Benchmarking

Comments

Grazie wrote on 7/24/2019, 5:15 AM

Not what is generally regarded as overclocking, as I understand it.

@Former user - So what is regarded as OC, generally or otherwise?

Former user wrote on 7/24/2019, 5:36 AM

@Grazie There are people here better placed than myself to answer that for sure, I think that @TheRhino has done some overclocking on his rig. My limited efforts amounts to letting the Asus motherboard Ai do it in the Bios.

If you want to do it manually then you really have to gen up on it. With modern motherboards having Ai its a way to put your toe in the water. You use more wattage, mostly run noisier and maybe not as stable.

@BruceUSA appears to be very good at it, without any reported issues.

It would be best to mainly keep this thread for the benchmarking testing, maybe pursue this topic in a new thread.

 

Grazie wrote on 7/24/2019, 5:45 AM

 

It would be best to mainly keep this thread for the benchmarking testing, maybe pursue this topic in a new thread.

@Former user Sure, but this thread wants us to note our OC, and thusly I wished to hear what that would mean, in direct relation to the OC column.

Former user wrote on 7/24/2019, 6:03 AM

@Grazie I agree entirely with that. I'm simply concerned that if answers to your general query about overclocking expanded into a full blown debate then it might be appropriate to start a new thread.

"Sure, but this thread wants us to note our OC" True and three options mostly cover that … No, Yes, and Partial.  You can expand on this if you feel a need to in the comments column.

Kinvermark wrote on 7/24/2019, 9:13 AM

Overclocking is a BIOS level clock and voltage (etc.) modification from the CPU's stock speeds. Turboboost is not a modification, it is a manufacturer intended stock speed. Thus, NOT an OVERclock.

 

Grazie wrote on 7/24/2019, 9:15 AM

@Kinvermark - Much appreciate your clarification.

Kinvermark wrote on 7/24/2019, 10:57 AM

No, I did not turn off the GPU. That (in my view) would be the wrong way to run the test. It's a render test, so we are testing various types of encoders. I called "CPU" "native" for that reason. There are no circumstances (again, IMO) where the GPU should be disabled - even when running a "native" mainconcept encode - except when the system is malfunctioning. Review frosty's comments for further explanation.

Howard-Vigorita wrote on 7/24/2019, 11:04 AM

@Howard-Vigorita - Hi there! Is Turbo Boost a form of Over-Clocking? If it is, then why am I reading “NO” under OC for your top three?

@Howard-Vigorita, I’m always willing to learn and even try adding some TurboBoost to my rig! 😏

The default for my 8-core Intel i9-9900k is to Turbo Boost to 5 ghz. Boost is not considered overclocking if it's standard stock issue. I lowered my boost rate due to core overheating during stress testing instead of implementing water cooling. Maybe underclocking? Your entry says you got a 10-core i9-9820X. I think the stock rates for that are 3.3 ghz base with 4.1 ghz boost. I suspect setting both rates to be equal might effectively shut off turbo boost but might be preferred if the base rate was overclocked instead. I personally would not choose to go that way myself because it tends to make the machine run hotter and fans louder when it's not doing anything .

BruceUSA wrote on 7/24/2019, 11:29 AM

Some clarify here guys. Your i9 9900k, 8 core with a base clock of 3.6ghz, From the factory, its can boost to max of 4.6ghz on all cores. But its can only boost to 4.8ghz on only 4 core and 1 core at max of 5ghz. When you OC the CPU it overclocking on all cores. Hope this help Grazie understand better. :) I like mine running all 16 cores at 4ghz. no less or more.

Intel i9 Core Ultra 285K Overclocked all P Cores @5.6, all E-Cores @5ghz               

MSI MEG Z890 ACE Gaming Wifi 7 10G Super Lan, thunderbolt 4                                

48GB DDR5 -8200mhz Overclocked @8800mhz                  

Crucial T705 nvme .M2 2TB Gen 5  OS. 4TB  gen 4 storage                    

RTX 5080 16GB  Overclocked 3.1ghz, Memory Bandwidth increased from 960 GB/s to 1152 GB/s                                                            

Custom built hard tube watercooling.                            

MSI PSU 1250W, Windows 11 Pro

 

Former user wrote on 7/25/2019, 6:06 AM

In the interests of transparency, it’s only right to bring existing contributors and potential new ones up to date with developments.

