...
TDT01 is a specialization course focusing on specialized topics within computer architecture. This year's edition will focus on Design, Microarchitecture, GPUs, and other advanced topics.
Course responsible: Rakesh Kumar
...
Meeting 4: Discussion 3 -
Meeting 5: Discussion 4 -
Reading list
Note: To access the full versions of the papers from IEEE and ACM you need to have an NTNU IP address. If you are not on campus, try VPN or a terminal server.
Design:
1. Is Dark Silicon Useful? Harnessing the Four Horsemen of the Coming Dark Silicon ApocalypseFireSim: FPGA-Accelerated Cycle-Exact Scale-OutSystem Simulation in the Public Cloud
2. Software and the Concurrency Revolution An Agile Approach to Building RISC-V Microprocessors
3. Needle : Leveraging Program Analysis to Analyze and Extract Accelerators from Whole Programs
Microarchitecture:BISMO: A Scalable Bit-Serial Matrix Multiplication Overlay for Reconfigurable Computing
4. Efficient Invisible Speculative Execution Through Selective Delay and Value Prediction
5. Intermittent Computing: Challenges and Opportunities
6. The load slice core microarchitecture 7.
Dynamo: a transparent dynamic optimization system 8 5. Boomerang: a Metadata-Free Architecture for Control Flow Delivery
6. Evolution of the Samsung Exynos CPU Microarchitecture
GPU:
7 9. Get out of the valley: Power-efficient address mapping for GPUs
108. HSM: A Hybrid Slowdown Model for Multitasking GPUs
119. MDM: The GPU Memory Divergence Model Model
Misc.:
10. Optimizing Bit-Serial Matrix Multiplication for Reconfigurable Computing
11. Understanding Selective Delay as a Method for Efficient Secure Speculative Execution
12. Needle : Leveraging Program Analysis to Analyze and Extract Accelerators from Whole ProgramsIntermittent Computing: Challenges and Opportunities