• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

Post your 7-Zip v22.01 scores

My old "mule" / X79-Platform Xeon 2697 V2 12c/24t 3,4Ghz(All-cores)......
74,2-Compressing
81,2-Decompressing
77,7-Overall

7_Zip_Benchmark.png
7_zip_memory.png
 
Last edited:
Oh terribly sorry my friend, I thought we were just playing around, I didn’t see you were keeping track until now

Was just testing settings.. 5 was faster than 10
Thanks, your scores have been updated!

Ran it on my $50 Windows tablet. Can only do x86 though, so I dunno if it can go on the scoreboard.
Those are certainly interesting results, but let's keep it 64-bit for consistency. I've got a few old systems running Win7 myself, it'll be interesting to see some single cores later :D
 
35.3
49.4
42.3

I really need to reconfigure my mem on my R5 1600 :D
 
Was just testing settings.. 5 was faster than 10 :)
But since you guys insist :laugh:

176.580
250.956
213.768


View attachment 252602
Did some more runs just for fun. ;)
This is one of the few benchmarks where Zen3 seem to benefit (compressing) from running the infinity fabric asynced from the memory speed, as long as you can run high enough MT/s..
(but there is a small reduction in decompressing speed)

5950x with infinity fabric @ 1900mhz and memory @ 4466MT/s with decently tight timings :toast:

Compressing = 182.202 gflops
Decompressing = 250.876 gflops
Total average = 216.539 gflops
1656340805424.png


Above 1900:3800 synced i get whea errors and reduced performance in everything but superlight Aida64 "benchmarks" on my current setup.
 
653K9rV.png

43.1
58.2
50.6

Seems better after reload and no alt tab games :D
 
5600 (I think 3800 CL15 may be faster, will test that tomorrow)

78.59
88.40
83.49

7Zip 5600 Bench 83 crop.jpg


ZenTimings_Screenshot 4000 CL16 GSKILL.png
 
New submissions currently on hold as the time limit for editing my post has expired :( Already contacted the forum mods about it.

EDIT:
I can add submissions again, awaiting your scores :)
Thanks @phill and @the54thvoid
 
Last edited:
That's cool, but has Igor Pavlov also fixed the vulnerabilities in his software? Has Igor Pavlov stopped living in the past and updated 7-Zip's context menu to support Windows 11's new context menu? No? Bandizip it is then.
 
Ran it on my Steam Deck. CPU-Z doesn't like this device at all.

Steam Deck CPU (AMD Van Gogh/Custom APU 0405), 16GB LPDDR5-5500.

25.4
36.2
30.8


TEToxc4.png
 
Just a placeholder... to show you a very ancient system! I'm under linux so I can't do a cpu_z memory shot... can't get decode-dimms to bloody work... so here's unofficial for
a Xeon 1680-v2 at 4.3, and quad channel at 1866:
7-zip1.png
 
Going to throw a spanner in the works here... In my opinion, an interesting comparison with Rocket Lake architecture (cypress cove cores) in relation to memory subsystem performance running in both gears 1 & 2.
For the record, I'd like to nominate gear 1 as the final result for myself, however keep in mind my 11700k does not have the best gear 1 IMC. Can't get the system to boot successfully, let alone run memory stress testing higher than 3466MHz on the RAM with gear 1. Also I only have Hynix DJR sticks to play with... :)

Gear 1 results;
85.0
103.3
94.2
7z_benchmark._G1 - Copy.JPG

Gear 2 results;
73.8
95.8
83.8
7z_benchmark._G2 - Copy.JPG
 
Phenom II X4 @ 4.18
18.9
25.1
22.0
phenom975.jpg
 
Going to throw a spanner in the works here... In my opinion, an interesting comparison with Rocket Lake architecture (cypress cove cores) in relation to memory subsystem performance running in both gears 1 & 2.
For the record, I'd like to nominate gear 1 as the final result for myself, however keep in mind my 11700k does not have the best gear 1 IMC. Can't get the system to boot successfully, let alone run memory stress testing higher than 3466MHz on the RAM with gear 1. Also I only have Hynix DJR sticks to play with... :)
Gear 1 results;
85.0
103.3
94.2
Gear 2 results;
73.8
95.8
83.8
Interesting results, pretty much the opposite findings compared to the results from my 16core Zen3...

Running the infinity fabric unsync'ed from the memory on Zen is almost the same as running Intel gen11 and gen12 in gear 2 mode. We get half the uclck-speed with AMD and with Intel you get half the memory-controller speed. (both should be comparable somewhat)

With synced 1900:1900:1900 (3800MT/s) i got the following:
Compressing = 176.580 gflops
Decompressing = 250.956 gflops
Total average = 213.768 gflops

Asynced 1900:1166:2233 (4466MT/s)
Compressing = 182.202 gflops
Decompressing = 250.876 gflops
Total average = 216.539 gflops

Strangely its only the compressing numbers that improve with higher bandwidth for me (at the expense of latency)
The only other benchmark i know i get higher results from running unsynced is y-cruncher where i get a small benefit above 1900:4466MT/s compared to 1900:3800MT/s synced.
Did you try/can you run higher memory clockspeeds in gear 2 mode ?

New submissions currently on hold as the time limit for editing my post has expired :( Already contacted the forum mods about it.

EDIT:
I can add submissions again, awaiting your scores :)
Thanks @phill and @the54thvoid
Please update my score :)
1656426294593.png
 
Last edited:
To be fair, nearly no one has.
And they are at least starting to add 7-zip to the new context menu. It now has at least the option to open things with 7zip when you right click on a file.
 
AMD 5900X
I used Nanazip (a 7-zip clone with a nice W11 integration). Hope that's OK.
It identifies W11 as W10.

91.590
167.020
129.305

Screenshot 2022-06-28 192106.jpg
Screenshot 2022-06-28 192115.jpg
 
i5 11400
55.7
64.9
60.3

11400.jpg


i3 12100f
46.8
44.3
45.6

12100f.jpg
 
I used Nanazip (a 7-zip clone with a nice W11 integration).
Nanazip is a fork of 7-zip so I'm afraid I can't include your results in the chart. But I'm sure Win10/11 users missing the Explorer context menu will find this app helpful :)
 
Back
Top