In a weblog submit, AMD’s Executive Vice President and Chief Technology Officer, Mark Papermaster introduced that AMD is formally becoming a member of Compute Express Link (CXL) Consortium. Previously AMD has joined totally different working teams which are addressing the issue of cache coherency inside numerous methods and accelerators, like CCIX, OpenCAPI and Gen-Z. CXL is initiative began by Intel that guarantees higher resolution to cache coherency downside and plenty of firms have joined, now together with AMD.
As workloads are cut up into totally different segments like machine studying and graphics processing that require a mixture of scalar and vector processing, accelerators for these workloads want higher inter connection to different processors similar to CPUs, the place coherency and reminiscence semantics are one of many largest challenges. AMD’s aim of becoming a member of the CXL working group is to assist advance the usual and probably combine it into future merchandise. making the idea of heterogeneous computing a bit nearer to actuality.
