| | | | | | |

Ryzen 9 7950X vs 3900


Description
The 7950X is based on Zen 4 architecture while the 3900 is based on Zen 2.

Using the multithread performance as a reference, the 7950X gets a score of 1597.8 k points while the 3900 gets 687.5 k points.

Summarizing, the 7950X is 2.3 times faster than the 3900. To get a proper comparison between both models, take a look to the data shown below.

Specs
CPUID
a60f12
870f10
Core
Raphael
Matisse
Architecture
Zen 4
Base frecuency
4.5 GHz
3.1 GHz
Boost frecuency
5.7 GHz
4.3 GHz
Socket
AM5
AM4
Cores/Threads
16/32
12/24
TDP
170 W
65 W
Cache L1 (d+i)
16x32+16x32 kB
12x32+12x32 kB
Cache L2
16x1024 kB
12x512 kB
Cache L3
131072 kB
4x16384 kB
Date
September 2022
September 2019
Mean monothread perf.
121.33k points
74.97k points
Mean multithread perf.
1597.79k points
687.5k points

Non-optimized benchmark
The benchmark in Mode 0 (FPU) measures cpu performance with non-optimized software. It uses the basic µinstructions from the i386 architecture with the i387 floating point unit. This mode is compatible with all CPUs so it's practical to compare very different CPUs
Monothread
7950X
3900
Test#1 (Integers)
6.05k
4.39k (x0.73)
Test#2 (FP)
23.41k
16.99k (x0.73)
Test#3 (Generic, ZIP)
16.49k
7.74k (x0.47)
Test#1 (Memory)
34.35k
24.57k (x0.72)
TOTAL
80.3k
53.69k (x0.67)

Multithread

7950X

3900
Test#1 (Integers)
128.82k
48.45k (x0.38)
Test#2 (FP)
465.79k
230.81k (x0.5)
Test#3 (Generic, ZIP)
289.43k
122.82k (x0.42)
Test#1 (Memory)
75.23k
54.73k (x0.73)
TOTAL
959.27k
456.8k (x0.48)

AVX optimized benchmark
The benchmark in mode II (AVX) is optimized to used 256 bits registers beside the first version of the Advanced Vector Extensions (AVX). The first AVX compatible CPU was released in 2011.
Monothread
7950X
3900
Test#1 (Integers)
26.86k
16.16k (x0.6)
Test#2 (FP)
30.05k
23.75k (x0.79)
Test#3 (Generic, ZIP)
16.15k
9.32k (x0.58)
Test#1 (Memory)
31.74k
23.17k (x0.73)
TOTAL
104.8k
72.4k (x0.69)

Multithread

7950X

3900
Test#1 (Integers)
430.6k
174.73k (x0.41)
Test#2 (FP)
545.91k
275.63k (x0.5)
Test#3 (Generic, ZIP)
289.57k
125.65k (x0.43)
Test#1 (Memory)
64.33k
44.17k (x0.69)
TOTAL
1330.41k
620.18k (x0.47)

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
7950X
3900
Test#1 (Integers)
40.61k
16.85k (x0.42)
Test#2 (FP)
31.82k
26.03k (x0.82)
Test#3 (Generic, ZIP)
16.63k
9.54k (x0.57)
Test#1 (Memory)
32.28k
22.55k (x0.7)
TOTAL
121.33k
74.97k (x0.62)

Multithread

7950X

3900
Test#1 (Integers)
683.17k
229.41k (x0.34)
Test#2 (FP)
563.45k
292.81k (x0.52)
Test#3 (Generic, ZIP)
286.77k
128.48k (x0.45)
Test#1 (Memory)
64.4k
36.8k (x0.57)
TOTAL
1597.79k
687.5k (x0.43)

Performance/W
7950X
3900
Test#1 (Integers)
4019 points/W
3529 points/W
Test#2 (FP)
3314 points/W
4505 points/W
Test#3 (Generic, ZIP)
1687 points/W
1977 points/W
Test#1 (Memory)
379 points/W
566 points/W
TOTAL
9399 points/W
10577 points/W

Performance/GHz
7950X
3900
Test#1 (Integers)
7124 points/GHz
3920 points/GHz
Test#2 (FP)
5582 points/GHz
6054 points/GHz
Test#3 (Generic, ZIP)
2917 points/GHz
2218 points/GHz
Test#1 (Memory)
5663 points/GHz
5243 points/GHz
TOTAL
21286 points/GHz
17435 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.


Test#1 (Integers) [points vs time]

grafica bm.hardlimit.com


Test#2 (FP) [points vs time]

grafica bm.hardlimit.com


Test#3 (Generic, ZIP) [points vs time]

grafica bm.hardlimit.com


Test#1 (Memory) [points vs time]

grafica bm.hardlimit.com

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.


Test#1 (Integers) [points vs time]

grafica bm.hardlimit.com


Test#2 (FP) [points vs time]

grafica bm.hardlimit.com


Test#3 (Generic, ZIP) [points vs time]

grafica bm.hardlimit.com


Test#1 (Memory) [points vs time]

grafica bm.hardlimit.com

Hardlimit Benchmark Central - Ver. 3.11.4