Scaling up silicon photonic-based accelerators: Challenges and opportunities

Al-Qadasi, Mohammed; Chrostowski, Lukas; Shastri, Bhavin J.; Shekhar, Sudip

doi:10.1063/5.0070992

Cited by 60 publications

(25 citation statements)

References 102 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Compared to digital accelerators in the latest CMOS processes, optical or optoelectronic MAC operations in SOI-based PIC platforms can process signals with lower latency and higher throughput 51 . The optical link budget and total energy efficiency are discussed in this section to shed light on the advantages and challenges of implementing MAC operations using our proposed photonic processing unit with IM-MRMs.…”

Section: Discussionmentioning

confidence: 99%

“…The optical link budget is calculated based on the following equation 51 : where is the optical power of the laser, is the insertion loss of the EAM, is the loss introduced by the attenuation of the single-mode optical fiber (SMF) and the fiber-to-chip coupling loss, is the silicon waveguide attenuation, is the splitter insertion and excess loss, is the insertion loss of the add-drop IM-MRR filter at the input vector wavelength, is the inter-channel crosstalk penalty, and is the out-of-band insertion loss when the resonant peak is not aligned with input wavelengths, and is the network penalty due to extinction ratio, crosstalk, and laser relative intensity noise (RIN). Considering = 0 dBm, = 1.6 dB 51 , = 6.2 dB 53 , = 2.5 dB/cm where L in centimeters is the waveguide spacing between the two MRMs, = 3.3 log dB, = 1.25 dB, = 3 dB, = 0.01 dB 51 , and = 4.8 dB 51 , the optical link budget as a function of the dimensions of the proposed photonic processing system is calculated and plotted in Fig. 12 a.…”

Section: Discussionmentioning

confidence: 99%

“…The loss can be compensated by either increasing the laser’s output power or adding SOAs to the system. It is worth mentioning that whole SOAs can pre-amplify the input signal, the non-linear gain-current curve would need to be calibration for 51 .

Figure 11 Schematic of the proposed IM-MRM-based photonic processing system, where a set of high-speed EAMs (or EOMs) are used for input data encoding.…”

Section: Discussionmentioning

confidence: 99%

See 2 more Smart Citations

Towards a high-density photonic tensor core enabled by intensity-modulated microrings and photonic wire bonding

Luan

Salmani

et al. 2023

Sci Rep

View full text Add to dashboard Cite

We propose a photonic processing unit for high-density analog computation using intensity-modulation-based microring modulators (IM-MRMs). The output signal at the fixed resonance wavelength is directly intensity modulated by changing the extinction ratio (ER) of the IM-MRM . Thanks to the intensity-modulated approach, the proposed photonic processing unit is less sensitive to the inter-channel crosstalk. Simulation results reveal that the proposed design offers a maximum of 17-fold increase in wavelength channel density compared to its wavelength-modulated counterpart. Therefore, a photonic tensor core of size 512 $$\times $$ × 512 can be realized by current foundry lines. A convolutional neural network (CNN) simulator with 6-bit precision is built for handwritten digit recognition task using the proposed modulator. Simulation results show an overall accuracy of 96.76%, when the wavelength channel spacing suffers a 3-dB power penalty. To experimentally validate the system, 1000 dot product operations are carried out with a 4-bit signed system on a co-packaged photonic chip, where optical and electrical I/Os are realized using photonic and electrical wire bonding techniques. Study of the measurement results show a mean squared error (MSE) of 3.09$$\times $$ × 10$$^{-3}$$ - 3 for dot product calculations. The proposed IM-MRM, therefore, renders the crosstalk issue tractable and provides a solution for the development of large-scale optical information processing systems with multiple wavelengths.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

Figure 11 Schematic of the proposed IM-MRM-based photonic processing system, where a set of high-speed EAMs (or EOMs) are used for input data encoding.…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

Towards a high-density photonic tensor core enabled by intensity-modulated microrings and photonic wire bonding

