| | | | | | |

A10 9630P vs Engineering Sample 506E8


Description
The 9630P is based on Excavator architecture while the 506E8 is based on Kaby Lake.

Using the multithread performance as a reference, the 9630P gets a score of 32.9 k points while the 506E8 gets 211.6 k points.

Summarizing, the 506E8 is 6.4 times faster than the 9630P. To get a proper comparison between both models, take a look to the data shown below.

Specs
CPUID
660f51
506e8
Core
Bristol Ridge
Kaby Lake
Architecture
Base frecuency
2.6 GHz
2.4 GHz
Boost frecuency
3.3 GHz
3.7 GHz
Socket
Socket FP4
LGA 1151
Cores/Threads
4/4
4/8
TDP
45 W
- W
Cache L1 (d+i)
2x96+4x32 kB
4x32+4x32 kB
Cache L2
2x1024 kB
4x256 kB
Cache L3
0 kB
8192 kB
Date
June 2016
June 2016
Mean monothread perf.
23.7k points
56.39k points
Mean multithread perf.
32.92k points
211.6k points

AVX2 optimized benchmark
The benchmark in mode III (AVX2), like AVX1, is optimized to used 256 bits registers beside the second version of the Advanced Vector Extensions (AVX). The first AVX2 compatible CPU was released in 2013.
Monothread
9630P
506E8
Test#1 (Integers)
8.52k
23.91k (x2.81)
Test#2 (FP)
11.04k
21.03k (x1.91)
Test#3 (Generic, ZIP)
2.36k
5.05k (x2.14)
Test#1 (Memory)
1.79k
6.4k (x3.57)
TOTAL
23.7k
56.39k (x2.38)

Multithread

9630P

506E8
Test#1 (Integers)
12.83k
95.3k (x7.43)
Test#2 (FP)
14.44k
91.94k (x6.37)
Test#3 (Generic, ZIP)
3.6k
22.06k (x6.12)
Test#1 (Memory)
2.05k
2.3k (x1.12)
TOTAL
32.92k
211.6k (x6.43)

Performance/W
9630P
506E8
Test#1 (Integers)
285 points/W