Results 1 - 10 of 17,942

Table 1. Area and delay of functional units for the proposed extensions as well as of the extended integer unit

in Instruction set extensions for efficient AES implementation on 32-bit processors
by Stefan Tillich, Johann Großschädl 2006
"... In PAGE 8: ...hese synthesis results include the complete area overhead of the extensions, e.g. new functional units, decoding of additional opcodes. The results are given in Table1 . Note that sbox4s indicates the three instructions sbox4s, isbox4s and sbox4r and that mixcol4s stands for the instructions mixcol4s and imixcol4s.... In PAGE 8: ... The MixColumns multiplier requires little area and has a shorter critical path than the S-boxes. The results in Table1 for the integer unit use the approach of Canright [4] for S-box extensions. Area overhead is calculated in relation to an integer unit without extensions and ranges between a factor of 1.... ..."
Cited by 1

TABLE 3.3: Architectural metrics for SPECInt95 with and without the operating system for both SMT and the superscalar. The maximum issue for integer programs is 6 instructions on the 8-wide SMT, because there are only 6 integer units.

in An Analysis of Software Interface Issues for SMT Processors
by Joshua Abram Redstone 2002

Table 2: Cycle count comparison of different types of machines. Floating point unit (FPU) utilization is givenas the averagenumber of floating point operations executed each cycle. Integer Unit (IU) utilization is calculated similarly.

in unknown title
by unknown authors 1992
"... In PAGE 5: ... Sequential mode operation provides an upper bound since only the parallelism within a single cluster can be exploited. Table2 shows the cycle counts for each machine mode. Floating point utilization is givenasthe averagenumberof floatingpoint operationsexecuted each cycle.... ..."
Cited by 84

Table 2: Cycle count comparison of different types of machines. Floating point unit (FPU) utilization is givenas the average number of floating point operations executed each cycle. Integer Unit (IU) utilization is calculated similarly.

in unknown title
by unknown authors 1992
"... In PAGE 5: ... Sequential mode operation provides an upper bound since only the parallelism within a single cluster can be exploited. Table2 shows the cycle counts for each machine mode. Floating point utilization is givenas the averagenumberof floatingpoint operations executed each cycle.... ..."
Cited by 84

Table 2: Results averaged over integer benchmarks for di erent integer function units using various techniques to exploit frequent-valued operands.

in A Detailed Study of Hardware Techniques That Dynamically Exploit Frequent Operands to Reduce Power Consumption in Integer Function Units
by Kaushal R. Gandhi, Kaushal R. G, Nihar R. Mahapatra
"... In PAGE 8: ... most number of cases of frequent valued operands at a ner granularity. Table2 tabulates the results. Our results show that power can be reduced by 10-15% for addition, comparison, and multiplication.... ..."

Table 4.1: SGI R5000 Integer addition. In the second table we can see that every two cycles it is started an operation, that is, the integer unit is a two stages integer unit. In the rst table we can see how the performance are increased by the reorganization of the code by the compiler when the instruction are independent. We can see that the architecture is not completely utilized. PIPELINE 1 2 3 4 5 6

in Performance Evaluation of Data Locality Exploitation
by Paolo D'Alberto, Mura Anteo Zamboni

Table 1. Unit-Integer rotation angles. N-1 Angle 1-N Angle

in Fast and Efficient Rotated Haar-like Features using Rotated Integral Images
by Chris Messom, Andre Barczak

Table 3.1: SEND Instructions The isnd instructions are used in the integer unit, while the fsnd instructions are for the oating-point unit. Only the isnd0 and isnd0o instructions and their oating- point unit counterparts are accessible to user-level threads. While in-order delivery 40

in Mechanisms for Efficient, Protected Messaging
by Whay Sing Lee

Table 6: Architectural metrics comparing Apache executing on an SMT to SPECInt95 on SMT and Apache on a superscalar. All applications are executing with the operating system. The maximum issue for integer programs is 6 instructions on the 8- wide SMT, because there are only 6 integer units.

in An Analysis of Operating System Behavior on a Simultaneous Multithreaded Architecture
by Joshua A. Redstone, Susan J. Eggers, Henry M. Levy 2000
"... In PAGE 8: ...ccess operations for Apache bypass the TLB, i.e., they specify physical addresses directly. Table6 shows architectural performance characteristics for Apache and compares them to the SPECInt workload in steady state. The chart also shows statistics for Apache running on a superscalar.... ..."
Cited by 35

Table 6: Architectural metrics comparing Apache executing on an SMT to SPECInt95 on SMT and Apache on a superscalar. All applications are executing with the operating system. The maximum issue for integer programs is 6 instructions on the 8- wide SMT, because there are only 6 integer units.

in An analysis of operating system behavior on a simultaneous multithreaded architecture
by Joshua A. Redstone, Susan J. Eggers, Henry M. Levy 2000
"... In PAGE 8: ...ccess operations for Apache bypass the TLB, i.e., they specify physical addresses directly. Table6 shows architectural performance characteristics for Apache and compares them to the SPECInt workload in steady state. The chart also shows statistics for Apache running on a superscalar.... ..."
Cited by 35
Results 1 - 10 of 17,942