Adaptive multicore scheduling for the LTE uplink

Abstract

The next generation cellular system of 3GPP is named Long Term Evolution (LTE). Each millisecond, a LTE base station receives information from up to one hundred users. Multicore heterogeneous embedded systems with Digital Signal Processors (DSP) and coprocessors are power efficient solutions to decode the LTE uplink signals in base stations. The LTE uplink is a highly variable algorithm. Its multicore scheduling must be adapted every millisecond to the number of connected users and to the data rate they require. To solve the issue of the dynamic deployment while maintaining low latency, one approach would be to find efficient on-the-fly solutions using techniques such as graph generation and scheduling. This approach is opposed to a static scheduling of predefined cases. We show that the static approach is not suitable for the LTE uplink and that present DSP cores are powerful enough to recompute an efficient adaptive schedule for the LTE uplink most complex cases in real-time.

Applying the adaptive Hybrid Flow-Shop scheduling method to schedule a 3GPP LTE physical layer algorithm onto many-core digital signal processors

“…The scheduling of the 3GPP LTE PUSCH decoding process has previously been discussed in the work of Pelcat et al [5]. In this work it was shown that static scheduling of PUSCH is unfeasible and that an adaptive scheduler is necessary.…”

“…In this work it was shown that static scheduling of PUSCH is unfeasible and that an adaptive scheduler is necessary. The applied list scheduling method provided very high-quality schedules, but the context of [5] was different to this work, as the assumed platform was isotropic and the resulting communications were not restricted in their direction.…”

“…For these operations, the parameters are highly variable during runtime (number of connected UEs, number of allocated Code Blocks (CBs), modulation order, and so on). The multi-core scheduling of this dynamic part must be adaptive [5]. A detailed graph description of this part is shown in Section III-B.…”

“…It has been shown previously [5] that the highly dynamic nature of the Uplink Physical Layer data processing (PUSCH) decoding algorithm executed in 3rd Generation Partnership Project (3GPP) Long Term Evolution (LTE) base stations does not allow the use of static multiprocessor schedules; fast adaptive scheduling is mandatory. [5] used a list scheduling method to schedule 3GPP LTE PUSCH decoding on a multicore DSP.…”

“…[5] used a list scheduling method to schedule 3GPP LTE PUSCH decoding on a multicore DSP. This method is not applicable to many-core adaptive scheduling (with hundreds of PEs), as 1/9 of the computation is dedicated to scheduling which involves 8 PEs.…”