In this section, we describe our heterogeneous LLC management mechanism that mitigates the performance impact of LLC sharing by throttling LLC accesses initiated by the GPU cores. HeLM exploits the memory access latency tolerance capability of the GPU cores and allows the GPU cores to yield LLC space to the cache sensitive CPU cores without significantly degrading their own performance. In HeLM, we manage the LLC occupancy of the GPU cores by allowing the GPU memory traffic to selectively bypass the LLC when: i) the GPU cores exhibit sufficient TLP to tolerate memory access latency; or ii) when the GPU application is not sensitive to LLC performance.