DocumentCode
3195134
Title
1 GHz HAL SPARC64R Dual Floating Point Unit with RAS features
Author
Naini, Ajay ; Dhablania, Atul ; James, Warren ; Sarma, Debjit Das
Author_Institution
HAL Comput. Syst., Campbell, CA, USA
fYear
2001
fDate
2001
Firstpage
173
Lastpage
183
Abstract
An IEEE compliant, 1 GHz Sparc64-V Floating-Point Unit (FPU) with reliability-accessibility-serviceability (RAS) features and partial support for denormal operands and results is presented. The FPU has two functional units, each with an adder (FADD) and a multiplier (FMUL). Additionally, one of the functional units contains a graphics unit (VIS). Two floating-point instructions can be scheduled out of order each cycle, one to each functional unit. A peak performance of 4 GFLOP is achieved by scheduling two floating-point multiply add (FMA) instructions each cycle. The FADD unit is fully pipelined and can execute an addition, subtraction, conversion, or compare instruction every cycle. The FMUL unit executes pipelined multiply instructions. Divide and square-root instructions are executed with multiple iterations through the multiplier pipeline. The VIS unit is also pipelined and executes SIMD fixed-point graphics instructions. The adder and multiplier have latencies of 3 and 4 cycles respectively. Novel techniques are presented in the adder and multiplier implementations to reduce area and cycle time. The FPU provides RAS support for enhanced server reliability by using selective parity error detection. The FPU has been implemented in 0.15 u, 6-layer metal CMOS technology
Keywords
adders; fault tolerant computing; floating point arithmetic; pipeline arithmetic; processor scheduling; 1 GHz; IEEE compliant HAL Sparc64-V Floating-Point Unit; SIMD fixed-point graphics instructions; adder; compare instructions; conversion instructions; denormal operands; divide and square-root instructions; floating-point instructions; floating-point multiply add instructions; graphics unit; latency; metal CMOS technology; multiple iterations; multiplier; multiplier pipeline; reliability-accessibility-serviceability features; scheduling; selective parity error detection; server reliability; square-root instructions; subtraction instructions; CMOS process; CMOS technology; Clocks; Delay; Graphics; Large-scale systems; Microprocessors; Out of order; Pipelines; USA Councils;
fLanguage
English
Publisher
ieee
Conference_Titel
Computer Arithmetic, 2001. Proceedings. 15th IEEE Symposium on
Conference_Location
Vail, CO
ISSN
1063-6889
Print_ISBN
0-7695-1150-3
Type
conf
DOI
10.1109/ARITH.2001.930117
Filename
930117
Link To Document