confirmed: AMD's big Navi launch to disrupt 4K gaming

I don't know what platform you are using to do your transcoding but many of them require you to make changes in settings or modify an .ini to utilize more than 8 cores, they will do so happily but most don't do it by default.

I mostly utilize ffmpeg manually from the command line, but sometimes use handbrake. Both under Linux.
 
I mostly utilize ffmpeg manually from the command line, but sometimes use handbrake. Both under Linux.
Have you been using the -threads command in ffmpeg? Depending on your OS you have to play with it a little otherwise it it notorious for loading one or 2 then trickling out over a few others but not using them too much. Handbrakes algorithms don't really scale beyond 6 threads maybe 8 per encode but you can take advantage of their batch encoding stuff to do multiples at once. However I believe that the .264 and .265 encoders in handbrake also have a -threads argument you can use to have it use more it is mentioned in a few places but I haven't played with it overly much (my Handbrake box is only 8 Thread).
 
Let us hope it is good. Nvidia is well on their way to their next generation though. My RTX 2070 is around 15 months old, and I didn't order when they were first available. I am hoping they can get something that can at least trade blows with the 2080Ti, as that will essentially become the RTX 3070 (or whatever it will be called) in terms of performance. Ideally you would want something a lot better than that, but I am not seeing that. If they can great, may they'd be able to compete with a hypothetical RTX 3080.

I don't seem them competing with Nvidia's future Ti or Titan, but they don't necessarily have to. If they can compete in most segments the price will still come down. If Nvidia is forced to put out the **70s at $400 again and the **80s at $550-600, it will be hard to sell a 20% performance jump for $1200 as an example.
 
Have you been using the -threads command in ffmpeg? Depending on your OS you have to play with it a little otherwise it it notorious for loading one or 2 then trickling out over a few others but not using them too much. Handbrakes algorithms don't really scale beyond 6 threads maybe 8 per encode but you can take advantage of their batch encoding stuff to do multiples at once. However I believe that the .264 and .265 encoders in handbrake also have a -threads argument you can use to have it use more it is mentioned in a few places but I haven't played with it overly much (my Handbrake box is only 8 Thread).

Default behavior for ffmpeg for x264 encodes is to use as many threads as are needed. I never have to specify threads and it will saturate my 12c/24t encoding machine.
 
Let us hope it is good. Nvidia is well on their way to their next generation though. My RTX 2070 is around 15 months old, and I didn't order when they were first available. I am hoping they can get something that can at least trade blows with the 2080Ti, as that will essentially become the RTX 3070 (or whatever it will be called) in terms of performance. Ideally you would want something a lot better than that, but I am not seeing that. If they can great, may they'd be able to compete with a hypothetical RTX 3080.

I don't seem them competing with Nvidia's future Ti or Titan, but they don't necessarily have to. If they can compete in most segments the price will still come down. If Nvidia is forced to put out the **70s at $400 again and the **80s at $550-600, it will be hard to sell a 20% performance jump for $1200 as an example.

These hypotheticals with NV being "forced" to do anything always read like - and then turn out to be - fanfiction. They are perpetually laps ahead of AMD. Problem is that unlike Intel, they never take their foot off the gas.

There will be no 3070 at $400 or 3080 at $550-600 simply because if AMD has an offering anywhere close, AMD will not be price that low. This fantasy about waiting for cheap AMD GPUs isn't going to happen. Rightly so, if AMD keeps building strong products, the market will buy and desperation pricing isn't needed.

As for 3080Ti, the 2080Ti proved that the market has a ferocious appetite for fastest performance at any price. So even if 3080Ti is $2000, they'll all be sold out. People can bang head into keyboard all they want about this being unfair somehow.

And the people that insist "but 2080Ti is only .1% of GPU market share, WGAF", the halo effect cannot be understated. In this market, it's everything, because it cascades perception of the whole product stack
 
Last edited:
Let us hope it is good. Nvidia is well on their way to their next generation though. My RTX 2070 is around 15 months old, and I didn't order when they were first available. I am hoping they can get something that can at least trade blows with the 2080Ti, as that will essentially become the RTX 3070 (or whatever it will be called) in terms of performance. Ideally you would want something a lot better than that, but I am not seeing that. If they can great, may they'd be able to compete with a hypothetical RTX 3080.

