The SPARC64 V microprocessor is designed for use in high-reliability, large-scale unix servers. In addition to implementing ECC for large SRAM arrays, the SPARC64 V microprocessor incorporates error detection and recovery mechanisms for processor logic circuits and smaller SRAM arrays. The effectiveness of these error recovery mechanisms was validated via accelerated neutron testing of Fujitsu’s commercial unix server, the PRIMEPOWER 650. Soft errors generated in SRAM arrays were completely recovered by the implemented hardware mechanisms, and only 6.4% of the estimated neutron-induced logic circuit faults manifested as errors, 76% of which were recovered by hardware. From these tests, the soft error failure rate of the SPARC64 V microprocessor due to atmospheric neutron hits was confirmed to be well below 10 FIT.

To predict application performance on an HPC system is an important technology for designing the computing system and developing applications. However, accurate prediction is a challenge, particularly, in the case of a future coming system with higher performance. In this paper, we present a new method for predicting application performance on HPC systems. This method combines modeling of sequential performance on a single processor and macro-level simulations of applications for parallel performance on the entire system. In the simulation, the execution flow is traced but kernel computations are omitted for reducing the execution time. Validation on a real terascale system showed that the predicted and measured performance agreed within 10% to 20 %. We employed the method in designing a hypothetical petascale system of 32768 SIMD-extended processor cores. For predicting application performance on the petascale system, the macro-level simulation required several hours.

Current high-end microprocessor designs focus on increasing instruction parallelism and clock frequency at the expense of power dissipation. This paper presents a case study of a different direction, a chip multiprocessor (CMP) with a smaller processor core than a baseline high-end 130-nm 64-bit SPARC server uniprocessor. We demonstrate that the size of the baseline processor core can be reduced by 2/3 using a combination of logical resource reduction and dense custom macros while still delivering about 70% of the TPC-C performance. Circuit speed is traded for power reduction by reducing the power supply from 1.0 to 0.8 V and increasing transistor channel lengths by 12.5% above the minimum. The resulting CMP with six reduced size cores and 4-MB L2 cache is estimated to run at 1.8 GHz while consuming less than 30% of the power compared to the scaled baseline dual-core processor running at 2.4 GHz. The proposed CMP is more than four times higher in TPC/W than the dual-core processor, facilitating the design of high-density servers.

A 5th generation SPARC64 processor is fabricated in 130nm SOI CMOS process with 8 layers of Cu metallization. It runs at 1.3GHz with 34.7W power dissipation in the laboratory. The chip contains over 190M transistors with 19M in logic circuits. The chip size is 18.14mm x 15.99mm. The error detection and recovery mechanism is implemented for execution units and data path logic circuits in addition to on-chip arrays to detect and recover from data logic error. This processor is developed by using mostly in-house CAD tools.