    """
    # Currently, input sequences can only contain all prompts
    # or all decoding. True if all sequences are prompts.
    is_prompt: bool
    slot_mapping: torch.Tensor
    seq_lens: Optional[List[int]]
    # Maximum query length in the batch. None for decoding.
    max_query_len: Optional[int]

Am I correct in my diagnosis? I can send a crash log if necessary, but I believe this is pretty self evidently an error (and a CI test should be added or the backend should be removed probably.)

[simon-mo]

cc @bigPYJ1151 @jikunshang @zhouyuan

I'm surprised this is the case since this backend should be already be tested on CPU machine

[WoosukKwon]

@Reichenbachian If I understand correctly, you are using the TORCH_SDPA backend with NVIDIA GPUs. The backend is for Intel (x86) CPUs and should not be used with NVIDIA GPUs. (cc @simon-mo)