7zip benchmark

# List compiled by Götz Hoffart, Freiburg                    # Columns:
# ... using '7za b'                                          #  1. Name of the system/model
# Thanks to PoC, Marek C., Alex B., Thomas K.,               #  2. Specific model
# Patrick K., Christian D., Dominik S., Dennis G.,           #  3. Type of CPU (shortened, see below table)
# Alexander B., Georg B., Paul W. and many others            #  4. MHz per CPU
#                                                            #  5. Number of Cores in total
# If you want to know which system ...                       #  6. Number of Threads (a dual core CPU with Hyperthreading
#   ... is the fastest in overall speed: look for "7z-MIPS"  #     = 2 Cores with 2 Threads each = 4 threads)
#   ... is the fastest per CPU core: look for "7z-MIPS/Core" #  7. Amount of RAM/memory built-in in GB
#                                                            #  8. Operating system used
#                                                            #  9. 7zip-MIPS in total, output of '7z b'. The higher the better.
#                                                            # 10. 7zip-MIPS per Core: to judge how fast a single core system
#                                                            #     theoretically would be. The higher the better.
#                                                            # 11. 7zip-MIPS per Thread: to judge how fast a thread in this system is.
#                                                            #     The higher the better.
#                                                            # 12. 7zip-MIPS per Core per MHz: how efficient is the system when taking
#                                                            #     into account the performance per Core and per MHz. The higher the better.
#                                                            #     Not available for CPUs that overclock to higher frequencies when possible.

Remarks:
M1          = Apple M1 ARM64
M2          = Apple M2 ARM64
M3          = Apple M3… you guessed it.
CD          = Intel Core Duo
C2D         = Intel Core 2 Duo

[1]: p7zip is relatively slow here due to small CPU caches (512 bytes), and only 16 MHz main/memory bus.


#### how to get benchmark results:
## 1) Preparation: close all running processes
## 2) If it's a laptop, attach it to PSU and/or set all energy saving
##    levels to “off”
## 3) run “7za b”
## 4) On Linux: run “uname -a” and “free” and “cat /proc/cpuinfo”
## 5) On macOS: run “sysctl hw.model hw.memsize hw.machine hw.ncpu machdep.cpu.brand_string hw.perflevel0 hw.perflevel1”
##    and “sw_vers”
## 6) send the result to "bench@hoffart.de"
####

• Log of 7zip Benchmark
• back to tiffy’s web page
• tableSorter by Christian Bach
• table live filter w3schools