Does Edius 9's GPU usage scale?

[Liverpool TV]

With the latest version of Edius 9 and its functions that allow for utilisation of GPUs, such as decoding of certain codecs and PP as a for instance.

Do these functions scale with the processing power of the GPU?

For Nvidia GPUs, what is more beneficial for E9, more Cuda cores or more V-RAM?

Also. Is there a preference to certain GPU types, or indeed do some simply not work? Maybe an Nvidia vs AMD thing?

I've been searching for many answers to questions I've got while trialling E9 but there seems to be some vagueness over certain stuff, some stuff I can't find and with changes as E9 evolves some information may simply be out of date. Hence all the questions.

Cheers,
Dave.

[Ron Evans]

I would like to find out if EDIUS 9 uses GPU's better too. I did not think it uses them any differently for normal decode and encode as EDIUS 8. I think QS being the only one used for timeline performance. One reason for me to stay with 8.53WG at the moment.

There are big timeline performance differences between EDIUS 8.53 and Resolve for instance with my system ( no QS ). Playback with an edited HQX HD file EDIUS resource use is as follows, CPU 11% and GPU 40% for Resolve same HQX file CPU 11% and GPU 12%. For a GH5 UHD 60P source file in a 1920x1080 timeline no effects, straight clip on the timeline, EDIUS CPU 67%, GPU 48% and Resolve, CPU 22%, GPU 19%. Both files of course play realtime through the same IP4K to the same monitor. Still puzzled as to what EDIUS is doing with the GPU as in the case of either the HQX file or the GH5 source file there are no GPU effects engaged it is pure file playback for the HQX file with scaling to 1920x1080 in the case of the GH5 file. Adding colour correction with PCC in EDIUS or nodes in Resolve do not change resource use significantly. Not sure if in EDIUS the GPU is feeding the IP4K somehow as in both the case of EDIUS and Resolve the PC screen refresh will be the same with the on screen monitor set to be the same size.

[Jerry]

With the latest version of Edius 9 and its functions that allow for utilisation of GPUs, such as decoding of certain codecs and PP as a for instance.

Do these functions scale with the processing power of the GPU?

For Nvidia GPUs, what is more beneficial for E9, more Cuda cores or more V-RAM?

Also. Is there a preference to certain GPU types, or indeed do some simply not work? Maybe an Nvidia vs AMD thing?

I've been searching for many answers to questions I've got while trialling E9 but there seems to be some vagueness over certain stuff, some stuff I can't find and with changes as E9 evolves some information may simply be out of date. Hence all the questions.

Cheers,
Dave.

Do these functions scale with the processing power of the GPU?

I can only go by experience on the same system. I went from a 980ti to a 1080ti and I did have improved performance with items that utilized both the added CUDA cores and Vram.
This also included the newer codecs found in the importer/exporter dropdown.

For Nvidia GPUs, what is more beneficial for E9, more Cuda cores or more V-RAM?

If you use a lot of 3rd party programs more CUDA cores will be welcomed.
Also, codecs that are receiving GPU assist would benefit from the cores.
Certain 3rd party plugins will also require higher ram counts to process more complicated images.
I have a lot of CUDA and heavy ram use 3rd party plugins installed. When I went from 6GB to 11GB, there was a marked difference.

For a general work experience a 6GB 1050 would be an example of an entry level card. Personally, I wouldn't install anything less than a 1070(ti). Keep in mind that Edius 9 now has are more codecs receiving GPU assist.


Also. Is there a preference to certain GPU types, or indeed do some simply not work? Maybe an Nvidia vs AMD thing?

Back in the Canopus Rex and DV Rex days, Radeon reigned supreme. Now, Nvidia has the edge. Premiere Pro also prefers Nvidia. Resolve works equally well with Nvidia or AMD.
GeForce models are more than sufficient to handle editing in Edius.

[Liverpool TV]

Do these functions scale with the processing power of the GPU?

I can only go by experience on the same system. I went from a 980ti to a 1080ti and I did have improved performance with items that utilized both the added CUDA cores and Vram.
This also included the newer codecs found in the importer/exporter dropdown.

For Nvidia GPUs, what is more beneficial for E9, more Cuda cores or more V-RAM?

If you use a lot of 3rd party programs more CUDA cores will be welcomed.
Also, codecs that are receiving GPU assist would benefit from the cores.
Certain 3rd party plugins will also require higher ram counts to process more complicated images.
I have a lot of CUDA and heavy ram use 3rd party plugins installed. When I went from 6GB to 11GB, there was a marked difference.

For a general work experience a 6GB 1050 would be an example of an entry level card. Personally, I wouldn't install anything less than a 1070(ti). Keep in mind that Edius 9 now has are more codecs receiving GPU assist.


Also. Is there a preference to certain GPU types, or indeed do some simply not work? Maybe an Nvidia vs AMD thing?

Back in the Canopus Rex and DV Rex days, Radeon reigned supreme. Now, Nvidia has the edge. Premiere Pro also prefers Nvidia. Resolve works equally well with Nvidia or AMD.
GeForce models are more than sufficient to handle editing in Edius.

Thanks for that very informative and descriptive real world info Jerry.

It's reassuring that Edius will scale with the GPU even if it doesn't utilise them extensively, at least for now.

It's all info that will help toward a decision for justifying the possible layout for E9.

Cheers,
Dave.

[trillium]

Those 1080i weigh about 2 lbs don't they?

[Jerry]

To be honest, I've never looked at the weight as a criteria. Length and height are the biggest concerns for some users with smaller cases.

[Liverpool TV]

To be honest, I've never looked at the weight as a criteria. Length and height are the biggest concerns for some users with smaller cases.

Hi Jerry.

I think his weight concern maybe because of the strain on the MB slot.

Some boards have re-enforced slots and obviously depending on the case and card etc. some cards can also be supported from the rear.

With those that have neither, the weight of the cards can become an issue.

Cheers,
Dave.

[Jerry]

There are brackets for stabilizing GPU's. Prices go from $9 up. The $70 carbon fiber versions are nice.

https://www.amazon.com/CloverTale-Ae.../dp/B077LGLN4G

Using a high quality case in the first place reduces the likelihood of sagging and extra costs later.

[Liverpool TV]

Using a high quality case in the first place reduces the likelihood of sagging and extra costs later.

Totally agree, or a MB with a re-enforced socket.