A Scalable Low Power Issue Queue for Large Instruction Window Processors

Large instruction windows and issue queues are key to exploiting greater instruction level parallelism in out-of-order superscalar processors. However, the cycle time and energy consumption of conventional large monolithic issue queues are high. Previous efforts to reduce cycle time segment the issue queue and pipeline wakeup. Unfortunately, this results in significant IPC loss. Other proposals which address energy efficiency issues by avoiding only the unnecessary tag-comparisons do not reduce broadcasts. These schemes also increase the issue latency.To address both these issues comprehensively, we propose the Scalable Lowpower Issue Queue (SLIQ). SLIQ augments a pipelined issue queue with direct indexing to mitigate the problem of delayed wakeups while reducing the cycle time. Also, the SLIQ design naturally leads to significant energy savings by reducing both the number of tag broadcasts and comparisons required.A 2 segment SLIQ incurs an average IPC loss of 0.2% over the entire SPEC CPU2000 suite, while achieving a 25.2% reduction in issue latency when compared to a monolithic 128-entry issue queue for an 8-wide superscalar processor. An 8 segment SLIQ improves scalability by reducing the issue latency by 38.3% while incurring an IPC loss of only 2.3%. Further, the 8 segment SLIQ significantly reduces the energy consumption and energy-delay product by 48.3% and 67.4% respectively on average.

A Predictive Perfomance Model for Superscalar Processors

Designing and optimizing high performance microprocessors is an increasingly difficult task due to the size and complexity of the processor design space, high cost of detailed simulation and several constraints that a processor design must satisfy. In this paper, we propose the use of empirical non-linear modeling techniques to assist processor architects in making design decisions and resolving complex trade-offs. We propose a procedure for building accurate non-linear models that consists of the following steps: (i) selection of a small set of representative design points spread across processor design space using latin hypercube sampling, (ii) obtaining performance measures at the selected design points using detailed simulation, (iii) building non-linear models for performance using the function approximation capabilities of radial basis function networks, and (iv) validating the models using an independently and randomly generated set of design points. We evaluate our model building procedure by constructing non-linear performance models for programs from the SPEC CPU2000 benchmark suite with a microarchitectural design space that consists of 9 key parameters. Our results show that the models, built using a relatively small number of simulations, achieve high prediction accuracy (only 2.8% error in CPI estimates on average) across a large processor design space. Our models can potentially replace detailed simulation for common tasks such as the analysis of key microarchitectural trends or searches for optimal processor design points.

Interdiffusion and the vacancy wind effect in Ni-Pt and Co-Pt systems

In this study, bulk and multifoil diffusion couple experiments were conducted to examine the interdiffusion process in Ni-Pt and Co-Pt binary alloy systems. Inter-, intrinsic-, and tracer-diffusion coefficients at different temperatures, and as a function of the composition, were estimated by using the experimental data. Results show that in both the alloy systems, Pt is the slower diffusing species, and hence the interdiffusion process is controlled by either Ni or Co. The thermodynamic driving force makes the intrinsic diffusion coefficients of Co and Ni higher in the range of 30-70 at.%. The low activation energy for Co and Ni impurity diffusion in Pt compared with Pt in Ni and Co indicates that the size of the atoms plays an important role. The vacancy wind effects on the diffusion process are examined in detail, and it was demonstrated that its contribution falls within the experimental scatter and hence can be neglected.

Ion exchange equilibria between (Mn,Co)O solid solution and (Mn,Co)Cr2O4 and (Mn,Co)Al2O4 spinel solid solutions at 1100oC

The compositions of the (Mn,Co)O solid solution with rock salt structure in equilibrium with (Mn,Co)Cr2O4 and (Mn,Co)Al2O4 spinel solid solutions have been determined by X-ray diffraction measurements at 1100° C and an oxygen partial pressure of 10–10 atm. The ion exchange equilibria are quantitatively analysed, using values for activities in the (Mn,Co)O solid solution available in the literature, in order to obtain activities in the spinel solid solutions. The MnAl2O4-CoAl2O4 solid solution exhibits negative deviations from Raoult's law, consistent with the estimated cation disorder in the solid solution, while the MnCr2O4-CoCr2O4 solid solution shows slightly positive deviations. The difference in the Gibbs free energy of formation of the two pure chromites and aluminates derived from the results of this study are in good agreement with recent results obtained from solid oxide galvanic cells and gas-equilibrium techniques.

Three phase equilibria for activity determinations: system Co Mn

A solid state e.m.f. cell with yttria-doped thoria as the electrolyte and a mixture of Cr + Cr sub 2 O sub 3 as the reference electrode, was employed for the measurement of the activity of manganese in the Co--Mn system at 1760 deg K, for 0.3 > X sub Mn > 0.05. The liquid alloy was contained in an alumina crucible and saturated with MnAl sub 2+2x O sub 4+3x . The cell can be represented by Pt, W, (Co--Mn) + MnAl sub 2+2x O sub 4+3x + Al sub 2 O sub 3 /ThO sub 2 --Y sub 2 O sub 3 /Cr + Cr sub 2 O sub 3 , Pt. The activity of manganese shows negative deviations from Raoult's law. The activities in the Co--Mn system are intermediate between those in the Fe--Mn and Ni--Mn systems. The Gibbs' energies of mixing in these systems follow the trends given by Miedema's model. 14 ref.--AA.