Luan

Salmani

et al. 2023

Sci Rep

View full text Add to dashboard Cite

show abstract

“…Compared to digital accelerators in the latest CMOS processes, optical or optoelectronic MAC operations in SOI-based PIC platforms can process signals with lower latency and higher throughput 50 . The optical link budget and total energy efficiency are discussed in this section to shed light on the advantages and challenges of implementing MAC operations using our proposed photonic processing unit with IM-MRMs.…”

Section: Discussionmentioning

confidence: 99%

“…The schematic of the N × M photonic processing system is illustrated in Figure 11. The optical link budget is calculated based on the following equation 50 :…”

Section: Discussionmentioning

confidence: 99%

Towards a High-density Photonic Tensor Core Enabled by Intensity-modulated Microrings and Photonic Wire Bonding

Luan

Salmani

et al. 2022

Preprint

View full text Add to dashboard Cite

We propose a photonic processing unit for high-density analog computation using intensity-modulation-based microringmodulators (IM-MRMs). The output signal at the fixed resonance wavelength is directly intensity modulated by changing theextinction ratio (ER) of the IM-MRM. Thanks to the intensity-modulated approach, the proposed photonic processing unit isless sensitive to the inter-channel crosstalk. Simulation results reveal that the proposed design offers a maximum of 17-foldincrease in wavelength channel density compared to its wavelength-modulated counterpart. Therefore, a photonic tensor coreof size 512 × 512 can be realized by current foundry lines. A convolutional neural network (CNN) simulator with 6-bit precisionis built for handwritten digit recognition task using the proposed modulator. Simulation results show an overall accuracy of 96.76%, when the wavelength channel spacing suffers a 3-dB power penalty. To experimentally validate the system, 1000 dotproduct operations are carried out with a 4-bit signed system on a co-packaged photonic chip, where optical and electrical I/Osare realized using photonic and electrical wire bonding techniques. Study of the measurement results show a mean squarederror (MSE) of 3.09 × 10^−3 for dot product calculations. The proposed IM-MRM, therefore, renders the crosstalk issue tractableand provides a solution for the development of large-scale optical information processing systems with multiple wavelengths.

show abstract

Neuromorphic Computing via Fission‐based Broadband Frequency Generation

Fischer,

Chemnitz,

Zhu

et al. 2023

Advanced Science

View full text Add to dashboard Cite

The performance limitations of traditional computer architectures have led to the rise of brain‐inspired hardware, with optical solutions gaining popularity due to the energy efficiency, high speed, and scalability of linear operations. However, the use of optics to emulate the synaptic activity of neurons has remained a challenge since the integration of nonlinear nodes is power‐hungry and, thus, hard to scale. Neuromorphic wave computing offers a new paradigm for energy‐efficient information processing, building upon transient and passively nonlinear interactions between optical modes in a waveguide. Here, an implementation of this concept is presented using broadband frequency conversion by coherent higher‐order soliton fission in a single‐mode fiber. It is shown that phase encoding on femtosecond pulses at the input, alongside frequency selection and weighting at the system output, makes transient spectro‐temporal system states interpretable and allows for the energy‐efficient emulation of various digital neural networks. The experiments in a compact, fully fiber‐integrated setup substantiate an anticipated enhancement in computational performance with increasing system nonlinearity. The findings suggest that broadband frequency generation, accessible on‐chip and in‐fiber with off‐the‐shelf components, may challenge the traditional approach to node‐based brain‐inspired hardware design, ultimately leading to energy‐efficient, scalable, and dependable computing with minimal optical hardware requirements.

show abstract

Scaling up silicon photonic-based accelerators: Challenges and opportunities

Cited by 60 publications

References 102 publications

Towards a high-density photonic tensor core enabled by intensity-modulated microrings and photonic wire bonding

Towards a high-density photonic tensor core enabled by intensity-modulated microrings and photonic wire bonding

Towards a High-density Photonic Tensor Core Enabled by Intensity-modulated Microrings and Photonic Wire Bonding

Neuromorphic Computing via Fission‐based Broadband Frequency Generation

Contact Info

Product

Resources

About