GPU speed database

questions about practical use of Neat Video, examples of use
jpsdr
Posts: 221
Joined: Mon Aug 11, 2008 7:33 am

Post by jpsdr »

On my PC i have a 720 nVidia card.

I didn't expect anything from it, but i've a surprising result.

I've perfomed performance tests using the feature include in NV.
With NV4, using the GPU produce a slight benefit, but with NV3, using GPU produce only negative effect.

Is there a (simple) explaination ?
NVTeam
Posts: 2745
Joined: Thu Sep 01, 2005 4:12 pm
Contact:

Post by NVTeam »

There were many changes in NV4 algorithms, both optimizations and new functionality (read additional computations) added, so some variations are to be expected.

Vlad
JDW
Posts: 17
Joined: Mon Sep 28, 2015 7:12 am

Post by JDW »

Wow. An utter dearth of results for v4.0 of Neat Video. But the bigger problem is, I cannot see what computer is being used. I consider that information important. Just telling me it is OS X is too vague. Can you please add specifics about the computer? Knowing the clock speed of the CPU would be nice too.

I was also shocked to see that the D300 (which is the video card used in the base model Mac Pro) is not "pro" at all compared to the R 290, which is used in the base model iMac. Are dual D300's really that slow, or is Neat Video merely optimized for the single GPU chip video cards in the iMac, rather than the expensive "pro" models? :-)

I'm also surprised to see no M295X speed profiles. A lot of iMac users doing video work I know chose the 295 over the 290.

Maybe you folks at Neat Video should knock of 5% for Neat Video in exchange for Speed Test results? Either there are few of us who purchased Neat Video, or the folks who did purchase it are not spending time on the benchmarks. As someone planning to buy a new Mac, knowing which model and what GPU based would accelerate my video work would be very, very helpful.

Thanks.
JDW
Posts: 17
Joined: Mon Sep 28, 2015 7:12 am

Always "Invalid"!!!

Post by JDW »

What is wrong with your database at the following URL?

fifonik.com/nv/

I downloaded NeatBench 64-bit for my Yosemite iMac and ran it in the Terminal, but pasting the results into the "NeatVideo Optimize log" field and then clicking "Save" yields an invalid error. I then open FCPX and open Preferences > Performance and then click Optimize Settings and then Copy to Clipboard, but pasting that info in doesn't work either! I still get an invalid error!

I also see that the NeatVideo benchmark running in the Terminal yields slightly different results than doing the same in FCPX. Why?

Anyway, here is my information so you can post it to your DB manually:

GPU Brand: AMD (actually ATI)

GPU model: ATI Radeon HD 4850 512MB (which is sadly unsupported)

NeatVideo version: 4.0

NeatVideo optimize log:

Frame Size: 1920x1080 progressive
Bitdepth: 32 bits per channel
Temporal Filter: Enabled
Radius: 2 frames
Dust and Scratches: Disabled
Neat Video 4.0.9 Pro plug-in for Final Cut

Detecting the best combination of performance settings:
running the test data set on up to 8 CPU cores

1 core: 0.647 frames/sec
2 cores: 1.25 frames/sec
3 cores: 1.58 frames/sec
4 cores: 1.93 frames/sec
5 cores: 1.99 frames/sec
6 cores: 2.07 frames/sec
7 cores: 2.13 frames/sec
8 cores: 2.13 frames/sec

Best combination: 7 cores

Your Name: JDW
System Info: 2.8GHz Intel Core i7, 16GB RAM, Yosemite 10.10.5

Now please fix your online DB so that when I buy a new computer and profile it I can easily enter the DB info myself without the silly INVALID error.

Thanks.
NVTeam
Posts: 2745
Joined: Thu Sep 01, 2005 4:12 pm
Contact:

Post by NVTeam »

> What is wrong with your database at the following URL?

Please note that the database is supported by a third party who is doing that in their spare time. I believe fifonik simply didn't yet have a chance to update the scripts there. Please be kind to your fellow forum member.

