Ryzen 9 3900 vs 3950X
Description
Both models 3900 and 3950X are based on Zen 2 architecture.
Zen 2 is fabricated on the 7nm process from TSMC and it’s the third generation of Zen CPUs. It comes with 64kB of L1 cache and 512kB of L2 cache per core. Zen 2 CPUs are divided into 3 categories: Matisse (desktop), Rome (Server) and Castle Peak (high-end desktop).
Using the multithread performance as a reference, the 3900 gets a score of 687.5 k points while the 3950X gets 959.8 k points.
Summarizing, the 3950X is 1.4 times faster than the 3900. To get a proper comparison between both models, take a look to the data shown below.
Zen 2 is fabricated on the 7nm process from TSMC and it’s the third generation of Zen CPUs. It comes with 64kB of L1 cache and 512kB of L2 cache per core. Zen 2 CPUs are divided into 3 categories: Matisse (desktop), Rome (Server) and Castle Peak (high-end desktop).
Using the multithread performance as a reference, the 3900 gets a score of 687.5 k points while the 3950X gets 959.8 k points.
Summarizing, the 3950X is 1.4 times faster than the 3900. To get a proper comparison between both models, take a look to the data shown below.
Specs
CPUID
870f10
870f10
Core
Matisse
Matisse
Base frecuency
3.1 GHz
3.5 GHz
Boost frecuency
4.3 GHz
4.7 GHz
Socket
AM4
AM4
Cores/Threads
12/24
16/32
TDP
65 W
105 W
Cache L1 (d+i)
12x32+12x32 kB
16x32+16x32 kB
Cache L2
12x512 kB
16x512 kB
Cache L3
4x16384 kB
4x16384 kB
Date
September 2019
November 2019
Mean monothread perf.
74.97k points
72.95k points
Mean multithread perf.
687.5k points
959.77k 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
3900
3950X
Test#1 (Integers)
16.85k
16.49k (x0.98)
Test#2 (FP)
26.03k
26.16k (x1.01)
Test#3 (Generic, ZIP)
9.54k
7.9k (x0.83)
Test#1 (Memory)
22.55k
22.4k (x0.99)
TOTAL
74.97k
72.95k (x0.97)
Multithread
3900
3950X
Test#1 (Integers)
229.41k
334.49k (x1.46)
Test#2 (FP)
292.81k
423.74k (x1.45)
Test#3 (Generic, ZIP)
128.48k
187.85k (x1.46)
Test#1 (Memory)
36.8k
13.69k (x0.37)
TOTAL
687.5k
959.77k (x1.4)
Performance/W
3900
3950X
Test#1 (Integers)
3529 points/W
3186 points/W
Test#2 (FP)
4505 points/W
4036 points/W
Test#3 (Generic, ZIP)
1977 points/W
1789 points/W
Test#1 (Memory)
566 points/W
130 points/W
TOTAL
10577 points/W
9141 points/W
Performance/GHz
3900
3950X
Test#1 (Integers)
3920 points/GHz
3509 points/GHz
Test#2 (FP)
6054 points/GHz
5566 points/GHz
Test#3 (Generic, ZIP)
2218 points/GHz
1681 points/GHz
Test#1 (Memory)
5243 points/GHz
4765 points/GHz
TOTAL
17435 points/GHz
15522 points/GHz
Monothread performance graph
Monothread performance graphics gives the performance vs time. They are useful to measure the time it takes to the CPU to reach the maximum performance.
Usually, CPU's performance will be steady during these tests but if it has a slow frequency strategy, the first samples will show a lower score.
Usually, CPU's performance will be steady during these tests but if it has a slow frequency strategy, the first samples will show a lower score.
Test#1 (Integers) [points vs time]
Test#2 (FP) [points vs time]
Test#3 (Generic, ZIP) [points vs time]
Test#1 (Memory) [points vs time]
Multithread performance graph
Multithread graphs measure the performance against a heavy load during certain time.
If CPU's TDP doesn't limit the frequency and the machine is properly cooled, performance should remain steady vs time. Otherwise, the performance score will oscillate or decrease over time.
If CPU's TDP doesn't limit the frequency and the machine is properly cooled, performance should remain steady vs time. Otherwise, the performance score will oscillate or decrease over time.
Test#1 (Integers) [points vs time]
Test#2 (FP) [points vs time]
Test#3 (Generic, ZIP) [points vs time]
Test#1 (Memory) [points vs time]
Hardlimit Benchmark Central - Ver. 3.11.4