What's new

Resolve Standard Candle Benchmark

I'd love to know what kind of results people are getting on the Standard Candle test these days. What kind of performance do the new MacPros offer? How about the current crop of GPUs? Anyone want to share? I'll start, with my slightly aging but still capable setup:

MacPro Early 2012
2.66GHz 12-core
32GB RAM
3 x GTX680 4GB (1 Internal, 2 in Cyclone expander)

SC Results:

09 Blur Nodes - 24fps
18 Blur Nodes - 24fps
30 Blur Nodes - 17fps
66 Blur Nodes - 8fps

01 TNR Node - 24fps
02 TNR Nodes - 24fps
04 TNR Nodes - 12fps
06 TNR Nodes - 8fps

By the way, the footage file required by the Standard Candle test, "Tutorial - Tracking.mov", no longer comes bundled with the Resolve 11 installer. It is only present in the v10 and earlier installers.
 
I've been commissioned to write a technical article on testing various GPU's 'horsepower' with Resolve and this 'Standard Candle' test looks ideal, except for the test footage not being readily available. Does anyone have tips or a recent update?
 
yes, thanks I have that but the essential video test file is no longer packed with the Resolve 11.xx installer. Fortunately, a user on the BMD forum sent me a copy. Now, I have the .drp file downloaded from Sascha's website but I cannot find any 'versions' included in the project Folder.

I would appreciate a 'walk through' from anybody who has actually used this test with Resolve 11 (Lite) on a PC
 
@Craig just right+click the clip in the color page and you can see the versions listed

Would you be able to upload the file for me/us? I tried the link on the BMD forum and it's no longer working (it seems)
 
Thanks guys. I finally have it working - after downloading both .drp and .mov files, import the .drp Project file first then 'right click' over the blank file in the Resolve media pool to locate where the 'Tutorial - Tracking' file is stored.
 
MacPro (Late 2013)
3.7Ghz QuadCore
32GB RAM
Two AMD FirePro D500's w/ 3GB of VRAM each
Mac OS X 10.10.3
Davinci Resolve 11.3

Results:
09 Blur @ 22fps
18 Blur @ 14fps
30 Blur @ 8fps
60 Blur @ 4fps

1 TNR @ 24fps
2 TNR @ 24fps
3 TNR @ 24fps
4 TNR @ 24fps
5 TNR @ 24fps
6 TNR @ 24fps
 
Robert Olding said:
MacPro (Late 2013)
3.7Ghz QuadCore
32GB RAM
Two AMD FirePro D500's w/ 3GB of VRAM each
Mac OS X 10.10.3
Davinci Resolve 11.3

Results:
09 Blur @ 22fps
18 Blur @ 14fps
30 Blur @ 8fps
60 Blur @ 4fps

1 TNR @ 24fps
2 TNR @ 24fps
3 TNR @ 24fps
4 TNR @ 24fps
5 TNR @ 24fps
6 TNR @ 24fps
Click to expand...

Are you using lite?
 
Hmm, just ran this on my nMP (8 core, dual D700's, 16GB or RAM), and everything reported 24FPS. Thing is, the playback definitely wasn't 24FPS for the more difficult nodes. Anyone have any idea why it's reporting a faster FPS than it's playing?
 
Nick Pasquariello said:
Hmm, just ran this on my nMP (8 core, dual D700's, 16GB or RAM), and everything reported 24FPS. Thing is, the playback definitely wasn't 24FPS for the more difficult nodes. Anyone have any idea why it's reporting a faster FPS than it's playing?
Click to expand...
You're not using an external monitor (interface card). You've got caching enabled?
 