Thank you for posting your Optimize results here. It is normal that the speed measured inside a host application (FCPX in this case) may be lower than the speed measured by Neat Bench, because Neat Bench works alone, without any interference from the host application. Also, Neat Bench may run the test in another bitdepth (like 8-bit vs 32-bit in FCPX), which makes a difference as well.

The speeds seem not very high for an i7 CPU, probably because it is earlier model from several years ago. The clock speed (2.8) plays a role too. I guess the memory speed is also not highest possible.

For comparison, a similar test from a recent MacPro:
CPU only (1 core): 1.45 frames/sec
CPU only (2 cores): 2.88 frames/sec
CPU only (3 cores): 4 frames/sec
CPU only (4 cores): 5.21 frames/sec
CPU only (5 cores): 6.41 frames/sec
CPU only (6 cores): 7.52 frames/sec
CPU only (7 cores): 8.55 frames/sec
CPU only (8 cores): 9.43 frames/sec
CPU only (9 cores): 10.4 frames/sec
CPU only (10 cores): 10.9 frames/sec
CPU only (11 cores): 11.5 frames/sec
CPU only (12 cores): 11.4 frames/sec
CPU only (13 cores): 11.8 frames/sec
CPU only (14 cores): 11.4 frames/sec
CPU only (15 cores): 11.5 frames/sec
CPU only (16 cores): 11.4 frames/sec
CPU only (17 cores): 11.5 frames/sec
CPU only (18 cores): 11.1 frames/sec
CPU only (19 cores): 11 frames/sec
CPU only (20 cores): 10.9 frames/sec
CPU only (21 cores): 10.8 frames/sec
CPU only (22 cores): 10.4 frames/sec
CPU only (23 cores): 10.4 frames/sec
CPU only (24 cores): 10.1 frames/sec
As you can see, the new hardware is about 4 times faster even when using only 8 cores. So the lower speeds you observe on your machine probably have to do with hardware capabilities and there is a way to address that.

Regards,
Vlad
JDW
Posts: 17
Joined: Mon Sep 28, 2015 7:12 am

Post by JDW »

I gained access to a Skylake, Oct. 2015 iMac 5K with high-end specs: 4GHz i7 QuadCore CPU, M395 GPU (2GB), and 1TB Flash storage. Here are the results from NeatBench (Neat Video 4.0.9) run via the Terminal:


Looking for NVIDIA CUDA-capable devices...
Failed to load CUDA driver ("/usr/local/cuda/lib/libcuda.dylib")
If you use an NVIDIA card, please install the latest CUDA driver from NVIDIA.

Looking for AMD OpenCL-capable devices...
OpenCL driver version: 20150921.192411
OpenCL initialized successfully.
Checking OpenCL GPU #1:
GPU device name is: AMD Radeon R9 M395 Compute Engine
2048 MB available during initialization
This device is not supported
Check failed - will not use the device


Neat Video benchmark:

Frame Size: 1920x1080 progressive
Bitdepth: 8 bits per channel
Quality Mode: Normal
Mix with Original: Disabled
Temporal Filter: Enabled
Radius: 2 frames
Dust and Scratches: Disabled
Slow Shutter: Disabled
Spatial Filter: Enabled
Frequencies: High, Mid, Low
Artifact Removal: Enabled
Edge Smoothing: Disabled
Sharpening: Disabled

Detecting the best combination of performance settings:
running the test data set on up to 8 CPU cores

1 core: 2.09 frames/sec
2 cores: 4.1 frames/sec
3 cores: 5.81 frames/sec
4 cores: 7.35 frames/sec
5 cores: 7.3 frames/sec
6 cores: 7.19 frames/sec
7 cores: 6.94 frames/sec
8 cores: 6.58 frames/sec

Best combination: 4 cores

- - - - - - - - - - - - - - - - -

I then ran it a second time and it gave me these results:

1 core: 2.13 frames/sec
2 cores: 4.08 frames/sec
3 cores: 5.71 frames/sec
4 cores: 7.25 frames/sec
5 cores: 7.35 frames/sec
6 cores: 7.19 frames/sec
7 cores: 6.94 frames/sec
8 cores: 6.41 frames/sec

