| | | | | | |

Ryzen 9 3900X vs Xeon E5-2670 v3


Description
The 3900X is based on Zen 2 architecture while the E5-2670 v3 is based on Haswell.

Using the multithread performance as a reference, the 3900X gets a score of 756.3 k points while the E5-2670 v3 gets 452.9 k points.

Summarizing, the 3900X is 1.7 times faster than the E5-2670 v3. To get a proper comparison between both models, take a look to the data shown below.

Specs
CPUID
870f10
306f2
Core
Matisse
Haswell-EP
Architecture
Base frecuency
3.8 GHz
2.3 GHz
Boost frecuency
4.6 GHz
3.1 GHz
Socket
AM4
LGA 2011-3
Cores/Threads
12/24
12/24
TDP
105 W
120 W
Cache L1 (d+i)
12x32+12x32 kB
12x32+12x32 kB
Cache L2
12x512 kB
12x256 kB
Cache L3
4x16384 kB
30720 kB
Date
July 2019
September 2014
Mean monothread perf.
72.51k points
36.41k points
Mean multithread perf.
756.3k points
452.9k 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
3900X
E5-2670 v3
Test#1 (Integers)
4.46k
3.52k (x0.79)
Test#2 (FP)
18.46k
8.68k (x0.47)
Test#3 (Generic, ZIP)
8.05k
2.83k (x0.35)
Test#1 (Memory)
23.71k
3.13k (x0.13)
TOTAL
54.68k
18.17k (x0.33)

Multithread

3900X

E5-2670 v3
Test#1 (Integers)
52.65k
32.7k (x0.62)
Test#2 (FP)
265.41k
105.18k (x0.4)
Test#3 (Generic, ZIP)
140.78k
35.18k (x0.25)
Test#1 (Memory)
46.91k
9.54k (x0.2)
TOTAL
505.76k
182.59k (x0.36)

SSE3 optimized benchmark
The benchmark in mode I (SSE) is optimized for the use of SIMD instructions with 128 bits register and the SSE set up to version 3. Nearly every modern CPU has support for this mode.
Monothread
3900X
E5-2670 v3
Test#1 (Integers)
17.35k
11.08k (x0.64)
Test#2 (FP)
25.87k
9.57k (x0.37)
Test#3 (Generic, ZIP)
8.72k
3.18k (x0.37)
Test#1 (Memory)
27.86k
3.28k (x0.12)
TOTAL
79.8k
27.11k (x0.34)

Multithread

3900X

E5-2670 v3
Test#1 (Integers)
201.57k
139.07k (x0.69)
Test#2 (FP)
284.41k
133.47k (x0.47)
Test#3 (Generic, ZIP)
145.16k
46.6k (x0.32)
Test#1 (Memory)
100.56k
6.65k (x0.07)
TOTAL
731.7k
325.78k (x0.45)

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
3900X
E5-2670 v3
Test#1 (Integers)
16.95k
11.18k (x0.66)
Test#2 (FP)
25.37k
10.06k (x0.4)
Test#3 (Generic, ZIP)
9.23k
3.21k (x0.35)
Test#1 (Memory)
24.83k
3.12k (x0.13)
TOTAL
76.38k
27.57k (x0.36)

Multithread

3900X

E5-2670 v3
Test#1 (Integers)
196.31k
140.98k (x0.72)
Test#2 (FP)
290.5k
145.37k (x0.5)
Test#3 (Generic, ZIP)
143.95k
45.62k (x0.32)
Test#1 (Memory)
64.92k
6.75k (x0.1)
TOTAL
695.69k
338.73k (x0.49)

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
3900X
E5-2670 v3
Test#1 (Integers)
16.73k
18.62k (x1.11)
Test#2 (FP)
24.74k
10.49k (x0.42)
Test#3 (Generic, ZIP)
8.82k
4.04k (x0.46)
Test#1 (Memory)
22.22k
3.26k (x0.15)
TOTAL
72.51k
36.41k (x0.5)

Multithread

3900X

E5-2670 v3
Test#1 (Integers)
248.24k
236.25k (x0.95)
Test#2 (FP)
309.66k
151.05k (x0.49)
Test#3 (Generic, ZIP)
140.01k
58.9k (x0.42)
Test#1 (Memory)
58.39k
6.7k (x0.11)
TOTAL
756.3k
452.9k (x0.6)

Performance/W
3900X
E5-2670 v3
Test#1 (Integers)
2364 points/W
1969 points/W
Test#2 (FP)
2949 points/W
1259 points/W
Test#3 (Generic, ZIP)
1333 points/W
491 points/W
Test#1 (Memory)
556 points/W
56 points/W
TOTAL
7203 points/W
3774 points/W

Performance/GHz
3900X
E5-2670 v3
Test#1 (Integers)
3638 points/GHz
6006 points/GHz
Test#2 (FP)
5377 points/GHz
3385 points/GHz
Test#3 (Generic, ZIP)
1917 points/GHz
1303 points/GHz
Test#1 (Memory)
4831 points/GHz
1052 points/GHz
TOTAL
15763 points/GHz
11745 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