Comparisons of AMD Microarchitectures
The power efficiency of the CPU / APUs were tested at stock speed and at about 2 GHz.
At stock speed, the aging Athlon 5050e was less power efficient than A8-3800 across nearly all scenarios except in BIOS while other platforms used UEFI. A10-5700 was rated at 65W Thermal Design Power (TDP) but when the CPU cores were loaded, it easily exceeded the 65W TDP target. Athlon 5350 drew the least power from socket because it was a low power design.
At about 2 GHz, A10-5700, A8-3800 tied in power consumption, but A8-3800 was faster in nearly all scenario at this clock speed. Athlon 5050e was not power efficient compared to newer APUs.
AMD’s mainstream processors have been disappointing most people in the recent years due to minimal processing power gained over the years. The results show that while A10-5700 was faster than A8-3800, it drew significantly more power with lower power efficiency. For some applications, A10-5700 even did not held advantages over Athlon 5050e for single-thread applications. It could be the reason that AMD appeared not to improve the microarchitecture after 2015.
Despite being the low power microarchitecture, Jaguar showed promise as it was power efficient. With higher clock speed and better power management, Jaguar could even surpass K10 in performance. If AMD was given the time to do so…
Link for discussion: http://bbs.hk-spot.com/viewthread.php?tid=92106 (Chinese, but English welcomed)