Best combination: 5 cores

- - - - - - - - - - - - - - - - -

A third run gave me this:

Detecting the best combination of performance settings:
running the test data set on up to 8 CPU cores

1 core: 2.11 frames/sec
2 cores: 4.15 frames/sec
3 cores: 5.81 frames/sec
4 cores: 7.3 frames/sec
5 cores: 7.3 frames/sec
6 cores: 7.14 frames/sec
7 cores: 6.9 frames/sec
8 cores: 6.45 frames/sec

Best combination: 4 cores

- - - - - - - - - - - - - - - - -

Running it from within FCPX 3 times yielded more consistent speeds, albeit a bit slower than the Terminal:

Frame Size: 1920x1080 progressive
Bitdepth: 8 bits per channel
Quality Mode: Normal
Mix with Original: Disabled
Temporal Filter: Enabled
Radius: 2 frames
Dust and Scratches: Disabled
Slow Shutter: Disabled
Spatial Filter: Enabled
Frequencies: High, Mid, Low
Artifact Removal: Enabled
Edge Smoothing: Disabled
Sharpening: Disabled
Neat Video 4.0.9 Pro plug-in for Final Cut

Detecting the best combination of performance settings:
running the test data set on up to 8 CPU cores

1 core: 2.07 frames/sec
2 cores: 4.1 frames/sec
3 cores: 5.75 frames/sec
4 cores: 7.04 frames/sec
5 cores: 7.09 frames/sec
6 cores: 6.9 frames/sec
7 cores: 6.62 frames/sec
8 cores: 6.25 frames/sec

Best combination: 5 cores

- - - - - - - - - - - - - - - - -

Within FCPX, even though "Enable AMD/ATI GPU Support" is check marked, it says "Check failed" and will use only CPU. So it would be nice to see the M395 and M395X supported.
NVTeam
Posts: 2745
Joined: Thu Sep 01, 2005 4:12 pm
Contact:

Post by NVTeam »

Thank you for reporting. The CPU-based results seem adequate. Regaring the GPU support, we will likely add it in the next update of Neat Video. Once the update becomes available, please re-run the test and also post the results in this thread.

Thank you very much.

Vlad
JDW
Posts: 17
Joined: Mon Sep 28, 2015 7:12 am

Post by JDW »

NVTeam wrote:Regaring the GPU support, we will likely add it in the next update of Neat Video. Once the update becomes available, please re-run the test and also post the results in this thread.
I'm sorry for my delayed reply, but I only today visited your website and noticed NeatVideo had been updated. I discovered that your Terminal app "NeatBench" does not recognize my late 2015 5K iMac17,1's GPU. Here is what I did:

1) Downloaded NeatBench again from this page:

https://www.neatvideo.com/nvforum/viewtopic.php?t=784

(OS X 64-bit)

2) Installed in my Applications folder and then launched it, which launches in the Terminal.

3) The Terminal responded as follows:

Code: Select all

Last login: Mon Dec 21 16:13:19 on ttys001
iMacJDW:~ james$ /Applications/Neat\ Video\ v4\ for\ Final\ Cut/NeatBench ; exit;
Neat Bench (Neat Image 7.6.1, Neat Video 3.6.1)
Copyright (c) 1999-2014 Neat Image team, Neat Video team, ABSoft.
All Rights Reserved.

Log will be saved to /Users/james/NeatBenchLog 2015-12-21 16-17-53.txt


GPU detection log:

Looking for NVIDIA CUDA-capable devices...
Failed to load CUDA driver ("/usr/local/cuda/lib/libcuda.dylib")
If you use an NVIDIA card, please install the latest CUDA driver from NVIDIA.

