Data Encoding Techniques for Reducing Energy Consumption in Network-on-Chip

Odd/Even Invert coding for phase change memory with thermal crosstalk

“…Nevertheless, solutions dealing with thermal crosstalk problem are still in their infancy stage. On contrary, a rich body of knowledge exist to tackle crosstalk problem in I/O bus encodings [38][39][40][41][42][43] . One can learn from such work to address PCM thermal crosstalk problem.…”

A design methodology and various performance and fabrication metrics evaluation of 3D Network-on-Chip with multiplexed Through-Silicon Vias

“…Each of the various low-power-coding techniques [35] that geared towards minimizing the number of transitions and hence power consumption can be applied here. We expect that applying such technique will reduce the power consumption of the TSVBOX more and makes it too close or even lower in consumption than the conventional case.…”

Design of Smart Power-Saving Architecture for Network on Chip

“…{select the channel to output} (6) else if ( ℎ 2 is tail and = ℎ 2 ) (7) {clear the channel and select the channel to output;} (8) else (9) {read back flit to virtual channel} (10) end while Algorithm 1: Channel sign up algorithm. (5) if (global) (6) { = global input virtual channel address} (7) end while / * Channel switch * / (8) Case (9) { 1 = local packet of } (10) end case (11) Case (12) { = global packet of } (13) end case Algorithm 2: VC arbitration algorithm.…”

“…Output: VCs clock (1) VC group is VCs group of 4 direction port (2) VC flag includes full flag and empty flag (3) Initial VCs clock and access VCs count and stage flag (4) follow LCR to arrangement all slots priority; (5) VC is VCs clock of each VC group //where 1 ≤ ≤ (6) Example VC group = East port (7) initial VC = 0; //where 1 ≤ ≤ (8) while (virtual channel be write) do (9) if (VC flag = empty) (10) {VC = system clock} (11) If (VC flag = full flag) (12) {VC = 0 and VC +1 = system clock} (13) end while (14) while (virtual channel be read) do (15) if ( with VC flags. The slots priority from LCR [2] and each VCs clock can be initialized at lines 4, 5, and 7.…”

“…Palma et al [8] use T-Bus-Invert technology to reduce the hamming distance transition activity rate to improve the power consumption. Jafarzadeh et al [9] use end-to-end data coding technology to minimize switching activity rate and routing path to improve NI power consumption. Lee et al [10] proposed buffer clock-gating architecture and used clock-gating to reduce the transmit power consumption when slots are empty and full.…”