I don't seem them competing with Nvidia's future Ti or Titan, but they don't necessarily have to. If they can compete in most segments the price will still come down. If Nvidia is forced to put out the **70s at $400 again and the **80s at $550-600, it will be hard to sell a 20% performance jump for $1200 as an example.

I'm curious about the "3070" as well, and that segment. My 2070 is nowhere even slightly "old" yet and has been great bang for the buck.
 
I mostly utilize ffmpeg manually from the command line, but sometimes use handbrake. Both under Linux.

I would say use Vidcoder but you use linux. They're has to be an equivalent for linux.
Vidcoder is based on handbrake, and is built for batch encodes. You drag and drop the videos you want and it can encodes 4 at a single time (4 instances of handbrake's ffmpeg), which would help speed up your situation immensely since you have a Threadripper.
I've been using it to encode my personal media library of family videos/blurays to x265 with much satisfaction with my 1700.
It makes converting a folder with hundreds of old huge .MPGs that came from tapes into tiny small x265 .MKVs effortless and quick.

You got a Threadripper!! Use them threads!
 
This is not confirmed. Better than a rumor but it's basically AMDs usual hype machine. It's confirmed when I have the card in my machine and the performance is the same as an actual review. But I believe it.
 
These hypotheticals with NV being "forced" to do anything always read like - and then turn out to be - fanfiction. They are perpetually laps ahead of AMD. Unlike Intel, they never take their foot off the gas.

There will be no 3070 at $400 or 3080 at $550-600 simply because if AMD has an offering anywhere close, AMD will not be pricing their GPUs anywhere near that low. This fantasy about wait for cheap AMD GPUs is just outdated.

As for 3080Ti, the 2080Ti proved that the market has a ferocious appetite for fastest performance at any price. So even if 3080Ti is $2000, they'll all be sold out. People can bang head into keyboard all they want about this being unfair somehow.

The people that scream "but 2080Ti is only .01% GPUs they're not a big deal" don't understand that the halo effect isn't just valuable in this market, it's everything.

We've already seen the price drop on the RTX 2060 to counter AMD's new offering. We can also look to the GTX 260 price drop in the past.

You can charge as much as you want and obviously someone will be happy to spend $10000 on the fastest GPU, but less people will buy it. If AMD gets something within 20% of the performance of the fasted GPU and prices it at $600-700, Nvidia will have to compete. There will of course be someone willing to spend $2000, but less will. Three times the price for a 20% performance jump will turn a lot off even at the high end.

A big problem with AMD has been timing. By the time they get out a good RTX 2080 TI competitor, Nvidia will probably have something new coming along. The RX 5700XT is great from a performance and price standpoint, but came a year after the original RTX 2070. I had already upgraded so it wasn't even an option.
 
Last edited:
I could see nVidia having to drop 2080 prices if AMD came within 20% of the performance of a 2080 Ti and priced it at 600/700 USD as the 2080 is within 30% of the 2080 Ti. That's without accounting for ray tracing which, if you actually want it, is nVidia's market for the taking, at the moment.
 
I'm more or less with everyone else that I'll believe it when I see it. I am on the side of optimism here as AMD for the most part has been making good strides in a positive direction and have more or less done what they said they would do for the past 2-3 years on the CPU side and have made great strides on the video card side (bungled 5600XT launch not withstanding).

However, I think that the assessment that nVidia's top card will likely have a 20-30% improvement over 2080Ti also in this year means that not only does Big Navi have to compete with 2080Ti, but it will have to compete with nVidia's Ampere. Big Navi would have to be a significant leap of almost 60% over Radeon VII to reach parity with a card that's 20-30% faster than a 2080Ti.
I mean, it's possible, but we've been disappointed before. And I won't hold my breath. Still, this card will definitely help the market by adding a non-nVidia alternative, which presumably at worst will be cost effective and excellent bang for the buck in its place in the market. Top dog? That is faaaar less certain.
 