Looking for AMD OpenCL-capable devices...
OpenCL initialized successfully.
Checking OpenCL GPU #1:
GPU device name is: ATI Radeon R9 M395X Compute Engine
This device is not supported
Check failed - will not use the device


Neat Video benchmark:

Frame: 1920x1080 progressive, 8 bits per channel, Radius: 1 frame
Running the test data set on up to 8 CPU cores

1 core: 2.28 frames/sec
2 cores: 4.59 frames/sec
3 cores: 6.49 frames/sec
4 cores: 8 frames/sec
5 cores: 8.26 frames/sec
6 cores: 8.26 frames/sec
7 cores: 8.06 frames/sec
8 cores: 7.14 frames/sec

Best combination: 5 cores


Log has been saved to /Users/james/NeatBenchLog 2015-12-21 16-17-53.txt

Press Enter to exit
HOWEVER...

When I opened FCPX and the Neat Video Prefs (Performance tab), I was able to get the benchmarks and performance optimization to work, as you can see here:

Code: Select all

Frame Size:  1920x1080 progressive
Bitdepth:  8 bits per channel
Quality Mode:  Normal
Mix with Original:  Disabled
Temporal Filter:  Enabled
    Radius:  2 frames
    Dust and Scratches:  Disabled
    Slow Shutter:  Disabled
Spatial Filter:  Enabled
    Frequencies:  High, Mid, Low
    Artifact Removal:  Enabled
    Edge Smoothing:  Disabled
    Sharpening:  Disabled
Neat Video 4.1.0 Pro plug-in for Final Cut

Detecting the best combination of performance settings:
running the test data set on up to 8 CPU cores and on up to 1 GPU
AMD Radeon R9 M395X Compute Engine: 4096 MB currently available, using up to 70%

CPU only (1 core): 2.03 frames/sec
CPU only (2 cores): 4.05 frames/sec
CPU only (3 cores): 5.65 frames/sec
CPU only (4 cores): 7.04 frames/sec
CPU only (5 cores): 7.09 frames/sec
CPU only (6 cores): 6.94 frames/sec
CPU only (7 cores): 6.8 frames/sec
CPU only (8 cores): 6.58 frames/sec
GPU only (AMD Radeon R9 M395X Compute Engine): 6.62 frames/sec
CPU (1 core) and GPU (AMD Radeon R9 M395X Compute Engine): 5.92 frames/sec
CPU (2 cores) and GPU (AMD Radeon R9 M395X Compute Engine): 5.78 frames/sec
CPU (3 cores) and GPU (AMD Radeon R9 M395X Compute Engine): 7.81 frames/sec
CPU (4 cores) and GPU (AMD Radeon R9 M395X Compute Engine): 8.7 frames/sec
CPU (5 cores) and GPU (AMD Radeon R9 M395X Compute Engine): 10.1 frames/sec
CPU (6 cores) and GPU (AMD Radeon R9 M395X Compute Engine): 10 frames/sec
CPU (7 cores) and GPU (AMD Radeon R9 M395X Compute Engine): 8.7 frames/sec
CPU (8 cores) and GPU (AMD Radeon R9 M395X Compute Engine): 8.93 frames/sec

Best combination: CPU (5 cores) and GPU (AMD Radeon R9 M395X Compute Engine)
So I can now achieve 10.1 fps on the late 2015 iMac with M395X GPU — a noticeable improvement.

Thank you!
Tim
Posts: 20
Joined: Mon Jun 18, 2012 12:54 am
Contact:

Post by Tim »

... but I only today visited your website and noticed NeatVideo had been updated. I discovered that your Terminal app "NeatBench" does not recognize my late 2015 5K iMac17,1's GPU
Neat Bench has been updated too. Please download the new version from https://www.neatvideo.com/download.html (at the bottom of the left column).

Kind regards,
Tim
Image Image Neat Video team
noise reduction for video and photos
jpsdr
Posts: 221
Joined: Mon Aug 11, 2008 7:33 am

Post by jpsdr »

CPU : 10 cores broadwell overclocked at 4.1GHz.
Memory : DDR4-3000.

