This is the beginning of a list of recommended papers on GPUs. In no way is this a complete or exhaustive list. Instead, it's intended to cover some of the most important, relevant papers.
- Textbook: Tor Aamodt, Wilson Fung, and Tim Rogers have published a Synthesis Lecture on general-purpose GPUs. It covers many, but not all of the papers discussed below.
- Textbook #2: Another good Synthesis Lecture on GPGPUs.
Foundational Papers
- Lindholm IEEE Micro '08: arguably the original paper introducing the idea of GPGPU computing
- Owens IEEE '08: another early GPGPU paper
- Nickolls IEEE Micro '10: one of the first key GPGPU papers.
- Keckler IEEE Micro '11: builds on Nickolls, talks about future of GPUs
- Villa SC '16: follow-up to the Keckler '11 paper
- Vijayaraghavan HPCA '17: introduces AMD's model for the future of GPUs (a dual to the Keckler '11 paper, in some regards)
Simulation
- Khairy Arxiv '18: updating GPGPU-Sim to faithfully model Pascal- and Volta-class GPUs
- Lew Arxiv '18: shows how to get machine learning workloads to run in GPGPU-Sim
- Gutierrez HPCA '18: AMD's new GPU (APU) model, in process of being integrated into gem5 public simulator
- Power IEEE CAL '15
- Leng ISCA '13
- Ubal PACT '12
- Bakhoda ISPASS '09: introduces GPGPU-Sim
Memory Consistency
Cache Coherence
Address Translation
Cache Optimizations
Scheduling
DRAM
OS Support