Cinebench R15 Benchmark Thread

[996GT2]

My 2700K @ 5.1 GHz doing 893cb. The 876 run was at 5.0 GHz. Will try a run a 5.2 or 5.3 if I get around to it.

Highest score so far for a 4C/8T system on AT?

 

[Tequila]

i7-4770K @ 4.2ghz, memory @2400mhz

 

[bradly1101]

 

[SlowSpyder]

My 2700K @ 5.1 GHz doing 893cb. The 876 run was at 5.0 GHz. Will try a run a 5.2 or 5.3 if I get around to it.

Highest score so far for a 4C/8T system on AT?

Nice score! In general would you say 2700K's seem to overclock a bit better than 2600K's? 5.2-5.3GHz would be pretty impressive! At 5.017GHz I scored 789... Intel has some pretty good IPC considering you still easily outscored me with a HT quad. I guess I'll just have to go for more clock speed.

 

[MichaelD]

Hmm. Not sure why my post disappeared...some trickery must be afoot. Here it is again.

Not sure why it says "3.30GHz". My CPU is OCed to 4.0GHz via the "Easy Overclock" utility on my Asrock P67 Extreme4 MB.

 

[AtenRa]

Core i7 3770K @ 4.44GHz

 

[rvborgh]

Here's the results from my 2009 vintage dual Opteron 8439 SE rig (2 x 6 K10 cores @ 2.8 ghz):

a shot of the CPU-Z (after i ran wprime32... but no changes hardware wise)

https://www.youtube.com/watch?v=O45Lqg9o6cE

 

[996GT2]

I wanted to try a run on my Bay Trail tablet, but it seems like there is no 32-bit build of R15.

 

[Edgemeal]

I wanted to try a run on my Bay Trail tablet, but it seems like there is no 32-bit build of R15.

CB_R15 is 64-bit only
btw, There is already 1-2 threads for posting CB 11.5 scores, heres one.

BTW,
CB_R15 scores updated.
CB_R15 Multi-CPU scores updated.

 

[Edgemeal]

My 2700K @ 5.1 GHz doing 893cb.

Sweet! :awe:

4C / 8T

 

[Shamrock]

AMD Phenom II 955 BE (stock clock)

326

 

[Bubbleawsome]

i5 4670k@4.4Ghz doing 652.

 

[jhu]

PC in signature. Didnt optimzie the run. All stock.

No AVX, FMA or AVX2 in R15. Pretty odd.

I don't see why? Using those instructions effectively requires writing the code in such a way as to make it easier for the compiler to use those instructions effectively. 3D image rendering isn't really amenable to this because it's not a very data parallel situation. Image rendering likes moar cores, not SIMD (eg, go grab the sources to povray or luxrender and compile with and without AVX/FMA. You'll find no significant difference in performance between the AVX and SSE2 versions). Image post processing, now that's amenable to both moar cores and SIMD.

 

[SlowSpyder]

Still working on getting the right combo, system is 'mostly' stable at this speed, still tweaking settings. I'm shooting for FlanK3r's score... maybe I can get lucky!

On a side note, at stock settings with Cool & Quiet enabled I idle around 115 watts. With these settings, my Kill-o-Watt shows 491 watts peak during this bench. :biggrin: I'm pretty happy with my cooling set up, but I think it is about maxed.

 

[Edgemeal]

On a side note, at stock settings with Cool & Quiet enabled I idle around 115 watts. With these settings, my Kill-o-Watt shows 491 watts peak during this bench. :biggrin: I'm pretty happy with my cooling set up, but I think it is about maxed.

491 watts dang, thats only 2.23 points per CB_R15 watt.
491-115 = 376W (838/376 = 2.228723404255319)

I knew FX's were inefficient but didn't think it be that bad. I sure hope AMD can come back strong in a couple years ( I was a huge AMD fan from K5 era up to end of Phenom era).

FWIW, tested my i5-3570, 8GB Ram, 1-SSD.
i5-3570K @ 3.61GHz, Idle:43W, CB_R15: 83W.
CB_R15 Usage: 40W (83-43=40).
CB_R15 Score: 511.
CB_R15 PPW: 12.775 (points per watt).

 

[SlowSpyder]

491 watts dang, thats only 2.23 points per CB_R15 watt.
491-115 = 376W (838/376 = 2.228723404255319)

I knew FX's were inefficient but didn't think it be that bad. I sure hope AMD can come back strong in a couple years ( I was a huge AMD fan from K5 era up to end of Phenom era).

FWIW, tested my i5-3570, 8GB Ram, 1-SSD.
i5-3570K @ 3.61GHz, Idle:43W, CB_R15: 83W.
CB_R15 Usage: 40W (83-43=40).
CB_R15 Score: 511.
CB_R15 PPW: 12.775 (points per watt).

503 watts with more NB speed now. I'm really pushing this system... brb, going to do a stock clock/voltage run and see what the Kill-o-Watt says. A good chunk of the idle wattage used is going to a pump and four 200mm fans as well. I'm sure with a good air cooler I'd be well under 100 watts idle.

 

[SlowSpyder]

503 watts with more NB speed now. I'm really pushing this system... brb, going to do a stock clock/voltage run and see what the Kill-o-Watt says. A good chunk of the idle wattage used is going to a pump and four 200mm fans as well. I'm sure with a good air cooler I'd be well under 100 watts idle.

