FPGA Prototyping of Emerging Manycore Architectures for Parallel Programming Research using Formic Boards

Performance evaluation of parallel software and architectural exploration of innovative hardware support face a common challenge with emerging manycore platforms: they are limited by the slow running time and the low accuracy of software simulators. Manycore FPGA prototypes are difficult to build, but they offer great rewards. Software running on such prototypes runs orders of magnitude faster than current simulators. Moreover, researchers gain significant architectural insight during the modeling process. We use the Formic FPGA prototyping board [1], which specifically targets scalable and cost-efficient multi-board prototyping, to build and test a 64-board model of a 512-core, MicroBlaze-based, non-coherent hardware prototype with a full network-on-chip in a 3D-mesh topology. We expand the hardware architecture to include the ARM Versatile Express platforms and build a 520-core heterogeneous prototype of 8 Cortex-A9 cores and 512 MicroBlaze cores. We then develop an MPI library for the prototype and evaluate it extensively using several bare-metal and MPI benchmarks. We find that our processor prototype is highly scalable, models faithfully single-chip multicore architectures, and is a very efficient platform for parallel programming research, being 50,000 times faster than software simulation.

A smart load appliance application using a single compressor fridge-freezer to support grid operations

This paper describes a fridge-freezer smart load model, which responds to external signals from the wholesale electricity market to support grid operations while switching the fridge-freezer on and off to maintain optimum operations for the owner. The key parameters of the model are the appliance dimensions, thermal mass, the fridge and freezer thermal time constants and the compressor power consumption. The model demonstrates that control strategies help to minimise load at times when the grid is under stress from high demand, and shift some load to a lower wholesale price or when there is excess renewable power. Three control strategies are proposed, based on peak shaving and valley filling, price signals and wind availability.

Single vocal cord irradiation:a competitive treatment strategy in early glottic cancer

INTRODUCTION: The treatment of choice for early glottic cancer is still being debated; ultimately it relies on the functional outcome. This paper reports on a novel sparing 4D conformal technique for single vocal cord irradiation (SVCI).

MATERIAL AND METHODS: The records of 164 T1a patients with SCC of the vocal cord, irradiated in the Erasmus MC between 2000 and 2008, were analyzed for local control and overall survival. The quality of life was determined by EORTC H&N35 questionnaires. Also the VHI (voice handicap index), and the TSH (thyroid stimulating hormone) blood levels, were established. On-line image guided SVCI, using cone beam CT or stereotactic radiation therapy (SRT) techniques, were developed.

RESULTS: A LC rate at five-years of 93% and a VHI of 12.7 (0-63) was determined. It appeared feasible to irradiate one vocal cord within 1-2mm accuracy. This way sparing of the contralateral (CL) vocal cord and CL normal tissues, could be achieved.

CONCLUSIONS: Given the accuracy (1-2mm) and small volume disease (CTV limited to one vocal cord), for the use of stereotactic RT techniques SVCI with large fraction sizes is currently being investigated in clinic. It is argued that hypofractionated SVCI can be a competitive alternative to laser surgery.

IMRT for image-guided single vocal cord irradiation

PURPOSE: We have been developing an image-guided single vocal cord irradiation technique to treat patients with stage T1a glottic carcinoma. In the present study, we compared the dose coverage to the affected vocal cord and the dose delivered to the organs at risk using conventional, intensity-modulated radiotherapy (IMRT) coplanar, and IMRT non-coplanar techniques.

METHODS AND MATERIALS: For 10 patients, conventional treatment plans using two laterally opposed wedged 6-MV photon beams were calculated in XiO (Elekta-CMS treatment planning system). An in-house IMRT/beam angle optimization algorithm was used to obtain the coplanar and non-coplanar optimized beam angles. Using these angles, the IMRT plans were generated in Monaco (IMRT treatment planning system, Elekta-CMS) with the implemented Monte Carlo dose calculation algorithm. The organs at risk included the contralateral vocal cord, arytenoids, swallowing muscles, carotid arteries, and spinal cord. The prescription dose was 66 Gy in 33 fractions.

