Scalable Matrix Multiplication with Hybrid CMOS-RSFQ Digital Signal Processor
Paper i proceeding, 2007

We report an RSFQ Digital Signal Processor design based on hybrid RSFQ-CMOS memory suitable for a general matrix-on-matrix multiplication algorithm. The design consists of an RSFQ Multiply-Accumulate Unit, memory caches and synchronization block, partitioned into multiple chips, and a large CMOS memory. The complexity of the RSFQ DSP is 10x10 multiplication, rounding to 14 bits, 18-bits accumulator and 4.4 Kb memory cache. The maximum simulated clock frequency is equal to 24 GHz for HYPRES 4.5 kA/cm2 process and optimum communication bandwidth with CMOS memory is 2 Gbps. The simplified version of the RSFQ DSP consisting of 4x4 MAC with rounding to 5 bits and 17x6 memory caches has been designed for HYPRES 4.5 kA/cm2 process and fabricated.

RSFQ

Multiply-Accumulate Unit.

hybrid memory

DSP

Författare

Irina Kataeva

Fasta tillståndets elektronik

Henrik Engseth

Fasta tillståndets elektronik

Samuel Intiso

Chalmers, Mikroteknologi och nanovetenskap (MC2)

Anna Kidiyarova-Shevchenko

Fasta tillståndets elektronik

oral presentation Applied Superconductivity Conference (ASC) Sept 2006, Seattle, will be published in IEEE Transactions on Applied Superconductivity June 2007

Vol. June

Ämneskategorier

TEKNIK OCH TEKNOLOGIER

Mer information

Skapat

2017-10-06