Cinebench R15 Benchmark Thread

[R0H1T]

AMD A8 5550M
Win8.1 x64

 

[Makaveli]

FYI the Durvelle guy who posted results in this thread is a banned AMD viral marketer who doesn't actually own what he claims. Hopefully the benchmarks weren't stolen from anyone here

lol serious is that why I haven't seen him around.

O wait I think I remember scores that were stolen from another forum that he posted as his......

 

[KILLER_K]

Time to get on the new Cinebench R15. Here is a link to it. Sorry for not posting any scores.
Cinebench R15 Download

 

[Galatian]

A few questions? Is it normal that Cinebench won't show my overclock? I got my i5-3570K running at 4,5 GHz, but it still says @3,4 GHz? Also why is my overclocked GTX670 only getting 104 FPS? Are Nvidia cards somehow crippled in OpenGL compared to AMD?

 

[Makaveli]

A few questions? Is it normal that Cinebench won't show my overclock? I got my i5-3570K running at 4,5 GHz, but it still says @3,4 GHz? Also why is my overclocked GTX670 only getting 104 FPS? Are Nvidia cards somehow crippled in OpenGL compared to AMD?

if I do a google search for Cinebench GPU scores I find this.

http://www.c4dcafe.com/ipb/Benchmarks/cinebench.html

Your scores seem to be normal and I don't see it favoring AMD cards.

As for what clockspeed the application shows does it matter?

This application has an issue with it not showing the proper clock speed.

You know what clock speed the system is running at and that is the reason people are posting cpu z shots also.

If your concerned its not running at these clock speeds when you run the benchmark just grab real temp or another application to see if its running full load and at what clock speed.

 

[Enigmoid]

i7-3630qm + 660m @ 1085/1250

 

[Ferzerp]

Did I do this right?

 

[Makaveli]

Did I do this right?

no that score it too low overclock that server please :biggrin:

 

[pw257008]

AMD A8 5550M
Win8.1 x64

I find this one pretty interesting. Don't have the pic because I'm on a different computer (I'll upload it soon enough), but I got a 154 with an A6-5200 clocked at 2Ghz w/o turbo. Single core score was somewhere around 40 or 41, so the multiplier was close to 3.8. So, at least in Cinebench, Jaguar has worse single-core IPC than Richland, but better multi-core IPC (Assuming that 2.8 Ghz held for both tests--if the A8 turboed to 3.1 on the single core test, then single core IPC is pretty much a draw). This gets me wondering what kind of upgrade AMD's next line of Cat cores will be, and if they can scale them up like someone suggested in another topic, akin to Intel with Core 2. Or, on the other hand, if Kaveri can scale performance down the line better than Richland.

BTW, I mean IPC here as a stand-in for performance per Hz, which I know is not totally accurate.

 

[KILLER_K]

Can't never have a thread with out someone photo shopping the image.......jp

Looks great for a server. Very nice score indeed.

Did I do this right?

 

[BonzaiDuck]

I'd never worked with Cinebench before -- I just saw the results of reviews for various hardware.

And since I don't want to fiddle with the screen-capture and whatever I need to do so that I can post it, I'll just report:

I7-2600K 4C/8T @ 4.60 Ghz; DDR3-1600 @ stock speed/timings and CMD=1; eVGA [NVidia] GTX 570 1,284MB in Z68 with PCIE 2.0 x16

OpenGL 103.64 fps
CPU 797 cb
CPU (Single Core) 162 cb
MP Ratio 4.92 x

===========
ADDENDUM & UPDATE:

Same processor, GFX and settings, with DDR3-1600 @ 1866 and CMD=2

OpenGL 108.31
CPU 806 cb
CPU (Single Core) 161 cb
MP Ratio 5.01 x

===========
ADD'L UPDATE:
Same processor @ 4.70 Ghz, GFX and settings, with DDR3-1600 @ stock & CMD=2

OpenGL 108.43
CPU 818 cb
CPU (Single Core) 165 cb
MP Ratio 4.96 x


===========
I'll be interested to see it among the others. I'm a bit puzzled, because R15 reports my processor only as "@ 3.40 Ghz," but the ranking data shows an i7-4770K Haswell at 4.40 Ghz. So . . . . I'm puzzled.

Am I supposed to follow some rules here? Or does the software pick up information about the Haswell that it misses for my Sandy? I'm puzzled . . .

 

[SPBHM]

I find this one pretty interesting. Don't have the pic because I'm on a different computer (I'll upload it soon enough), but I got a 154 with an A6-5200 clocked at 2Ghz w/o turbo. Single core score was somewhere around 40 or 41, so the multiplier was close to 3.8. So, at least in Cinebench, Jaguar has worse single-core IPC than Richland, but better multi-core IPC (Assuming that 2.8 Ghz held for both tests--if the A8 turboed to 3.1 on the single core test, then single core IPC is pretty much a draw). This gets me wondering what kind of upgrade AMD's next line of Cat cores will be, and if they can scale them up like someone suggested in another topic, akin to Intel with Core 2. Or, on the other hand, if Kaveri can scale performance down the line better than Richland.