RESULTS: For the conventional plans and coplanar and non-coplanar IMRT plans, the population-averaged mean dose ± standard deviation to the planning target volume was 67 ± 1 Gy. The contralateral vocal cord dose was reduced from 66 ± 1 Gy in the conventional plans to 39 ± 8 Gy and 36 ± 6 Gy in the coplanar and non-coplanar IMRT plans, respectively. IMRT consistently reduced the doses to the other organs at risk.

CONCLUSIONS: Single vocal cord irradiation with IMRT resulted in good target coverage and provided significant sparing of the critical structures. This has the potential to improve the quality-of-life outcomes after RT and maintain the same local control rates.

Predictive Factors for Local Control in Primary and Metastatic Lung Tumours after Four to Five Fraction Stereotactic Ablative Body Radiotherapy: A Single Institution's Comprehensive Experience

AIMS: We report the outcomes of a large lung stereotactic ablative body radiotherapy (SABR) programme for primary non-small cell lung cancer (NSCLC) and pulmonary metastases. The primary study aim was to identify factors predictive for local control.

MATERIALS AND METHODS: In total, 311 pulmonary tumours in 254 patients were treated between 2008 and 2011 with SABR using 48-60 Gy in four to five fractions. Local, regional and distant failure data were collected prospectively, whereas other end points were collected retrospectively. Potential clinical and dosimetric predictors of local control were evaluated using univariate and multivariate analyses.

RESULTS: Of the 311 tumours, 240 were NSCLC and 71 were other histologies. The 2 year local control rate was 96% in stage I NSCLC, 76% in colorectal cancer (CRC) metastases and 91% in non-lung/non-CRC metastases. Predictors of better local control on multivariate analysis were non-CRC tumours and a larger proportion of the planning target volume (PTV) receiving ≥100% of the prescribed dose (higher PTV V100). Among the 45 CRC metastases, a higher PTV V100 and previous chemotherapy predicted for better local control.

CONCLUSIONS: Lung SABR of 48-60 Gy/four to five fractions resulted in high local control rates for all tumours except CRC metastases. Covering more of the PTV with the prescription dose (a higher PTV V100) also resulted in superior local control.

WHIRLBOB, the Whirlpool Based Variant of STRIBOB: Lighter, Faster, and Constant Time

WHIRLBOB, also known as STRIBOBr2, is an AEAD (Authenticated Encryption with Associated Data) algorithm derived from STRIBOBr1 and the Whirlpool hash algorithm. WHIRLBOB/STRIBOBr2 is a second round candidate in the CAESAR competition. As with STRIBOBr1, the reduced-size Sponge design has a strong provable security link with a standardized hash algorithm. The new design utilizes only the LPS or ρ component of Whirlpool in flexibly domain-separated BLNK Sponge mode. The number of rounds is increased from 10 to 12 as a countermeasure against Rebound Distinguishing attacks. The 8 ×8 - bit S-Box used by Whirlpool and WHIRLBOB is constructed from 4 ×4 - bit “MiniBoxes”. We report on fast constant-time Intel SSSE3 and ARM NEON SIMD WHIRLBOB implementations that keep full miniboxes in registers and access them via SIMD shuffles. This is an efficient countermeasure against AES-style cache timing side-channel attacks. Another main advantage of WHIRLBOB over STRIBOBr1 (and most other AEADs) is its greatly reduced implementation footprint on lightweight platforms. On many lower-end microcontrollers the total software footprint of π+BLNK = WHIRLBOB AEAD is less than half a kilobyte. We also report an FPGA implementation that requires 4,946 logic units for a single round of WHIRLBOB, which compares favorably to 7,972 required for Keccak / Keyak on the same target platform. The relatively small S-Box gate count also enables efficient 64-bit bitsliced straight-line implementations. We finally present some discussion and analysis on the relationships between WHIRLBOB, Whirlpool, the Russian GOST Streebog hash, and the recent draft Russian Encryption Standard Kuznyechik.