1 day ago
Responsibilities
- Design and implement optimized C++ kernel libraries for NPU and GPU platforms.
- Develop vectorized code using SIMD and instruction-level parallelism for VLIW vector processor units.
- Profile existing kernels, identify bottlenecks, and tune critical sections for higher throughput.
- Develop C++ and Python CPU models for machine-learning operators and write unit and integration tests.
- Validate kernel accuracy and performance across different hardware platforms.
- Collaborate with hardware engineers, machine-learning researchers, and software engineers to integrate kernels into the software stack.
- Document kernel designs and performance improvements and maintain code through Git and pull requests.
Requirements
- Excellent C, C++, and Python programming skills.
- Understanding of SIMD, tensor, and VLIW processor architectures and parallel computing.
- Experience with vectorized programming and performance optimization.
- BS, master's, or PhD degree in computer science, electrical engineering, or a related field, with approximately 10, 8, or 5 years of experience respectively.
- Familiarity with TensorFlow or PyTorch is preferred.
- Experience with silicon bring-up and pre-silicon validation on emulation platforms is preferred.
- Knowledge of cache hierarchy and memory access patterns is desirable.
Benefits
- AMD benefits are described in the AMD benefits at a glance materials.
- The position is an existing vacancy.
About AMD
We care deeply about transforming lives with AMD technology to enrich our industry, our communities, and the world. Our mission is to build great products that accelerate next-generation computing experiences – the building blocks for the data center, artificial intelligence, PCs, gaming and embedded. Underpinning our mission is the AMD culture. We push the limits of innovation to solve the world’s most important challenges. We strive for execution excellence while being direct, humble, collaborative, and inclusive of diverse perspectives. AMD together we advance_