Big difference. Everything set to auto, Cool & Quiet enabled, turbo core enabled, I idle about 120 watts (I never looked with this new motherboard, my MSI was ~115, this ASRock looks to use a few watts more). Load was 320 watts running this bench. Scorred 663 for those 320 watts.

What was crazy was the temp difference. With 1.575v and 24.5 x 218 Coretemp has gone as high as 66C. During the factory setting run, max temp was 33C. (I doubt that is accurate, but it was the same utility on both the factory and overclocked bench, so it should be good enough for this comparrison)

LLC was set at 50% on both runs, I probably should have disabled that, could have saved a few more watts.

 

[jhu]

Still working on getting the right combo, system is 'mostly' stable at this speed, still tweaking settings. I'm shooting for FlanK3r's score... maybe I can get lucky!

On a side note, at stock settings with Cool & Quiet enabled I idle around 115 watts. With these settings, my Kill-o-Watt shows 491 watts peak during this bench. :biggrin: I'm pretty happy with my cooling set up, but I think it is about maxed.

Damn, that's 376W difference! My stock FX8350 is 186W difference between idle and full load. Another 200W for and one extra GHz…

 

[SlowSpyder]

Damn, that's 376W difference! My stock FX8350 is 186W difference between idle and full load. Another 200W for and one extra GHz…

Dropped the voltage down to 1.3625, closer to what an FX8350 would use. Idle is still hanging around 120 watts. Load was 290 watts max I saw during the Cinebench run. Scored 686, never went over 29C according to Coretemp.


CPU's power use scale linearly with frequency at a given voltage, as I understand it (add 5% clock, get 5% more power use). But when you add voltage, power use jumps up in a hurry. I don't plan on running 5.3GHz for my everyday clock, but I want to see what the chip can do in my setup. I'm pretty sure I'll settle on ~5.2GHz at a lower voltage and have a more 'sane' machine.

I can do 4.9GHz on the stock 1.5v, going to lock it at 4.4GHz (factory base frequency) and leave turbo on, see how far I can undervolt... just want to see how far I can take this CPU going the other direction.

*edit - 1.3375v run. 280 watts on the Kill-o-Watt during the Cinebench run. Dropped another 4C according to Coretemp to 25C. Scored 689.

*edit 2 - 1.3250v run 278 watts on the Kill-o-Watt during the bench. Temp stayed the same as the previous, showing 25C. Scored the same 689.

*edit 3 - 1.3000v run. Jumped two settings down in the BIOS to 1.3000. 266 watts on the Kill-o-Watt during this run. Temp maxed at 24C. Scored 692. Kind of surprised to see it still going! Basically I have a 10% overclocked FX8350 here, using a good deal less wattage!

*edit 4 - 1.2875v run. 262 watts on the Kill-o-Watt. 23C max temp. Scored 690.

*edit 5 - 1.2750v run. 258 watts on the Kill-o-Watt. 22C max temp. Scored 687.

*edit 6 - 1.2625v 253 watts on the Kill-o-Watt. 21C max temp. Score a little lower at 679.

*edit 7 - 1.2500v run. 250 watts on the Kill-o-Watt. 21C again. Scored 690. That's ~5.27 points per CB15 watt used. Big improvement. Idle is about the same at ~118 watts, 250 under load in this bench at these settings.

*edit 8 - 1.2375v run. 246 watts max shown. 20C. Scored 689.

*edit 9 - 1.2250v run. 242 watts max on the Kill-o-Watt. 19C. Scored 688.

 

[Bubbleawsome]

I get crazy varied results. Anyone else? My best run was at 4.3Ghz and I can't match it again at 4.5. Some at the same frequency vary up to 50 points.

 

[Edgemeal]

I get crazy varied results. Anyone else? My best run was at 4.3Ghz and I can't match it again at 4.5. Some at the same frequency vary up to 50 points.

Na, my CB_R15 scores don't vary by more then 1 point between runs.
I can only guess you have something running in the background that was doing some work (antivirus, etc) at same time or possibly some CPU throttling due to heat?

 

[SlowSpyder]

I get crazy varied results. Anyone else? My best run was at 4.3Ghz and I can't match it again at 4.5. Some at the same frequency vary up to 50 points.

I just ran it 10 times or so at the same clockspeed, results were pretty consistent. Think you are throttling maybe?

 

[Bubbleawsome]

No throttle, very few programs in the background. They never used more than ~1% of CPU time. :/

 

[Edgemeal]

*edit 9 - 1.2250v run. 242 watts max on the Kill-o-Watt. 19C. Scored 688.

242-120=122W
688/122 = 5.64
Interesting, From 2.23 to 5.64 = 152.91% increase!

Still stable?

 

[SlowSpyder]

242-120=122W
688/122 = 5.64
Interesting, From 2.23 to 5.64 = 152.91% increase!

Still stable?

Stable running that bench, I didn't try any hardcore stability tests though. I was surprised to see all eight cores load up and keep on finishing the bench at lower and lower voltages. I'll probably pick up where I left off sometime tomorrow just to satisfy my curiosity.