GPU Benchmark Results

• Target profile: nvidia-gpu
• Suite: gpu/einsum
• Suite file: benchmarks/gpu/einsum.yaml
• Timestamp: 2026-10-06T19:36:19.553911+00:00
• tenferro-rs commit: 5cf78c7ec0ad9516dd78bab546d5e7bd42fa3102

GPU Information

• Device: cuda:0
• Name: NVIDIA A100 80GB PCIe
• UUID: GPU-530977e1-4968-9283-4129-9fbec3e66542
• Memory: 80 GiB
• Driver version: 580.126.09
• CUDA version: 13.0
• CUDA runtime: 12.9
• cuDNN version: 92700

CPU Information

• Model: AMD EPYC 7713P 64-Core Processor
• Vendor: AuthenticAMD
• Logical CPUs: 64
• Sockets: 1
• Cores per socket: 64
• Threads per core: 1
• NUMA nodes: 1
• Python platform: Linux-6.8.0-101-generic-x86_64-with-glibc2.39

Median time is reported in milliseconds for ok records. Inputs are prepared on the GPU before timed runs; initial host-to-device transfer is outside the timed region. Timed runs include the host API call and backend-native device synchronization. tenferro-rs CUDA uses the explicit tenferro-rs synchronize API without downloading result tensors in the timed region. Dense and einsum inputs use the same deterministic benchmark generator in the Rust and Python runners; host-to-device layout conversion remains outside the timed region. tenferro-rs uses native column-major GPU tensors; PyTorch and vendor-wrapper columns use their native row-major framework tensors unless noted. The cuSOLVER column is torch.linalg with preferred_linalg_library=cusolver; for SVD it pins driver=gesvd as a QR-based cuSOLVER comparison. tenferro-rs CUDA SVD uses its backend default driver policy, currently gesvdj for matrices with both dimensions at most 1024 and gesvd otherwise. Non-ok cells show the structured backend status.

gpu/einsum / einsum / allocating output