The Co Cr S ternary diagram at 1223 K and applications to corrosion

Ternary phase relations in the Co-Cr-S system at 1223 K were determined using microprobe analysis of quenched samples. The results are consistent with the data available on the binary systems. A complete solid solution exists between cobalt monosulfide and chromium monosulfide. The CoCr2S4 thiospinel is the only ternary compound formed. A sulfur potential diagram was constructed for the region involving equilibrium between alloy and monosulfide based on thermodynamic data on the Co-Cr, Co-S, and Cr-S binary systems and the ternary information obtained in this study. The sulfidation behavior of Co-Cr alloys reported in the literature is discussed in light of the sulfur potential diagram.

Thermodynamic properties and phase boundaries of Co O solutions

The thermodynamic properties of liquid unsaturated Co--O solutions have been determined by electrochemical measurements using (Y sub 2 O sub 3 )ThO sub 2 as solid electrolyte. The cell can be represented as, Pt. MoO sub 2 + Mo | (Y sub 2 O sub 3 )ThO sub 2 | O sub Co , tungsten, Pt, Emf of the cell was measured as a function of oxygen concentration in liquid Co at 1798, 1873 and 1973K. Least-mean squares regression analysis of the experimental data gives for the free energy of solution of diatomic oxygen in liquid Co Delta G exp 0 sub O(Co) = --84935--7.61 T ( plus/minus 400) J/g-atom and self interaction parameter for oxygen epsilon exp O sub O = --97240/T + 40.52 ( plus/minus 1) where the standard state for O is an infinitely dilute solution in which the activity is equal to atomic percent. The present data are discussed in comparison with those reported in the literature and the phase diagram for the Co--O system. 18 ref.--AA.

Phase relations and activities in the Co Ni O system at 1373 K

The tie-lines delineating equilibria between CoO-NiO and Co-Ni solid solutions in the ternary Co-Ni-O system at 1373 K have been determined by electron microprobe andedax point count analysis of the oxide phase equilibrated with the alloy. The oxygen potentials corresponding to the tie-line compositions have been measured using a solid oxide galvanic cell with calcia-stabilized zirconia electrolyte and Ni + NiO reference electrode. Activities in the metallic and oxide solid solution have been derived using a new Gibbs-Duhem integration technique. Both phases exhibit small positive deviations from ideality; the values ofG E/X 1 X 2 are 2640 J mol−1 for the metallic phase and 2870 J mol−1 for the oxide solid solution.

CO oxidation by CeO(2)-Al(2)O(3)-CeAlO(3) hybrid oxides

A modified solution combustion technique was successfully used to synthesize sub-10 nm crystallites of hybrid CeO(2)-Al(2)O(3)-CeAlO(3). The fuel in the solution combustion was tuned to obtain mixed oxides and solid solutions of the compound. The compounds were characterized by X-ray diffraction, transmission electron microscopy and X-ray photoelectron spectroscopy. XRD and TEM analysis showed the substitution of Al(3+) ions in the CeO(2) matrix when a combination of glycine, urea, hexamine and oxalyl dihydrazide was used as fuel for the synthesis. The compounds showed high activity for CO oxidation and the activity of the compounds was dependent upon the composition of the oxide.

Phase equilibria in the Co Ni F system and activities in (CoxNi1 x)F2 solid solutions at 1373 K

The tie-lines delineating equilibria between CoF2-NiF2 and Co-Ni solid solutions in the ternary Co-Ni-F system at 1373 K have been determined by electron microprobe and EDAX point count analysis of the equilibrated phases. Activities in the fluoride solid solution have been derived from the knowledge of activitycomposition relation in the metallic solid solution and tie-line data,using a modified form of the Gibbs-Duhem integration. The fluorine potentials corresponding to the tie-line compositions have been calculated.The excess Gibbs' energy of mixing for the fluoride solid solution derived from the present data can be represented by the expression

La-Doped Ba2In2O5 Electrolyte: Pechini Synthesis, Microstructure, Electrical Conductivity, and Application for CO Gas Sensing

Dense (Ba1―xLax)2In2O5+x (BLIO) electrolytes with different compositions (x = 0.4, 0.5, 0.6) were fabricated using powders obtained by the Pechini method. The formation of BLIO powders was investigated by using X-ray diffraction and scanning electron microscopy coupled with energy-dispersive X-ray spectroscopy. The calcination temperature and time were optimized. The sintered (Ba1―xLax)2In2O5+x electrolytes showed a relative density greater than ∼97%, and the major phase of three electrolyte compositions was indexed as a cubic perovskite. The electrical conductivity of BLIO ceramics at elevated temperatures in air was measured by ac-impedance spectroscopy. The activation energies for conduction in BLIO were 102 kJ mol―1 between 473 and 666 K and 118 kJ mol―1 between 769 and 873 K, which are comparable to that for 8 mol % yttria-stabilized cubic zirconia. Mixed-potential gas sensors utilizing BLIO-based electrolytes exhibited good sensitivity to different CO concentrations from ∼100 to ∼500 ppm and excellent selectivity to methane at around 873 K.