That would do it. I had my monitor packed away. Works now that I've pulled it out. So, what I would consider a middle-of-the-road grading machine:

new Mac Pro
8-Core 3GHz Intel Xeon E5
16GB 1866 MHz DDR3 ECC RAM
Dual AMD D700 GPUs with 6GB VRAM each
Media stored on internal PCIe SSD (read speed around 950MB/s)
Running 10.10.3 and Resolve 11.3

Results:
09 Blur @ 24fps
18 Blur @ 19fps
30 Blur @ 12fps
60 Blur @ 6fps

1 TNR @ 24fps
2 TNR @ 14fps
3 TNR @ 9fps
4 TNR @ 7fps
5 TNR @ 6fps
6 TNR @ 5fps
 
Home brew z87 'Haswell' 3.6Ghz 4 Core i7 PC, 16GB RAM, 2 x 256GB SSDs Striped RAID 0 under Win 7 64bit OS

Resolve 'lite' 11.2 with 1 x Maxwell GTX 970 4GB G1:

11 Blur @ 24fps
13 Blur @ 19fps
15 Blur @ 17fps
18 Blur @ 14fps
30 Blur @ 09fps
 
I have to wonder if the D700's do still have an issue with Resolve? I was surprised to see my home box beat the nMacPro in TNR rendering. The processor is almost 5 years old and the ram isn't as fast and while the Titan is fast, I would think TWO D700s would be faster. Not complaining, but I was surprised. In a week I'll have an updated test of my PC with the new 980 Ti. Should help people decide if the upgrade is worth it...but I don't think this particular test will support that.


nMPro (10.10.4)
8Core/32GB/Dual D700
Media on Pegasus R6 TB2 RAID
Resolve 11.3.1

9 Blur - 24
18 Blur - 16.5
30 Blur - 11
66 Blur - 5.5

1 TNR - 24
2 TNR - 13
4 TNR - 7
6 TNR - 4
_______________________________

Windows 7 Pro
i7 990X
24GB/EVGA Titan 6GB Superclocked
Media on internal 4 disk RAID
Resolve 11.3.1

9 Blur - 24
18 Blur - 14
30 Blur - 9
66 Blur - 4

1 TNR - 24
2 TNR - 21
4 TNR - 11
6 TNR - 7.5
 
Early 2008 Mac Pro
Dual 4-Core 3.2GHz Intel Xeon
32GB 667 MHz DDR2 RAM
NVIDIA GTX 760 GPU with 2GB RAM
Media stored on internal SSD (read speed around 250MB/s)
Running 10.9.5 and Resolve 11.3

Results:
09 Blur @ 15fps
18 Blur @ 9fps
30 Blur @ 5fps
66 Blur @ 2fps

1 TNR @ 19fps
2 TNR @ 12fps
4 TNR -Your GPU memory is full
6 TNR -Your GPU memory is full

This is not surprising. It might be time to consider a nMPro .o_O
 
I wish I could edit/add to my post to save from adding a new one...but here we are. Received my new 980Ti and put it to the test.

Windows 7 Pro
i7 990X
24GB
Media on internal 4 disk RAID
Resolve 11.3.1
Same Nvidia driver use for both tests.

Old EVGA Titan 6GB Superclocked / New EVGA Superclocked ACX2 980 Ti
9 Blur - 24 / 24
18 Blur - 14 / 20
30 Blur - 9 / 13
66 Blur - 4 / 6

1 TNR - 24 / 24
2 TNR - 21 / 24
4 TNR - 11 / 14
6 TNR - 7.5 / 9

So I'm now seeing better performance on my home PC than my work nMacPro. With dual D700s. Which is still a bit of a surprise. My work nMacPro cost $8900. I built my PC for $3200. Adding the $1K for the 980Ti (cdn) I'm still half the price of my trashcan. Fairly pleased about that. However, having said that...I just paid that $1k to gain 4-6 fps in playback performance inResolve! HA (24-35% boost actually) At least my machine will playback 4K raw footage at full res debayer now. That was really my goal.

Heading out on a long weekend trip, but when I get back, I'm going to test using both cards installed. Provided I can figure out the power distribution on my power supply.
 
You must log in or register to reply here.
Back
Top