Unverified50% confidenceReleaseExact time
NVIDIA introduced the Global Performance Tuner in TensorRT 11.1.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/23/2026
First Seen
Valid until: 1/21/2027
Sources
Related Entities
Related Claims
VerifiedTensorRT is a high-performance inference optimization engine provided by NVIDIA that compiles PyTorch/ONNX models into GPU-optimized execution engines, typically improving inference speed by 2–5×.76% similarUnverifiedNVIDIA has launched the TensorRT Model Connect tool.75% similarUnverifiedNVIDIA NIM内置了TensorRT-LLM等推理加速引擎75% similarUnverifiedNVIDIA提出了非均匀张量并行（Nonuniform Tensor Parallelism）技术，用于在GPU故障时最大化LLM训练的有效产出（Goodput）73% similarUnverifiedNVIDIA在Ampere架构（RTX 30系，2020年）引入专用INT8 Tensor Core使INT8推理速度相比FP16提升约2倍，NVFP4是为Blackwell（RTX 50系，2025年）定制的4位浮点格式73% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1090989API
curl https://kongchang.com/api/v1/knowledge/claims/1090989MCP
get_claim(id=1090989)