Last week I did one of my occasional CPU family string searches and found two Geekbench results of Carrizo. They're clearly labeled "FX-8800P" and "A10-8700P". High view counts suggest some tech forum activity around these results. Due to seemingly active power management the listed clocks don't tell us much about the true clock frequency range.
The Aspire E5 series seems to be designed for 15-19W TDP, which makes it likely that those models tested here also fit into this TDP range. Some Carrizo Sandra results also show a TDP of 15W.
The leaked 3D Mark result of a "FX-8800P" lists a 2.1GHz base clock and a 1.7GHz "turbo". The 1.7GHz appears again in the Geekbench 3 result. Using the Carrizo ES Sandra results listed in the table below, we might speculate, that the first entry belongs to the FX-8800P. The A10-8700P might correlate with the lower 3 ES results (likely one system reported with different clocks).
To compare these two Carrizo APUs to some existing Kaveri models at roughly similar TDP ratings I used some Geekbench3 results of A10-7300 and FX-7500 based notebooks. Picking the right models as tested under unknown conditions (power management settings, etc.) is difficult. So I just took the best results (all using the 32 bit client) with a single core result >= 1000 and put them into some score sorted charts. The dots represent single systems while the horizontal lines show the single Carrizo based results.
While the A10-8700P really shines, the FX-8800P doesn't by plainly looking at the scores. But the TDP of these models while containing the FCH is likely 15W, which is 5W less than the 20W TDPs of the two Kaveri APUs. From user "Abwx" at Anandtech forums I also learned that the FX based system used an older BIOS.
Edit: Graphs exchanged (twice) and some explainations added.