Cinebench "Extreme Edition" MOD by HwGeek

[EarthDog]

@Wavetrex I think it would be helpful, for comparison's sake, to require users to display CPUz's first tab and Memory tabs so users can have a comparison...

Im DLing and running now.


ALso, THANK YOU THANK YOU THANK YOU for not starting this thread as a shit pile of results!!!


EDIT: Why are we including results without ANY screenshot???

 

[Psychoholic]

9900K@5ghz -- 3200CL14 -- Corsair H110i

 

[EarthDog]

7960X @ 4.4 GHz 16/c16t, DDR4 @ 3600 CL16

I'll eventually throw one up there at the same speeds+ with all threads enabled.

 

[Tomgang]

I7 980X @ 4.4 GHz 6C/12T and triple channel memory DDR3 @ 1600 MHz, Air cooling (Noctua NH-D14 with noctua nf-f12 ippc-3000 pwm fans) score = 261

 

[Wavetrex]

@Wavetrex1) I think it would be helpful, for comparison's sake, to require users to display CPUz's first tab and Memory tabs so users can have a comparison...

2) ALso, THANK YOU THANK YOU THANK YOU for not starting this thread as a shit pile of results!!!
3) EDIT: Why are we including results without ANY screenshot???

1) Nah, it's a free country... ^.^
2) Seems right to have it ordered by score, no ?
3) As long as they seem to match the known abilities of that CPU, why not.

If someone comes with a very similar config but with screenshots, and has a totally different score, I will delete the lines that are dubious.

p.s. Poor AMD FX users

p.p.s. I have a working Pentium 4 board somewhere in the house. I will try to get it online and obtain an amazing score of .... something under 20

 

[EarthDog]

1) Nah, it's a free country... ^.^
2) Seems right to have it ordered by score, no ?
3) As long as they seem to match the known abilities of that CPU, why not.

If someone comes with a very similar config but with screenshots, and has a totally different score, I will delete the lines that are dubious.

1. Not sure what the country has to do with it... just trying to give users a better idea of the system if it includes those apps in the screeny. Results can vary and seeing a couple more detailed stats can help clarify at a glance. Like for example, if I didn't mention it, CPUz would have told you my result is w/o HT on...when others look at my CPU and think the result is terribly low. It works the other way too... 9900K scoring 1000, another scores 1050 at the same clock... memory speed/timings could be higher/tighter, etc. Just good stuff to have.
2. You misunderstood. Too often here people post a results thread without a scoreboard so its like hundreds of posts of, say, CPUz results without any way to reference them. This is great!
3. Sorry, that's the Hwbot/extreme bencher in me coming out. And as a Dad.... trust, but verify.

 

[MrGenius]

I'm pretty surprised I was able to get it to run @ 5.1GHz. But not too surprised than running it on Windows 7 gave me 3 more points than on Windows 10(all other things being equal). Just like the non-Extreme version in that regard.

i7-3770K(4C/8T) @ 5.1GHz, DDR3-2400 Dual Channel, Enermax Liqmax II 240

https://hwbot.org/submission/4065539_

 

[FireFox]

@Wavetrex

Memory here:

Cooler: Waterchiller but was running in Watercooler mode, who needs to turn a Waterchiller to run Cinebench, right?

 

[Wavetrex]

@MrGenius
I wonder how is that chip still alive... 1.68 V on water. You are completely insane, bonkers !

Anyway, 1 result per computer in the table, replaced your 5.0 with the new 5.1

@EarthDog
By "free country" I mean everybody is free to post whatever they want. That doesn't mean I will add it to the table if it doesn't seem legit.

 

[DR4G00N]

Cooler: Waterchiller but was running in Watercooler mode, who needs to turn a Waterchiller to run Cinebench, right?

My 8700K does so it doesn't cook itself. Quick run of normal R15 @ 5.1 1.35V on custom water cores will pass 80c. AVX load will reach near 100c and be very unstable.

p.p.s. I have a working Pentium 4 board somewhere in the house. I will try to get it online and obtain an amazing score of .... something under 20

Probably less than that, I will guess 1 to 5pts.

 

[Tomgang]

@MrGenius
I wonder how is that chip still alive... 1.68 V on water. You are completely insane, bonkers !