Kinvermark mostly set the ball rolling on this new 4K test, see his thread below, I just did most of the preparation ... https://www.vegascreativesoftware.info/us/forum/feature-request-standardized-4k-test-project--116193/

Through PM's and thread postings it later became clear that he wanted it hosted off site.  No doubt he has his own reasons for this preference.  He asked Dr Zen, Derek Moran to do this.  Derek agreed and it’s close to being agreed, once Magix agree.

With all due respect to Dr Zen, who is a superb choice, I think it would have been better to keep it in house, after all the users here are the ones that have done all the work forwarding their test data.  I previously made the point that if Magix could allow this benchmarking resource to be “pinned”, prevented from sinking, and updated by a moderator or moderators then IMHO that would be ideal, after all, if I was say unwell for a while, there would be a delay in updating.

I get it though that the Moderators have already probably got enough on their plate as it is without adding to it.  Magix's thinking? might well be that this is something that the forum users want, have created themselves and so let them get on with it, if so thats understandable.

To save you trawling through the above thread i’ve copied and pasted a few of the pertinent comments, from page 4 ...

 

“Dr Zen

Hello everyone

Kinvermark asked me a few days ago, if I was willing and able to host the Vegas standardized 4k test project.

I said I could do that - no problemo.

I can host the test on my website and create a dedicated page with permanent download links, however I am not the architect of this project. I have already told Kinvermark that someone must help me with this and write up the instructions/wording and data sheets for me.

Which ever people are in charge of this project, you need to write up exactly how you want this web page to look. Maybe create a Word document and then send it to me. I can then copy and paste this into my website.

I will then ask the leaders of this project to inspect the new web page before it does live to the world.

It is not my job to be creating Spreadsheets for this project - that is something you must do yourself.

I'm still not 100% sure of what I am expected to do, other than offer my website to host the project.

Regards

Derek”

 

JN-

“Hi Derek, the zip file for the project, a user data input text file, a spreadsheet and a suggested render template, FHD, are now available at the link provided. No one is really in charge of the project, it just grew out of a thread. It was Kinvermarks idea to host it elsewhere. For sure you are a great choice, superb site and source of really lovely tutorials. Since its Magix's project still, essentially, albeit a 4K version, I would think you might definitely need to get Magix's permission to host it.

That probably wouldn’t? be the case if it was just left where it is now, residing as a link on this Magix forum. https://www.vegascreativesoftware.info/us/forum/benchmarking--116285/

Something to note ... So far there has been a relatively dismal interest in this amongst the community, ok, a lot of interest but from a very small number of users.”

 

If it all goes as Kinvermark plans then Dr Zen can simply access and port the existing resources located at this threads first post download link. He can then integrate as he chooses into his sites web page.

 

 

 

Kinvermark wrote on 7/25/2019, 8:44 AM

@Former user

I don't mind at all if you want to keep it here. 😀 Just didn't expect anyone to volunteer to host the downloads and keep the spreadsheet updated.

Former user wrote on 7/25/2019, 9:25 AM

@Kinvermark You’ve already set it in motion. Dr Zen is awaiting a response from @VEGASDerek. Lets see how that pans out. It’s not ideal either way that a single user like myself, or an off-site location hosts it. As I said in my previous post it would be best if the moderators pinned and updated it, but let’s wait and see.

Kinvermark wrote on 7/25/2019, 10:18 AM

OK. I don't think the moderators will update the spreadsheet or pin the thread - but I will be happy to be proven wrong! Still leaves the issue of a permanent download link for the test project. Sure, let's wait and see...

Former user wrote on 7/25/2019, 1:35 PM

@Howard-Vigorita I only now realised that I have some of your "Encode mode" entries still as "MC". Can you let me know which to change to i.e 4 available options ... Nvenc, QSV, VCE or CPU, thanks.

Item numbers ...

8 =?

10 =?

24 =?

28 =?

34 =?

36 =?

40 =?

41 =?

Howard-Vigorita wrote on 7/25/2019, 3:02 PM

@Former user That change is not authorized. That is an abbreviation for the Mainconcept Encode Mode option I selected from the Preset menu. I only get 3 choices: Mainconcept, AMD VCE, and Intel QSV. If you prefer to spell it out for clarity, I'd be fine with that. But I do not agree to your changing to something I did not select because that would be false and misleading. I realize you don't get that option yourself if you have an Nvidia card and do not have an AMD card. I don't get the Nvec option because I have an AMD card and no Nvidia card. Nobody gets a CPU option there in the Magix AVC preset group... anyone reporting that option might be using the Sony AVC with "Use legacy GPU" checked in general options... because I think that's the only place it still appears. Here's a screen shot of the Magix AVC menu I get with the Encode Mode dropdown displayed.