Frame Size: 1920x1080 progressive
Quality Mode: High
Mix with Original: Disabled
Temporal Filter: Enabled
Radius: 3 frames
Dust and Scratches: Disabled
Slow Shutter: Disabled
Spatial Filter: Enabled
Frequencies: High, Mid, Low, Very Low, Ultra Low
Artifact Removal: Disabled
Edge Smoothing: Disabled
Sharpening: Enabled
Neat Video 4.2.3 Pro plug-in for VirtualDub

Run 1 :

Code: Select all

Detecting the best combination of performance settings:
running the test data set on up to 20 CPU cores and on up to 1 GPU
GeForce GTX 1060 3GB: 2698 MB currently available (3072 MB total), using up to 70%

CPU only (1 core): 0.417 frames/sec
CPU only (2 cores): 0.838 frames/sec
CPU only (3 cores): 1.24 frames/sec
CPU only (4 cores): 1.62 frames/sec
CPU only (5 cores): 2 frames/sec
CPU only (6 cores): 2.37 frames/sec
CPU only (7 cores): 2.67 frames/sec
CPU only (8 cores): 2.99 frames/sec
CPU only (9 cores): 3.25 frames/sec
CPU only (10 cores): 3.52 frames/sec
CPU only (11 cores): 3.52 frames/sec
CPU only (12 cores): 3.62 frames/sec
CPU only (13 cores): 3.64 frames/sec
CPU only (14 cores): 3.68 frames/sec
CPU only (15 cores): 3.69 frames/sec
CPU only (16 cores): 3.7 frames/sec
CPU only (17 cores): 3.72 frames/sec
CPU only (18 cores): 3.66 frames/sec
CPU only (19 cores): 3.69 frames/sec
CPU only (20 cores): 3.72 frames/sec
GPU only (GeForce GTX 1060 3GB): 4.05 frames/sec
CPU (1 core) and GPU (GeForce GTX 1060 3GB): 3.73 frames/sec
CPU (2 cores) and GPU (GeForce GTX 1060 3GB): 3.27 frames/sec
CPU (3 cores) and GPU (GeForce GTX 1060 3GB): 3.8 frames/sec
CPU (4 cores) and GPU (GeForce GTX 1060 3GB): 4.13 frames/sec
CPU (5 cores) and GPU (GeForce GTX 1060 3GB): 4.76 frames/sec
CPU (6 cores) and GPU (GeForce GTX 1060 3GB): 5 frames/sec
CPU (7 cores) and GPU (GeForce GTX 1060 3GB): 5.15 frames/sec
CPU (8 cores) and GPU (GeForce GTX 1060 3GB): 5.56 frames/sec
CPU (9 cores) and GPU (GeForce GTX 1060 3GB): 5.78 frames/sec
CPU (10 cores) and GPU (GeForce GTX 1060 3GB): 5.85 frames/sec
CPU (11 cores) and GPU (GeForce GTX 1060 3GB): 6.13 frames/sec
CPU (12 cores) and GPU (GeForce GTX 1060 3GB): 6.02 frames/sec
CPU (13 cores) and GPU (GeForce GTX 1060 3GB): 6.06 frames/sec
CPU (14 cores) and GPU (GeForce GTX 1060 3GB): 5.99 frames/sec
CPU (15 cores) and GPU (GeForce GTX 1060 3GB): 6.02 frames/sec
CPU (16 cores) and GPU (GeForce GTX 1060 3GB): 6.06 frames/sec
CPU (17 cores) and GPU (GeForce GTX 1060 3GB): 5.24 frames/sec
CPU (18 cores) and GPU (GeForce GTX 1060 3GB): 5.95 frames/sec
CPU (19 cores) and GPU (GeForce GTX 1060 3GB): 5.92 frames/sec
CPU (20 cores) and GPU (GeForce GTX 1060 3GB): 5.95 frames/sec

