Calibrating a predictive cache emulator for SoC design

Abstract

Pre-fetching in a memory hierarchy is known to alleviate the "memory wall" paradigm but its use is impeded because of the difficulty to estimate efficiency when used in a complex system such as a SoC (System on Chip) or NoC (Network on Chip). Therefore, some methods are needed to evaluate the benefit of pre-fetching at the earliest possible stage in a design flow to help the designer choose architectural parameters or transform the application algorithm.In this paper we show that the emulation platform implementing the nD-AP Cache (n-Dimensional Adaptive and Predictive Cache) allows to perform a platformindependent measurement of this cache efficiency. The nD-AP Cache performs pre-fetching in multidimensional arrays which are commonly used in image processing and multimedia applications. The obtained metric can be used to extrapolate the cache performance in a much broader system configuration. The method to compute this metric is the calibration process .The performed benchmarks show that the calibration process is confident. Also, we measured that the nD-AP Cache is two times faster than a standard PowerPC 2-way set associative cache in the context of an image processing kernel.

Exploration of 3D grid caching strategies for ray-shooting

“…Also, we make the assumption that the processing unit is able to perform one fetch each clock cycle and sometimes waits for a datum when it is not present in the cache. In [20], we have shown that the performances measured on an emulation platform made of a PowerPC connected to a 2D-AP Cache 3 fit the ones obtained by simulation. More precisely the cache efficiency can be normalised by the ratio between the fetch period and the background memory latency.…”

“…The 3D-AP Cache may be linked to a classic general purpose processor (or a DSP) or to an hardware accelerator dedicated to ray-shooting. We have shown in [20] that a 2D-AP Cache connected to a processor allows better performances than a standard cache. Performing ray-shooting with programmable processors connected to 3D-AP Cache should still benefit of the pre-fetching mechanisms.…”