Howard-Vigorita wrote on 7/25/2019, 3:06 PM

@JN_  @Kinvermark If all you need is a place to park it, Google offers 15 gigs for free if you sign up.  If you already have a YouTube account, you're already most of the way there. There's no expiration but then again very few things in life are forever.  Personally I'd recommend against parking it on a commercial site because their interest would legitimately be commercial and the fact that it's there would reasonably be perceived as a community endorsement... which may or may not be cool.

If Magix has IP issues, and there are a number of possibilities, there are possible work arounds, depending on the facts.  (Btw, this is just a discussion of possible issues, not legal advice.  Consult a lawyer licensed to practice in the relevant jurisdiction for that.)  For instance Magix might object to their product names being used in a way seeming to imply they are the project creators, or that they endorse it, or that the project fairly or optimally represents the merit or performance of their product or it's abilities to edit, view, or render real world 4K video shot with a video camera.  Then again they may want to adopt and represent all of that.  You might succeed in short circuiting the issue by altering or removing any titles that might give those impressions or look like a Magix Vegas Video advertisement. Or perhaps prominently display some kind of disclaimer.

Then there's the footage.  Somebody shot it. There might have been a full production crew too.  If you look at the RedCar footage that was included by Sony, it has title and credit screens suggesting they contracted with numerous folks to create it and then preserved all the logos, trademarks, and credits in the clips issued with their Press Release distribution. Stuff like that might have been spelled out in the terms of an engagement contract between Sony and the content creators.  The facts and circumstances surrounding these clips would be very relevant.  Particularly the question of whether Trademarks were removed without the consent of creators or whether indicia of origin were removed contrary to law.  If the clips came from Magix, they probably know all the answers and might agree to hold a 3rd Party hosting site harmless and assume liability for any content creator claims.  If not, that can be fixed easily enough.  Just delete the footage track.  I did some render tests and found the video footage track only contributes about 20% to 4K render time.  And around 1% to the Region 1 fps viewing speed; meaning that it I mute some of the noise generator clips, Region 1 is able to play at a full 25 fps rate rather than dropping to 0.286 fps.

Don't get me wrong about what I just said.  I'm not knocking the validity of this benchmark as a hardware and software stress test for Vegas video platforms.  It's actually brilliant to create such a test without the need for space consuming, impractical to download, 4K camera footage.  The use of a Vegas noise generator in 16 places is kind of like compressing video down to the the instructions necessary to recreate it instead of downloading and storing all the bits and frames on your hard drive and inserting them into the time line. But you really need to go through the rigor of that substitution to determine if it's at all like editing, viewing, or rendering an actual video file.  But the fact that it works so well proves it's not at all like putting a lossy-compressed video file in its place.  If you inspect  the original RedCar project... clip 3 is the meaty 50 mbps mxf video clip while clips 1 and 2 are much more compressed but otherwise exact copies of it.  The compressed copies not only demonstrate the versatility of Vegas to simultaneously handle multiple video formats on its timeline, they also illustrate how well lossy-compressed video plays and renders.  This is dramatically illustrated if you replace clips 1 and 2 with clip 3.  (I actually did that in the torture test version of the 4K benchmark I worked on.)  It's also a demonstration of smart substitution... because clips 1 and 2 are used in circumstances where their compression does not detract from quality: they are used in places where they are shrunk into a small p-i-p box (ie, the seating segment) and as an overlay with high transparency. The fact that the Vegas Noise generator does not behave like this at all points to how generated media could be more intelligently used in a real life project to yield a performance boost that might be important in order to achieve full playback speed during authoring:  just solo the noise generator tracks, render them to a compressed format, then replace the original tracks on the timeline. Kind of like making a proxy then tossing the original. Since the noise generator output is visually sparse, meaning most of the noise that's there is invisible, it will compress really well with no quality loss in the display or rendered output. Maybe Vegas 17 will wise up and do this automatically. There's a much worse media generator that floats my boat and that I treat this way so it doesn't sink my ship:  it's the title generator, animation section. Rolling Glow and Enlarge preset.  

Just food for thought.

