Algorithm Acceleration for RISC-V Processors using High-Level Synthesis - Russell Klein, Siemens EDA
Algorithm Acceleration for RISC-V Processors using High-Level Synthesis - Russell Klein, Siemens EDA
Increasing the performance and efficiency of a RISC-V designs can be done by moving compute intensive functions from software, running on the CPU, into hardware accelerators. These accelerators can take the form of new instructions, co-processors, or bus based peripherals. One way to create this type of accelerator is to use High-Level Synthesis (HLS). HLS takes a C or C function and synthesizes a functionally equivalent Register Transfer Level (RTL) hardware description. This RTL description can be used to create the hardware accelerator for an ASIC or FPGA. This session will show you how to create different types of accelerators using HLS and how to integrate them into RISC-V processor designs. An AI inferencing algorithm and a an encryption algorithm will be used as examples to show how the accelerator can impact the performance and power consumption of a RISC-V design.
For more info about RISC-V, a fre
1 view
0
0
3 years ago 00:26:10 1
Algorithm Acceleration for RISC-V Processors using High-Level Synthesis - Russell Klein, Siemens EDA
5 years ago 00:30:04 6
SLIDE: Smart Algorithms over Hardware Acceleration for Large-Scale Deep Learning with Beidi Chen...
6 years ago 00:39:38 1
Visualization Class 12-The Shear Warp Algorithm
5 years ago 00:25:55 1
Edwin Black: The Algorithm Ghetto, Cashless Society & Newgenics
6 years ago 00:29:50 8
Performance Tuning Architectures with the Intel® FPGA Deep Learning Acceleration Suite
7 years ago 00:45:34 1
Visualization Lecture 13.2. The Shear Warp Factorization Algorithm, Run Length Encoding
6 years ago 00:03:28 1
3D Tracking with IMU
9 years ago 00:27:05 37
Using SGEMM and FFTs to Accelerate Deep Learning (Gian Marco Iodice, ARM)
3 years ago 00:03:55 1
Clusta - Accelerate [UTM-Records]
9 months ago 00:10:20 1
La chute s’accélère…
2 years ago 00:53:09 1
Accelerating Transformers via Kernel Density Estimation Insu Han