AMD Ryzen 5 7600X gmpbench submission (result 8010.2) -- new: 9176.2
hermann at stamm-wilbrandt.de
hermann at stamm-wilbrandt.de
Sat Jul 8 21:14:50 CEST 2023
On 2023-07-08 16:48, Torbjörn Granlund wrote:
> [snip]
>
> GMPbench: 8010.2
>
> Interesting! I would hope Zen 4 could perform much better than that,
> though.
>
> Could you please repeat the test with the most latest GMP snapshot?
>
> https://gmplib.org/download/snapshot/gmp-next/
>
THANK YOU!
I installed latest snapshot:
gmp-6.2.99-20230706074853.tar.zst
./configure seems to know about zen4:
...
configure: summary of build options:
Version: GNU MP 6.2.99
Host type: zen4-pc-linux-gnu
...
The result is 14.5% increase in GMPbench value, new value is 9176.2 !
New submission:
A. office PC with 7600X CPU:
https://github.com/Hermann-SW/7600X#details-of-pc
B. AMD Ryzen 5 7600X (4700MHz-5300MHz)
I did run "perf stat -e cycles,task-clock ./runbench" with slightly
worse results, but it reported more than 5.3GHz CPU frequency.
C. Using gmp version: gmp-6.2.99-20230706074853.tar.zst snapshot
D. no modifications
E. gcc version 11.3.0 (Ubuntu 11.3.0-1ubuntu1~22.04.1)
F. see log below
hermann at 7600x:~/gmpbench-0.2$ ./runbench
***** GMPbench version 0.2 *****
Using default CFLAGS = "-O3 -fomit-frame-pointer"
Using default CC = "gcc"
Using default LIBS = "-static -lgmp"
Using compilation command: gcc -O3 -fomit-frame-pointer foo.c -o foo
-static -lgmp
You may want to override CC, CFLAGS, and LIBS
Using gmp version: 6.2.99
Compiling benchmarks
Running benchmarks (propagated score accuracy exceeds printed
intermediates)
Category base
Program multiply (weight=1)
GMPbench.base.multiply(128) 2.185e+08
GMPbench.base.multiply(512) 5.43e+07
GMPbench.base.multiply(8192) 5.792e+05
GMPbench.base.multiply(131072) 1.005e+04
GMPbench.base.multiply(2097152) 390
GMPbench.base.multiply(128,128) 1.909e+08
GMPbench.base.multiply(512,512) 3.822e+07
GMPbench.base.multiply(8192,8192) 3.731e+05
GMPbench.base.multiply(131072,131072) 6699
GMPbench.base.multiply(2097152,2097152) 250
GMPbench.base.multiply(15000,10000) 1.884e+05
GMPbench.base.multiply(20000,10000) 1.472e+05
GMPbench.base.multiply(30000,10000) 9.288e+04
GMPbench.base.multiply(16777216,512) 1586
GMPbench.base.multiply(16777216,262144) 59
GMPbench.base.multiply 1.2534e+05
Program divide (weight=1)
GMPbench.base.divide(8192,32) 3.527e+06
GMPbench.base.divide(8192,64) 3.788e+06
GMPbench.base.divide(8192,128) 2.644e+06
GMPbench.base.divide(8192,4096) 8.85e+05
GMPbench.base.divide(131072,65536) 9215
GMPbench.base.divide(8388608,4194304) 52.7
GMPbench.base.divide(8192,8064) 3.943e+07
GMPbench.base.divide(16777216,262144) 30.2
GMPbench.base.divide 1.0769e+05
Program gcd (weight=0.5)
GMPbench.base.gcd(128,128) 1.179e+07
GMPbench.base.gcd(512,512) 1.17e+06
GMPbench.base.gcd(8192,8192) 3.692e+04
GMPbench.base.gcd(131072,131072) 447
GMPbench.base.gcd(1048576,1048576) 24.4
GMPbench.base.gcd 22334
Program gcdext (weight=0.5)
GMPbench.base.gcdext(128,128) 4.039e+06
GMPbench.base.gcdext(512,512) 9.925e+05
GMPbench.base.gcdext(8192,8192) 2.654e+04
GMPbench.base.gcdext(131072,131072) 316
GMPbench.base.gcdext(1048576,1048576) 15.8
GMPbench.base.gcdext 13965
GMPbench.base 62004
Category app
Program rsa (weight=1)
GMPbench.app.rsa(512) 7.785e+04
GMPbench.app.rsa(1024) 1.726e+04
GMPbench.app.rsa(2048) 2684
GMPbench.app.rsa 15334
Program pi (weight=1)
GMPbench.app.pi(10000) 2589
GMPbench.app.pi(100000) 107
GMPbench.app.pi(1000000) 6.28
GMPbench.app.pi 120.27
GMPbench.app 1358
GMPbench: 9176.2
hermann at 7600x:~/gmpbench-0.2$
Regards,
Hermann Stamm-Wilbrandt.
More information about the gmp-devel
mailing list