If either of the above were causing the performance problem (and the scheduling hack is highly suspect), this should have brought the 7950x3d to parity with the 7800x3d (higher really, since it can boost 200MHz higher). The simulation test disabling the non-vcache ccd should have addressed both of these, sharing power only on the 8 enabled cores and no need for AMD's hack. Second, AMD's software hack to address scheduling for the heterogeneous cores. First, they have the same power limit of 120W, while the 7950x3d has to share that amongst 16 cores instead of just 8. There are two limiting factors that affect the 7950x3d over the 7800x3d. There's no fundamental reason the 7950x3d should be slower than the 7800x3d, especially as much slower (in games) as this review is showing. Click to expand.I'd really like to see more investigation into this.
0 kommentar(er)