BTW, I mean IPC here as a stand-in for performance per Hz, which I know is not totally accurate.

MP ratio should be lower for richland because it doesn't have 4 "full cores" like Kabini, and because turbo runs the CPU at higher clock for ST it's even worse.

at the same clock I would expect Richland to beat Kabini for ST but loose for MT.

just for fun, tested my "Core 2 Quad" locked at 2GHz score was 223 for MT and 57 for ST.

 

[pw257008]

MP ratio should be lower for richland because it doesn't have 4 "full cores" like Kabini, and because turbo runs the CPU at higher clock for ST it's even worse.

at the same clock I would expect Richland to beat Kabini for ST but loose for MT.

just for fun, tested my "Core 2 Quad" locked at 2GHz score was 223 for MT and 57 for ST.

I thought Kabini had a similar Module design to the main line?

And Core 2 Quad is great. Nothing pushes me more away from a Pentium or lower than experiencing the longevity of C2Q systems vs. C2D. Of course, I've got a dual core in my laptop (as do my dad and brother--my mom's sadly stuck with a single core HT Atom), but hyper-threading changes the longevity calculation.

 

[pw257008]

In a similar vein, anyone try underclocking their Kaveri A10's yet to get a feel for possible notebook performance?

 

[pw257008]

And my main system:

i5-2410m (2.3 base, 2.9 boost)
ST:95
MT:232
Couldn't do the OpenGL portion (well, the whole racing sim ran through, but the image reproduction test failed, so I guess I get no score)

Since I'm not in Windows, here's a quasi-screenshot (text documents are real trustworthy, I know):

 

[BonzaiDuck]

And my main system:

i5-2410m (2.3 base, 2.9 boost)
ST:95
MT:232
Couldn't do the OpenGL portion (well, the whole racing sim ran through, but the image reproduction test failed, so I guess I get no score)

Since I'm not in Windows, here's a quasi-screenshot (text documents are real trustworthy, I know):

You're so kind . . . so kind . . . :\

 

[Burpo]

Xeon X5650 4.14ghz

 

[Burpo]

Just made the "4200 Club"
Xeon X5650 @ 4.2Ghz
12Gb Patriot Sector 7 RAM 1600 9-9-9-24

 

[BallaTheFeared]

^ Weak, check out my ride!

 

[lopri]

My main rig of 3 years:

1045T @3.90 GHz

My 2nd rig of 3 years:

2500K @4.80 GHz

2500K performs better on this version than it did on 11.5, it seems. They were roughly equal in MP scores.

 

[BonzaiDuck]

2500K performs better on this version than it did on 11.5, it seems. They were roughly equal in MP scores.

I think you OC'd the 2500K to 4.8Ghz? You beat me by six points on "CPU Single Core."

I finally decided to "re-certify" my system @ 4.60Ghz to run at 4.7. Did the preliminary Prime95 sFFT and lFFT today, each for a nominal 4 hours, temperatures at room-ambient 79F show an average for all four cores at ~69C -- load voltage (sFFT) average is 1.34V and lFFT is about 1.345V. The brief unloaded "turbo" voltage idle before settling back to EIST shows as 1.38V.

I posted my results in an addendum to my original post above . . . I think I can beat the i7-4770K @ 4.4Ghz by setting the RAM to DDR3-1866, but -- other things to do first.

[Incidentally, finally took the trouble to read the FAQs about posting images directly, and screen capture is not a problem for me. Next time! For my text results, you can "believe it or not." However, I think a lot of Sandy veterans here would see it as "nothing they didn't know." I'm a cautious over-clocker, Iyam Iyam . . . ]

 

[lopri]

Extra 150 MHz didn't do much for the 1045T. (at least it matches non-HT, 4-core Sandy/Ivy in MP score if clocks are normalized)

I think you OC'd the 2500K to 4.8Ghz?

Yes, my 2500K runs @4.80 GHz.

P.S. If past versions are any indication, Cinebench doesn't respond well to memory (latency and bandwidth).

 

[Tristor]

Sorry for zombifying the thread, but since I finally got my build finalized I wanted to contribute as well.

 

[moonbogg]

3930k @ 4.5 = 1139
gtx 670 = 119

 

[The|Hunter]

Cpu 4.7Ghz 1.284v, cache 4.2Ghz 1.135v, ram oc 2200mhz CL9/1T

4.8Ghz gets single 193cb, had only OC per core that time..

580gtx replacement and from what I saw this Cinebench15 openGL bench really benefits from Haswell's new extensions