Best combination: CPU (11 cores) and GPU (GeForce GTX 1060 3GB)
Run 2

Code: Select all

CPU only (1 core): 0.416 frames/sec
CPU only (2 cores): 0.836 frames/sec
CPU only (3 cores): 1.23 frames/sec
CPU only (4 cores): 1.62 frames/sec
CPU only (5 cores): 2 frames/sec
CPU only (6 cores): 2.38 frames/sec
CPU only (7 cores): 2.67 frames/sec
CPU only (8 cores): 2.99 frames/sec
CPU only (9 cores): 3.26 frames/sec
CPU only (10 cores): 3.5 frames/sec
CPU only (11 cores): 3.53 frames/sec
CPU only (12 cores): 3.6 frames/sec
CPU only (13 cores): 3.62 frames/sec
CPU only (14 cores): 3.64 frames/sec
CPU only (15 cores): 3.72 frames/sec
CPU only (16 cores): 3.69 frames/sec
CPU only (17 cores): 3.65 frames/sec
CPU only (18 cores): 3.66 frames/sec
CPU only (19 cores): 3.7 frames/sec
CPU only (20 cores): 3.66 frames/sec
GPU only (GeForce GTX 1060 3GB): 4.07 frames/sec
CPU (1 core) and GPU (GeForce GTX 1060 3GB): 3.75 frames/sec
CPU (2 cores) and GPU (GeForce GTX 1060 3GB): 3.27 frames/sec
CPU (3 cores) and GPU (GeForce GTX 1060 3GB): 3.8 frames/sec
CPU (4 cores) and GPU (GeForce GTX 1060 3GB): 4.15 frames/sec
CPU (5 cores) and GPU (GeForce GTX 1060 3GB): 4.78 frames/sec
CPU (6 cores) and GPU (GeForce GTX 1060 3GB): 5 frames/sec
CPU (7 cores) and GPU (GeForce GTX 1060 3GB): 5.15 frames/sec
CPU (8 cores) and GPU (GeForce GTX 1060 3GB): 5.59 frames/sec
CPU (9 cores) and GPU (GeForce GTX 1060 3GB): 5.81 frames/sec
CPU (10 cores) and GPU (GeForce GTX 1060 3GB): 6.02 frames/sec
CPU (11 cores) and GPU (GeForce GTX 1060 3GB): 5.92 frames/sec
CPU (12 cores) and GPU (GeForce GTX 1060 3GB): 5.52 frames/sec
CPU (13 cores) and GPU (GeForce GTX 1060 3GB): 5.85 frames/sec
CPU (14 cores) and GPU (GeForce GTX 1060 3GB): 6.1 frames/sec
CPU (15 cores) and GPU (GeForce GTX 1060 3GB): 6.1 frames/sec
CPU (16 cores) and GPU (GeForce GTX 1060 3GB): 6.06 frames/sec
CPU (17 cores) and GPU (GeForce GTX 1060 3GB): 6.1 frames/sec
CPU (18 cores) and GPU (GeForce GTX 1060 3GB): 5.99 frames/sec
CPU (19 cores) and GPU (GeForce GTX 1060 3GB): 5.92 frames/sec
CPU (20 cores) and GPU (GeForce GTX 1060 3GB): 5.95 frames/sec

Best combination: CPU (14 cores) and GPU (GeForce GTX 1060 3GB)
Is it normal to have such differences ?
Last edited by jpsdr on Thu Dec 15, 2016 8:53 am, edited 1 time in total.
NVTeam
Posts: 2745
Joined: Thu Sep 01, 2005 4:12 pm
Contact:

Post by NVTeam »

Yes, that is normal.

Could you also run NeatBench please?
It would run the filter with default filter settings making those easier to compare with other setups.

Thank you,
Vlad
jpsdr
Posts: 221
Joined: Mon Aug 11, 2008 7:33 am

Post by jpsdr »