Anyway, 1 result per computer in the table, replaced your 5.0 with the new 5.1

@EarthDog
By "free country" I mean everybody is free to post whatever they want. That doesn't mean I will add it to the table if it doesn't seem legit.

Just one little thing. Cut you change my CPU name from i7-780K to I7 980X, cause that is what my CPU really is.

EDIT: It has now been fixed

 

[MrGenius]

@MrGeniusYou are completely insane, bonkers !

Card-carrying, certified, crazier than a shithouse rat!

I'm might be able to run it with a little less volts...but I figured that oughta be enough to get the job done. I was right!

 

[FireFox]

It works the other way too... 9900K scoring 1000, another scores 1050 at the same clock...

I have thought it is possible

 

[storm-chaser]

Intel Core 2 quad Q9650 @ 4.275Ghz on air cooling (Ultra 120 extreme)

 

[FireFox]

My 8700K does so it doesn't cook itself. Quick run of normal R15 @ 5.1 1.35V on custom water cores will pass 80c. AVX load will reach near 100c and be very unstable.

That is a Mother of CPU

 

[Hardi]

2700X @ 4.25GHz / 16GB DDR4 3600 c16 / NH-D15

 

[kniaugaudiskis]

Here's my attempt: Intel Xeon E5-2696 V3 @3.33GHz on all cores (running cool and quiet with BeQuiet! Dark Rock 3), quad-channel DDR4-2155MHz CL11

 

[er557]

I really like the performance of the workstation, windows 19H1 thread scheduler seems to be well optimized, and the turbo hack undervolts by -80mv, all is dead stable, due to running the latest microcode update

 

[Wavetrex]

@er557 The scaling doesn't seem to be great though on dual-socket.

The other guy with a Xeon E5-18 cores seems to have higher "IPC" by 19-20% (Aka, your score is not more than double, as it should with two CPU's and higher clocks, but instead only 76% higher)

 

[er557]

Of course, but the reason for that is NOT dual socket scaling, but the fact that his cpu is of higher end model, not as power efficient as mine, his cpu turbo's on default three multipliers higher, 3.8ghz instead of 3.5 on mine, and that carries on to turbo modding as well, as he has a higher headroom. his scores will be consistently 5-8~ percent higher than mine on single cpu. The proc also costs much more as well, no bargain basement there. So if you carry this to dual socket, the difference is what you stated. Dont forget he also overclocks his bclk 100-> 104, not possible on my workstation board. that nets even more core speed. Actually, the dual socket scaling is really good, it is generally exactly twice or better , than single cpu.

Edit: There is a choice to use either NUMA mode, or Simultaneous multi processing. I chose NUMA in bios, in order to have better latencies in games and access to all memory quicker by both cpus. It is possible that without numa, dual cpus may provide more consistent scaling in benchmarks.

 

[kniaugaudiskis]

Dont forget he also overclocks his bclk 100-> 104, not possible on my workstation board. that nets even more core speed.

Nice to see a fellow Haswell Xeon user here Mine is actually at 101MHz BLCK, going any higher causes a slight crackling in my bluetooth headphones, and if I go over 102MHz my Samsung 960EVO 1TB M.2 SSD goes from PCI-E 3.0 to 2.0 mode.

The CPU does 3.33GHz on all cores and 3.83GHz on a single core.

 

[ruff0r]

Custom Water loop. Beating an 4770k not bad.

 

[MrGenius]

Beating an 4770k not bad.

No K. But still not bad.

EDIT: Well...excuse me! It is a K. But running stock. So...pretty much not a K.

 

[cdawall]

Going to work on tweaking 4.22ghz stable for this bench I think power draw might be too high for the board.

4.09ghz, 3066 quad channel ram, custom loop

 

[er557]

That overclock has net you a 30% performance gain on the xeons with this benchmark, which come to think about it, is the exact clock frequency difference, as the intel xeons do 3.0ghz on all cores when multi threaded stressed(with mod active). You do need to watch the power draw from the board, even though asrock boards have excellent vrm's and mosfets. If your psu has a windows utility to watch power draw that would be helpful, or the actual tdp in say hwinfo64.
I for instance cant do that, as the TDP limit being hacked , is falsely reported to the system as being 50w at all times (same principle as powercut).