AMD RX 7900XT, 90% to 130% faster than 6900XT, MCM, Q4 2022.

[melmac]

Volta never became a gaming gpu because according to jensen it was too expensive to produce. So presumably that's what that "poor volta" thing was about as it was widely expected to be a gaming gpu at some point but it never happened.

Found an article with quotes saying as much: https://www.techspot.com/news/70584-nvidia-volta-gaming-gpus-not-foreseeable-future.html

That poor Volta came out long before it was known that there was going to be no Volta Gaming GPUs. Whether AMD knew or not or whether they intended it or not, doesn't really matter. That "Poor Volta" raised the expectation levels of the performance of Vega. It was terrible marketing that completely backfired.

AMD have 50 years of history, the majority of it making very good and efficient products, my knowledge of it spans 25 years.

Due to a combination of a few mistakes and a lack of money AMD had a few years of products that were not good, so did Intel, so did Nvidia.

They were working through their problems and getting past them before this current generation of GPU's, i saw no reason why AMD's bad run should continue when they had already got past it.

You said it was nonsense when people thought that RDNA 2 would use a lot more power based on AMD's presentation slides. People were basing that on the last few GPU releases, not what AMD released 40 years. Vega, Polaris, Fury, Hawaii. Vega, Polaris and Fury were all hyped up based on AMD slides or AMD marketing. And they weren't all that great when it came to power efficiency either.

So, it's not really surprising that people were taking AMD's slides with a pinch of salt and questioning how they were going to achieve that 50% over RDNA 1 without increasing the power consumption.

AMD did really well with RDNA 2. It's the first time they are competing from the top down and without any issues. Now they have to do what they have done on the CPU side. And that's continue to release good products.

 

[Gerard]

That poor Volta came out long before it was known that there was going to be no Volta Gaming GPUs. Whether AMD knew or not or whether they intended it or not, doesn't really matter. That "Poor Volta" raised the expectation levels of the performance of Vega. It was terrible marketing that completely backfired.

Pretty obvious they knew as they'll hear about stuff from the competing company long before it comes public knowledge. Obviously an internal dig that people took the wrong way.

 

[humbug]

That poor Volta came out long before it was known that there was going to be no Volta Gaming GPUs. Whether AMD knew or not or whether they intended it or not, doesn't really matter. That "Poor Volta" raised the expectation levels of the performance of Vega. It was terrible marketing that completely backfired.



You said it was nonsense when people thought that RDNA 2 would use a lot more power based on AMD's presentation slides. People were basing that on the last few GPU releases, not what AMD released 40 years. Vega, Polaris, Fury, Hawaii. Vega, Polaris and Fury were all hyped up based on AMD slides or AMD marketing. And they weren't all that great when it came to power efficiency either.

So, it's not really surprising that people were taking AMD's slides with a pinch of salt and questioning how they were going to achieve that 50% over RDNA 1 without increasing the power consumption.

AMD did really well with RDNA 2. It's the first time they are competing from the top down and without any issues. Now they have to do what they have done on the CPU side. And that's continue to release good products.

Have you learned not to dismiss AMD's slides out of hand?

I mean, why would they lie?

BTW. Raja now works for Intel. He's now in-charge of their graphics department.

 

[gpuerrilla]

BTW. Raja now works for Intel.

See their dGPU release is going to plan..

 

[humbug]

See their dGPU release is going to plan..

Exactly, poor Intel

 

[melmac]

Have you learned not to dismiss AMD's slides out of hand?

I mean, why would they lie?

BTW. Raja now works for Intel. He's now in-charge of their graphics department.

And yet, nothing you are saying changes anything that I have said. After Hawaii, Fury, Polaris, Vega, people were right to question whether RDNA 2 cards would perform as advertised without using a ton of power. It wasn't nonsense to be wary.

