Silicon photonic memory interconnect for many-core architectures

A scalable and flexible memory interconnect is a key component for a many-core architecture to take full advantage of the high-bandwidth of multiple memory stacks. In this paper, we discuss both technological and architectural challenges of these processor-to-memory interconnects, and focus on two important issues of many-core memory accesses: traffic hotspots and non-uniform memory access (NUMA). We propose a reconfigurable Silicon photonic memory interconnect based on 2.5D stacking that can direct memory traffic to any memory interface on the processor, thus alleviating the two aforementioned effects in addition to providing high bandwidth. Simulations based on a 16-core 4-memory model show that the proposed architecture can lead to up to 2× STREAM speedup over fixed connections in both hotspot and NUMA scenarios. We also demonstrate the proposed architecture using a four-port Silicon photonic demultiplexer and a 4×4 synthesizable on-chip fabric called OpenSoC. The FPGA-emulated system demonstrates dynamic memory rewiring through wavelength routing, and achieves a reconfiguration time of 5 microseconds.

[1]  C R Doerr,et al.  Monolithically Integrated 40-Wavelength Demultiplexer and Photodetector Array on Silicon , 2011, IEEE Photonics Technology Letters.

[2]  Xuezhe Zheng,et al.  Silicon Photonic Interconnects for Large-Scale Computer Systems , 2013, IEEE Micro.

[3]  H. Thacker,et al.  A tunable 1x4 silicon CMOS photonic wavelength multiplexer/demultiplexer for dense optical interconnects. , 2010, Optics express.

[4]  Peter De Dobbelaere,et al.  Packaging of silicon photonics systems , 2014, OFC 2014.

[5]  Michal Lipson,et al.  10-Gb/s WDM optically-connected memory system using silicon microring modulators , 2012, 2012 38th European Conference and Exhibition on Optical Communications.

[6]  Mahmut T. Kandemir,et al.  Optimizing off-chip accesses in multicores , 2015, PLDI.

[7]  Alexander V. Rylyakov,et al.  Monolithic Silicon Integration of Scaled Photonic Switch Fabrics, CMOS Logic, and Device Driver Circuits , 2014, Journal of Lightwave Technology.

[8]  Bruce Jacob,et al.  The structural simulation toolkit , 2006, PERV.

[9]  Xuezhe Zheng,et al.  Submilliwatt, ultrafast and broadband electro-optic silicon switches. , 2010, Optics express.

[10]  Vivien Quéma,et al.  Traffic management: a holistic approach to memory placement on NUMA systems , 2013, ASPLOS '13.

[11]  Wim Bogaerts,et al.  Photonics-CMOS 3D integration: copper through-silicon-via approach , 2009 .

[12]  Rajeev J. Ram,et al.  Single-chip microprocessor that communicates directly using light , 2015, Nature.

[13]  K. Bergman,et al.  Thermal stabilization of a microring modulator using feedback control. , 2012, Optics express.

[14]  John Wawrzynek,et al.  Chisel: Constructing hardware in a Scala embedded language , 2012, DAC Design Automation Conference 2012.

[15]  Simon D. Hammond,et al.  End-to-End Modeling and Optimization of Power Consumption in HPC Interconnects , 2016, 2016 45th International Conference on Parallel Processing Workshops (ICPPW).

[16]  Jie Sun,et al.  Adiabatic thermo-optic Mach-Zehnder switch. , 2013, Optics letters.

[17]  Reetuparna Das,et al.  Application-to-core mapping policies to reduce memory system interference in multi-core systems , 2013, 2013 IEEE 19th International Symposium on High Performance Computer Architecture (HPCA).

[18]  John Shalf,et al.  OpenSoC Fabric: On-Chip Network Generator: Using Chisel to Generate a Parameterizable On-Chip Interconnect Fabric , 2014, NoCArc '14.

[19]  Yang Liu,et al.  A reconfigurable and redundant optically-connected memory system using a silicon photonic switch , 2014, OFC 2014.

[20]  Yang Liu,et al.  Programmable wavelength locking and routing in a silicon-photonic interconnection network implementation , 2015, 2015 Optical Fiber Communications Conference and Exhibition (OFC).

[21]  Yasuhiko Arakawa,et al.  Silicon photonics for next generation system integration platform , 2013, IEEE Communications Magazine.

[22]  Mahmut T. Kandemir,et al.  Network footprint reduction through data access and computation placement in NoC-based manycores , 2015, 2015 52nd ACM/EDAC/IEEE Design Automation Conference (DAC).

[23]  Reetuparna Das,et al.  Application-to-core mapping policies to reduce memory interference in multi-core systems , 2012, 2012 21st International Conference on Parallel Architectures and Compilation Techniques (PACT).

[24]  D.A.B. Miller,et al.  Rationale and challenges for optical interconnects to electronic chips , 2000, Proceedings of the IEEE.