| | | | | | |

Xeon E5-4640 vs Ryzen 9 3900X


Description
The E5-4640 is based on Sandy Bridge architecture while the 3900X is based on Zen 2.

Using the multithread performance as a reference, the E5-4640 gets a score of 610.2 k points while the 3900X gets 695.7 k points.

Summarizing, the 3900X is 1.1 times faster than the E5-4640 . To get a proper comparison between both models, take a look to the data shown below.

Specs
CPUID
206d7
870f10
Core
Sandy Bridge-EP
Matisse
Architecture
Base frecuency
2.4 GHz
3.8 GHz
Boost frecuency
2.8 GHz
4.6 GHz
Socket
LGA 2011
AM4
Cores/Threads
8 /16
12/24
TDP
95 W
105 W
Cache L1 (d+i)
8x32+8x32 kB
12x32+12x32 kB
Cache L2
8x256 kB
12x512 kB
Cache L3
20480 kB
4x16384 kB
Date
May 2012
July 2019
Mean monothread perf.
13.88k points
72.51k points
Mean multithread perf.
610.16k points
756.3k 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
E5-4640
3900X
Test#1 (Integers)
1.66k
4.46k (x2.68)
Test#2 (FP)
5.26k
18.46k (x3.51)
Test#3 (Generic, ZIP)
2.13k
8.05k (x3.78)
Test#1 (Memory)
2.53k
23.71k (x9.37)
TOTAL
11.58k
54.68k (x4.72)

Multithread

E5-4640

3900X
Test#1 (Integers)
77.64k
52.65k (x0.68)
Test#2 (FP)
212.2k
265.41k (x1.25)
Test#3 (Generic, ZIP)
113.11k
140.78k (x1.24)
Test#1 (Memory)
7.54k
46.91k (x6.22)
TOTAL
410.48k
505.76k (x1.23)

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
E5-4640
3900X
Test#1 (Integers)
6.19k
17.35k (x2.8)
Test#2 (FP)
5.89k
25.87k (x4.39)
Test#3 (Generic, ZIP)
2.11k
8.72k (x4.12)
Test#1 (Memory)
2.51k
27.86k (x11.11)
TOTAL
16.7k
79.8k (x4.78)

Multithread

E5-4640

3900X
Test#1 (Integers)
244.88k
201.57k (x0.82)
Test#2 (FP)
215.35k
284.41k (x1.32)
Test#3 (Generic, ZIP)
114.64k
145.16k (x1.27)
Test#1 (Memory)
6.76k
100.56k (x14.87)
TOTAL
581.63k
731.7k (x1.26)

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
E5-4640
3900X
Test#1 (Integers)
5.01k
16.95k (x3.38)
Test#2 (FP)
4.5k
25.37k (x5.63)
Test#3 (Generic, ZIP)
2.18k
9.23k (x4.24)
Test#1 (Memory)
2.19k
24.83k (x11.35)
TOTAL
13.88k
76.38k (x5.5)

Multithread

E5-4640

3900X
Test#1 (Integers)
251.29k
196.31k (x0.78)
Test#2 (FP)
236.61k
290.5k (x1.23)
Test#3 (Generic, ZIP)
114.89k
143.95k (x1.25)
Test#1 (Memory)
7.36k
64.92k (x8.82)
TOTAL
610.16k
695.69k (x1.14)

Performance/W
E5-4640
3900X
Test#1 (Integers)
2645 points/W
1870 points/W
Test#2 (FP)
2491 points/W
2767 points/W
Test#3 (Generic, ZIP)
1209 points/W
1371 points/W
Test#1 (Memory)
78 points/W
618 points/W
TOTAL
6423 points/W
6626 points/W

Performance/GHz
E5-4640
3900X
Test#1 (Integers)
1790 points/GHz
3686 points/GHz
Test#2 (FP)
1609 points/GHz
5515 points/GHz
Test#3 (Generic, ZIP)
777 points/GHz
2007 points/GHz
Test#1 (Memory)
782 points/GHz
5398 points/GHz
TOTAL
4957 points/GHz
16605 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