RapydMark CPU benchmark

[igor_kavinski]

RapydMark for Windows (frikiscape.com)

Synthetic highly parallel benchmark.




My company's dual socket Ivy Bridge-E beat Comet Lake in total time, only due to more cores. Should be fun to see how 5950X and i9-12900K fare in this benchmark. My guess is that 5950X will win.

Update: 5950X did win! But i9-12900K put up a good fight despite fewer cores.

Performance comparison chart so far thanks to Makaveli:



Public view-only link for the sheet: RapydMark64 - Google Drive

CCD comparison by Det0x:

1 CCD -> 8 cores / 8 threads = 368.309 seconds
1 CCD -> 8 cores / 16 threads = 230.797 seconds
2 CCD's -> 16 cores / 32 threads = 135.978 seconds

Interesting HT/no-HT comparison:

User​	CPU​	Cores​	Threads​	Time​
JoeRambo​	i9-12900K​	8​	16​	212.656​
JoeRambo​	5800X​	8​	16​	263.424​
Det0x​	5950X​	8​	16​	230.797​
Det0x​	5950X​	8​	8​	368.309​
JoeRambo​	i9-12900K​	8​	8​	328.88​

 

[moinmoin]

My company's dual socket Ivy Bridge-E beat Comet Lake in total time, only due to more cores.

In your screenshot Ivy Bridge-E didn't have to do the "Cathedral 3D Test" which appears to make up much of the difference.

Other than that interesting test suite. Anybody familiar about its dependency (or lack thereof) on the memory subsystem?

 

[igor_kavinski]

In your screenshot Ivy Bridge-E didn't have to do the "Cathedral 3D Test" which appears to make up much of the difference.

The server didn't have a 3D card. It's about 457 sec for the server vs. 472 sec for Comet Lake (minus Cathedral 3D's time).

 

[Det0x]

No idea if this is good or bad but i'm getting this with my tuned 5950x.
Normal ambient temp with all fans on auto



*edit*

Didn't notice stress level high, new runs

 

[igor_kavinski]

I think it's best not to select the Cathedral test since it's always around 88 secs. So that means 5950X completed all of the tests in less than a minute! It must have been fun seeing it crunch through them so quickly. I have a feeling 12900K can't beat that.

 

[nicalandia]

How good is this for IPC testing? What part of the CPU are tested? FPU? Memory subsystems?

This is what I got from a computer at work(32bits and without cathedral), I am running the 64bit one without cathedral now.




64bits

No idea if this is good or bad but i'm getting this with my tuned 5950x.

Such impressive Beast

 

[igor_kavinski]

How good is this for IPC testing? What part of the CPU are tested? FPU? Memory subsystems?

This is what I got from a computer at work(32bits and without cathedral), I am running the 64bit one without cathedral now.

View attachment 53334


64bits

View attachment 53335

Stress level should be high for comparable results. But not that bad for only two cores.

18% improvement going from 32-bits to 64-bits. Sounds about right for mostly pure number crunching workloads with a few memory bound ones.

 

[nicalandia]

Stress level should be high for comparable results. But not that bad for only two cores.

18% improvement going from 32-bits to 64-bits. Sounds about right for mostly pure number crunching workloads with a few memory bound ones.

It's a work computer, can't stress it too much or it will crash.

 

[igor_kavinski]

It's a work computer, can't stress it too much or it will crash.

I'm guessing you are not working in an IT capacity in your job? If my company forced me to work on a Core i3, I would do everything to kill it so I could get a newer one with more speed and more cores. Seriously, put in a request for an upgrade. Even a 10th gen i3 or Ryzen 3 would make a huge difference in responsiveness.

 

[Makaveli]

Here are my results without cathedral

Medium


High

 

[nicalandia]

Here are my results without cathedral

View attachment 53353

Any notes on your CPU config? Stock? PBO?

 

[Makaveli]

Any notes on your CPU config? Stock? PBO?

PBO + C0 is on single core boost to 5.05Ghz.

 

[Hitman928]

Here are my results without cathedral

View attachment 53353

What do you get with stress level high?

 

[Makaveli]

What do you get with stress level high?

I just noticed it on medium I've updated my original post with the high results.

 

[igor_kavinski]

Despite a 100MHz frequency deficit, the L4 cache likely helps the i7-5775C post 7% performance gain but man, it's still eye-wateringly slow. Funny thing is, I don't "feel" this slowness in general usage which so far doesn't involve much outside of browsing and watching movies.

 

[Tim_Derr]

 

[Det0x]

@JoeRambo How does the 12900k fare in this benchmark ?

 

[JoeRambo]

@Det0x I will run once i get home, i will most likely do an 8C P-core run only tho, would need to re-enable E-Cores and dial in OC or risk misrepresenting ADL

 

[Zucker2k]

@Det0x I will run once i get home, i will most likely do an 8C P-core run only tho, would need to re-enable E-Cores and dial in OC or risk misrepresenting ADL

Maybe @AdamK47 would do us the initial favor?

 

[JoeRambo]

Maybe @AdamK47 would do us the initial favor?

His system would surely represent things as Intel's marketing intended more than my 5Ghz 8C 30MB cache abomination.

 

[Det0x]

@Det0x I will run once i get home, i will most likely do an 8C P-core run only tho, would need to re-enable E-Cores and dial in OC or risk misrepresenting ADL

Great, looking forward to your results

I also did some new runs to check the core/thread scaling in this benchmark.

1 CCD -> 8 cores / 16 threads = 230.797 seconds



2 CCD's -> 16 cores / 32 threads = 135.978 seconds



Seems like we are seeing around ~70% scaling with core/threads in this benchmark

I think it's best not to select the Cathedral test since it's always around 88 secs. So that means 5950X completed all of the tests in less than a minute! It must have been fun seeing it crunch through them so quickly. I have a feeling 12900K can't beat that.

After some testing i found out that the Cathedra 3d test run dont actally count towards to "totalt time", so it dont matter if you run it or not..

 

[JoeRambo]

A no cathedral high run on well fed P-Core only 12900K:



EDIT: putting results of 8C vs 8C side by side i don't even know what to think about this benchmark and whatever virtues it has:

 

[Ed1]

here results on stock 12600k (didn't do last test, Cathedral 3d).
I must of done something wrong, only took 48s

Tests report generated by RapydMark v1.2a (64 bits) https://www.frikiscape.com/rapydmark/
=====================================
Date: 11-29-2021 (m-d-y)
Time: 10:34:20
CPU: 12th Gen Intel(R) Core(TM) i5-12600K (10 Cores)
Threads available: 16
GPU: NVIDIA GeForce GTX 970
Resolution: 1600 x 1200
Operating System: Windows 10 Pro Build 19044
System memory: 16193 MB
Stress level: Medium
System ID: F50165F2
=====================================
Results of all tests:

Total time (all tests): 48.177 seconds

 

[JoeRambo]

I must of done something wrong, only took 48s

That is a medium run, need to check High checkbox.

 

[Ed1]