The follow on, Big Navi Everafter is supposedly so powerful that it comes with a warning that small pets will actually begin to cook if staring at it for prolonged periods of time.

(no mention of harm to small children or infants though)
 
I think beating Nvidia's next halo card is probably out of the question.

But getting close to a 2080 Ti with a more affordable (but still probably high) cost seems possible.

Hopefully AMD won't go higher than $700 or $800, if they can do that it would be a huge win.
 
Remember when high end was $500. Those were the good ol' days.
Yeah, I think I got an ATI 9700 Pro when it came out for $399 and that seemed like a HUGE sum for a graphics card back then.

How times have changed (even accounting for inflation, $1,200 graphics cards are still insane).
 
Last edited:
hm

AMD does not seem to be exaggerating lately when it comes to performance, they've pretty much over delivered on every promise for ryzen, ryzen+, and ryzen 2.

RTG has a very bad history of overpromising, but Navi seemed to be on the money and I don't remember any crazy fud about it so hopefully there's some truth in this.
 
"disrupt" 4k gaming?? LOL. Sounds like a headline designed to influence investors...

'Disrupting' the 4k gaming market would be coming out with 8k gaming at 144hz for $500... or a streaming service for 4k gaming with no latency for $9.99 a month, all games included.. stuff like this would shake up the market.

Not holding my breath that an AMD GPU is going to accomplish those things anytime soon.

yea stuff like that will also wont last because business'es are suppose to make profit you know. Or the services go under lol.
 
I believe Nvidia will probably make ray tracing faster. I highly doubt they are going to deliver 50% more performance at 50% less power, that shit just doesn't add up unless they are just talking about ray tracing numbers.
 
hm

AMD does not seem to be exaggerating lately when it comes to performance, they've pretty much over delivered on every promise for ryzen, ryzen+, and ryzen 2.

RTG has a very bad history of overpromising, but Navi seemed to be on the money and I don't remember any crazy fud about it so hopefully there's some truth in this.

Even the 5500XT and the 5600XT were not over promised.
 
  • Like
Reactions: N4CR
like this
So, how much as the 8800 Ultra?

8800 ultra was over $800. AMD didn't have decent competition for it as the 2900 was a flop.

x1950xt was faster than the 7900gtx and launched at $449 (2006)

Even the 1080 launched at $599.
 
What exactly is confirmed here? Big Navi to be presented at computex? It may outperform a 2080ti? Ok, but then Nvidia will drop a 3080. Dont get me wrong, the 2080ti released at $1200 was a slap in the face to me, so much so that I upgraded to a 5700xt from a 980ti. Big navi needs to be $599 and beat a 2080ti with a big navi super waiting in the wings.
 
What exactly is confirmed here? Big Navi to be presented at computex? It may outperform a 2080ti? Ok, but then Nvidia will drop a 3080. Dont get me wrong, the 2080ti released at $1200 was a slap in the face to me, so much so that I upgraded to a 5700xt from a 980ti. Big navi needs to be $599 and beat a 2080ti with a big navi super waiting in the wings.

Where do people get these prices from????
 
... the 2080 Ti has been out at $1200 for a long time and most folks who want the performance are not willing to spend the price

apparently that does seem to be the case:

" Is the 2080 TI worth the price?
But it does so with a mammoth cost increase that is completely unjustifiable. If money is no object, the 2080 Ti is a great GPU.
But given that just 1.49 percent of Steam users own a GTX 1080 Ti — and that GPU is $500 less than the RTX 2080 Ti — it's clear
that money is an object for the vast majority of gamers."

https://www.extremetech.com/gaming/...d-rtx-2080-ti-review-you-cant-polish-a-turing
 
I'll believe it when I see it. AMD has over hyped under-performing video cards one too many times to be trusted until the product is in the hands of unbiased reviewers.

Exactly.
 
Back
Top