Former user wrote on 7/25/2019, 3:46 PM

@Former user That change is not authorized. That is an abbreviation for the Mainconcept Encode Mode option I selected from the Preset menu. I only get 3 choices: Mainconcept, AMD VCE, and Intel QSV. If you prefer to spell it out for clarity, I'd be fine with that. But I do not agree to your changing to something I did not select because that would be false and misleading. I realize you don't get that option yourself if you have an Nvidia card and do not have an AMD card. I don't get the Nvec option because I have an AMD card and no Nvidia card. Nobody gets a CPU option there in the Magix AVC preset group... anyone reporting that option might be using the Sony AVC with "Use legacy GPU" checked in general options... because I think that's the only place it still appears. Here's a screen shot of the Magix AVC menu I get with the Encode Mode dropdown displayed.

@Howard-Vigorita This is becoming wearisome.

Mainconcept MC is as I understand it the name of a third party company whose render codecs were bought in by Sony previously, and now used by Magix. When users used it without HW Acc. we just reported it mostly as Cpu. So MC or Cpu entries in encode mode mean the same.”

I'm quoting myself from earlier above, I have also given previous examples where users have historically given Cpu as encode mode, i’m not going to repeat all of that again.

That would be false and misleading “ Far from my intention, instead I think the Magix Mainconcept MC entry is maybe a bit ambiguous, so my attempt at clarity.

As I understand it there are a few ways to encode, render out a file, sometimes with or sometimes without HW Acceleration. With HW Acc. on/off it can be Nvenc, Vce and Qsv for Asic HW encodes. With HW Acc. on/off it can be MC and Cuda for Cpu and Gpu encodes.

Entering Cpu assists, focuses the viewer, user, in understanding that it was a Cpu encode, either with or without HW Acc.

Entering Mainconcept MC is of course what was selected to encode, but the viewer then has to in his/her mind translate to Cpu.

So it goes in as a Cpu encode because that’s what was used to do the encoding.

If your still very unhappy with this and want your results removed ( I sincerely hope not) I will of course do so, but for my part I'm attempting to create consistency within this project and also backwards in relating to the many Red Car testings and other tests that have gone before.

Former user wrote on 7/25/2019, 4:02 PM

@Howard-Vigorita As for hosting the project, @Dr Zen is waiting for a reply. If the answer is yes from Magix he will host it, I'll then remove the zip file from this thread, unless MagixDerek says its ok, the rest I can leave. If the answer is no, because of legalese, then Dr Zen won’t, in that case I'll then ask MagixDerek, for this 4K version here, if its ok, if the answer is no then someone better start putting together a new project.

Howard-Vigorita wrote on 7/25/2019, 10:16 PM
@Howard-Vigorita This is becoming wearisome... the viewer then has to in his/her mind translate ....

You certainly read my mind. Please remove my results. I support your good work organizing and promoting this but not misrepresenting results for settings. Especially my own. I will not be a part of that.

fr0sty wrote on 7/25/2019, 10:58 PM

Mainconcept IS CPU ENCODING. I don't know how many times we have to say that. The timeline rendering will be GPU accelerated as long as you have it enabled in preferences, but the encoding by the Mainconcept encoder is CPU based, just like every other encoder in Vegas Pro is except for NVENC/AVC/Quicksync/Sony AVC.

Systems:

Desktop

AMD Ryzen 7 1800x 8 core 16 thread at stock speed

64GB 3000mhz DDR4

Geforce RTX 3090

Windows 10

Laptop:

ASUS Zenbook Pro Duo 32GB (9980HK CPU, RTX 2060 GPU, dual 4K touch screens, main one OLED HDR)

Grazie wrote on 7/25/2019, 11:26 PM

@Kinvermark 

Mainconcept IS CPU ENCODING. I don't know how many times we have to say that.

I’ve understood this. Then why is this being brought up again, not @Kinvermark, as part of this BENCHMARKING thread? There has to be a reasonable reason to discuss it? What’s the issue here?

Kinvermark wrote on 7/25/2019, 11:31 PM

Yes, let's keep this all in perspective. We have a few dozen results being seen by about the same number of moderately interested forum members. No big deal if someone is slightly "mislead" by minuscule sematic differences. The benchmark methodology is in its infancy and will no doubt need tweaking over time - assuming it even survives to adolescence. :)

Kinvermark wrote on 7/25/2019, 11:35 PM

@Grazie

Oops, crossed posts. In short: the render templates available don't actually say "CPU" they say "Mainconcept" for encoder. Thus Howard-Vigoritas objection. Practically it is the same thing - thus JN_'s point of view.

Grazie wrote on 7/26/2019, 12:16 AM

@Kinvermark - Oh yes indeed. So, it’s in these perambulating discussions we’re getting convergence of ideas, founded in evidenced based results. I can dig that! 😎