Result with NeatBench

Code: Select all

CPU only (1 core): 1.93 frames/sec
CPU only (2 cores): 3.98 frames/sec
CPU only (3 cores): 5.88 frames/sec
CPU only (4 cores): 7.81 frames/sec
CPU only (5 cores): 9.62 frames/sec
CPU only (6 cores): 11.4 frames/sec
CPU only (7 cores): 13 frames/sec
CPU only (8 cores): 14.5 frames/sec
CPU only (9 cores): 15.9 frames/sec
CPU only (10 cores): 16.9 frames/sec
CPU only (11 cores): 16.9 frames/sec
CPU only (12 cores): 17.5 frames/sec
CPU only (13 cores): 17.5 frames/sec
CPU only (14 cores): 17.2 frames/sec
CPU only (15 cores): 17.5 frames/sec
CPU only (16 cores): 17.2 frames/sec
CPU only (17 cores): 16.9 frames/sec
CPU only (18 cores): 16.7 frames/sec
CPU only (19 cores): 15.6 frames/sec
CPU only (20 cores): 15.6 frames/sec
GPU only (GeForce GTX 1060 3GB): 10.6 frames/sec
CPU (1 core) and GPU (GeForce GTX 1060 3GB): 9.43 frames/sec
CPU (2 cores) and GPU (GeForce GTX 1060 3GB): 11.4 frames/sec
CPU (3 cores) and GPU (GeForce GTX 1060 3GB): 13.3 frames/sec
CPU (4 cores) and GPU (GeForce GTX 1060 3GB): 14.9 frames/sec
CPU (5 cores) and GPU (GeForce GTX 1060 3GB): 16.9 frames/sec
CPU (6 cores) and GPU (GeForce GTX 1060 3GB): 17.9 frames/sec
CPU (7 cores) and GPU (GeForce GTX 1060 3GB): 19.2 frames/sec
CPU (8 cores) and GPU (GeForce GTX 1060 3GB): 21.3 frames/sec
CPU (9 cores) and GPU (GeForce GTX 1060 3GB): 21.3 frames/sec
CPU (10 cores) and GPU (GeForce GTX 1060 3GB): 23.3 frames/sec
CPU (11 cores) and GPU (GeForce GTX 1060 3GB): 22.7 frames/sec
CPU (12 cores) and GPU (GeForce GTX 1060 3GB): 23.3 frames/sec
CPU (13 cores) and GPU (GeForce GTX 1060 3GB): 22.7 frames/sec
CPU (14 cores) and GPU (GeForce GTX 1060 3GB): 22.2 frames/sec
CPU (15 cores) and GPU (GeForce GTX 1060 3GB): 22.7 frames/sec
CPU (16 cores) and GPU (GeForce GTX 1060 3GB): 22.7 frames/sec
CPU (17 cores) and GPU (GeForce GTX 1060 3GB): 22.7 frames/sec
CPU (18 cores) and GPU (GeForce GTX 1060 3GB): 22.7 frames/sec
CPU (19 cores) and GPU (GeForce GTX 1060 3GB): 22.2 frames/sec
CPU (20 cores) and GPU (GeForce GTX 1060 3GB): 21.7 frames/sec
NVTeam
Posts: 2745
Joined: Thu Sep 01, 2005 4:12 pm
Contact:

Post by NVTeam »

Thank you!

That looks like the fastest results from any CPU so far.

Vlad
Zach
Posts: 38
Joined: Sat Jun 01, 2013 12:37 pm

Post by Zach »

It really seems like we are hitting a plateau in performance at around 6 - 10 cores though.

From what I can recall of my results, results from my brothers machine (offline) and a few others posted here, we all fall somewhere between 20 - 23 FPS optimal performance. I know my machine peaks around 21fps or so on Neatbench and its only a lowly 6 core / 12 thread w/ a GTX 1070.

I know memory timing has been suggested to be a big performance impact as well. I'm really curious if anyone out there is running DDR4-3000Mhz or higher and what their gains look like vs the 2100 - 2800 Mhz range.
jpsdr
Posts: 221
Joined: Mon Aug 11, 2008 7:33 am

Post by jpsdr »

Zach wrote:It really seems like we are hitting a plateau in performance at around 6 - 10 cores though.
Personnaly, i would say more around 10-16.
Zach wrote: I'm really curious if anyone out there is running DDR4-3000Mhz
To see the real impact, you have to check the CPU only test, because the GPU will "mess" the result. You have very good result, even with a "low" CPU, because of your GPU.
Post Reply