Post your 7-Zip v22.01 scores

[Paramount]

Compr = 106.550
Decompr =109.528
Total = 108.039

 

[QuietBob]

Athlon II X2 215 @ 3.8 GHz

 

[InVasMani]

147000K 6600MT/s CL28.

 

[mrcardio33x]

CPU: BCLK 105 - LLC7 - Scaler X6 - (-) 175 MHz or 5125MHz, Stable, SHA3, etc.

 

[xrror]

Revisting the FX 8310 today, actually figured out it needed more uncore voltage. So now can actually run 2600+ uncore and run at 1866 memory...

Put both the old and new pics, might be invalid though cause this is 7zip 24.01 (not 22) not sure if that changes the benchmark?

Old bench from before with slow uncore and memory at 1600:

new with actual uncore at actual speed and memory at 1866:

It's weird utilization goes down, maybe throttling?

 

[Paramount]

7900X
PBO 2 / CO -30 all cores
+200Mhz
scalar 6x
power limit 155w

 

[xrror]

Socket 1366 45nm Bloomfield time...
Can't seem to get a sweet spot for core clock, so just playing around on memory side lol

Bloomfield don't care 'bout no volts lol

wish I could get uncore up more but what's that smell? oh yea burning IMC =D


------------ forum merged these posts so different machine (gulftown 32nm hex core, but identical mobos though) below ------

Last socket 1366 for me today - for some reason not having luck with higher BCK on this combo, and uncore isn't anything great on this chip.

Mostly a comparison check between 7zip 22.01 and 24.01 but I need to recheck the number to be sure this rig hasn't changed since last time.

------- last edit add ----
so i realized i could just download 7zip 22.01 from his github so just downgraded and ran it:

doesn't seem like much difference between 22.01 and 24.01 horray

 

[Paramount]

7900X
PBO 2 / CO -30 all cores
+200Mhz
scalar 6x
power limit 155w

View attachment 335926

6200/2066 FCLK improve result a bit

 

[AMF]

 

[DaemonForce]

Ryzen workstation with 50 days of uptime, a week of random update nags and all browsers/streams closed:
Compression: 61.43G
Decompress: 71.27G
Total Result: 66.35G

FX Rack with 4 days of uptime and a bunch of junk HDDs connected:
Compression: 28.65G
Decompress: 39.69G
Total Result: 34.17G

The eMachines/Acer with 2 days of uptime and the 20TB hookup:
Compression: 1.54G
Decompress: 2.48G
Total Result: 2.01G

Fast af boiiiiiii! (=ﾟωﾟ)

 

[AMF]

 

[xrror]

The 1600AF returns with... actual memory timings this time - though ironically it seems slower?

--- crosscheck with v22.01 - seems about same

 

[Zyll Goliat]

X299 I9-7940X slightly OC/Undervolted on Air

Compr = 125.944
Decompr =130.841
Total = 128.393

 

[QuietBob]

Fast af boiiiiiii! (=ﾟωﾟ)

That energy-efficient Athlon is just as fast (read: slow) at 1600 MHz as my regular Sempron at 1862 MHz. Though I would've expected it to do better with DDR2 in dual channel.

And still twice as fast as my Celeron D 326, proudly the slowest CPU in the chart

 

[DaemonForce]

That energy-efficient Athlon is just as fast (read: slow) at 1600 MHz as my regular Sempron at 1862 MHz.

Because of the nature of 7-Zip benchmark tool having numerous possible bench configurations, it's interesting to pull down something newer like my 3600 and do direct compares. The Sempron might be just as fast (or slow) but the same can be said about my FX-8370 and everything before it.

Pulling one thread and noting the clock as 2.5x the Athlon, something very interesting happens when looking at the frequencies as scaled:
The compress, decompress and total results of the Athlon 64 were
1.54 2.48 2.01
The results of Ryzen downscaled to match the speed become
2.32 3.06 2.69
However it only takes ~86 seconds to achieve that where the Athlon 64 goes just over 5min. At scale the Ryzen would take 214sec.
The ratings are still very close but the time to get there marks real improvement. Ryzen is amazing at any speed.
It's hard to believe that just 15 years ago, that Athlon 64 was advertised to be the heart of a POWERFUL SLEEK PC.
It's also just as surprising that even without modern day virtualization instructions it can handle Win10, Server 2016/2019/2022 and Azure just fine but crashes 4x loading Ubuntu desktop.
This is a very oddball timeline. :/

Though I would've expected it to do better with DDR2 in dual channel.

Never found the settings for it. It would have been nice to get the full 800MT speeds out of it. I can always overclock this thing without any thermal or electrical consequences but any attempts to restore the CPU Freq are met with a crash.

 

[MaddoggMiranda]

Compr = 75.309
Decompr =223.948
Total = 149.628
2350%

@QuietBob

 

[xrror]

More old school hi-jinx:

Sadly 4Ghz requires a jump down to 1600 memory and more voltage than I really want to run on the 'ol chip (>1.45v core. 1.30v uncore) and it still wasn't really stable. There's probably a way to do it but I'm no socket AM3 master. Being stuck at the x13.5 mult kinda puts a damper on finding an optimal mix past this. Probably not ideal dragging 4 sticks of RAM along for the ride either lol (though surprisingly didn't seem to change much going to just two sticks?)

---
okay yea had to try - but it's not happy about it - not really stable and volts kinda high. I do NOT consider it a valid result, this was just more "what if?"

note: pasted in the correct memory cpuz tab cause i screwed it up originally (both were just showing cpu tab originally)
also answered own question on four sticks - had to pull 2 out for it to even post. Guessing we're on the edge of what IMC can handle or I need to slow IMC down.