 

[humbug]

And yet, nothing you are saying changes anything that I have said. After Hawaii, Fury, Polaris, Vega, people were right to question whether RDNA 2 cards would perform as advertised without using a ton of power. It wasn't nonsense to be wary.

Sure, but don't you think there comes a point where we should move past the past?

 

[melmac]

Sure, but don't you think there comes a point where we should move past the past?

Two points, first, what reason would they have had for moving past the past before the RDNA 2 release?

Second. It's up to AMD. As I said above AMD have to do the same as they have done and are doing on the CPU side of things. Release good products consistently for a few generations. If AMD had released the first generation of Ryzen CPUs, then followed up with a failure like bulldozer. Nothing would have changed for them. It's going to be the same on the GPU end. RDNA 3 is going to have to be as good or better than RDNA 2.

RDNA 2 is great but it will be for nothing if RDNA 3 sucks.

 

[humbug]

Two points, first, what reason would they have had for moving past the past before the RDNA 2 release?

Second. It's up to AMD. As I said above AMD have to do the same as they have done and are doing on the CPU side of things. Release good products consistently for a few generations. If AMD had released the first generation of Ryzen CPUs, then followed up with a failure like bulldozer. Nothing would have changed for them. It's going to be the same on the GPU end. RDNA 3 is going to have to be as good or better than RDNA 2.

RDNA 2 is great but it will be for nothing if RDNA 3 sucks.

A change of management, the new management having several years history of publishing accurate if not actually sandbagged product slides.

I think that's good reason to move on.

 

[melmac]

A change of management, the new management having several years history of publishing accurate if not actually sandbagged product slides.

I think that's good reason to move on.

Really Humbug? If we weren't discussing AMD's GPUs, you wouldn't be saying that. You would need to see results before you changed your mind about how good or bad they are.

 

[humbug]

Really Humbug? If we weren't discussing AMD's GPUs, you wouldn't be saying that. You would need to see results before you changed your mind about how good or bad they are.

Really Humbug? If we weren't discussing AMD's GPUs, you wouldn't be saying that.

I don't get where you are coming from with that.

The HD 4000 and HD 5000 series were good GPU's.
The HD 6000 series was a decent architecture.
GCN was not bad
The 290X was a pretty good GPU it just had a bad choice reference cooler.

The Fury cards, yeah they weren't good, this is when you had Raja Kadury in-charge, he insisted "The Radeon Technologies Group" be run as a separate entity to AMD, this is also where you had all the "Poor Volta" crap.

Well it was not long before he took an extended break and then resigned.
As soon as he left Lisa took over, scrapped the "The Radeon Technologies Group" thing, its now back to just AMD GPU Division.
Her first order of business was to look at Vega, (Fury Architecture) as it was being used in APU's, with in months AMD announced a new line of APU's with a revised Vega architecture that had a 50% performance per watt improvement, and it actually did.

RDNA1 "50% performance per watt improvement" which is exactly what we got.

Given also that the GPU divition was now being run by the same people that brought us Ryzen CPU's there was plenty to suggest the days of "Poor Volta" had gone with Raja and his Poor Fury X.

 

[five8five]

Delete

 

[Mel_P]

Doesn't 90% faster mean 10% slower?

 

[Wrinkly]

Doesn't 90% faster mean 10% slower?

It means RDNA3 will have 190% performance when compared to RDNA2's 100%. Or RDNA2+90%.

 

[five8five]

Bus width, memory and MCM rumours...

https://videocardz.com/newz/amd-nav...-feature-384-bit-memory-bus-24gb-gddr6-memory

 

[five8five]

Doesn't 90% faster mean 10% slower?

10% slower would be 90% of the original performance.

 

[humbug]

90% to 130% faster

Its not 90% of the performance, its 90% on top of the performance.

 

[Troezar]

90% to 130% faster

Its not 90% of the performance, its 90% on top of the performance.

Yes they should just say 1.9-2.3x the performance to save confusion.

 

[gpuerrilla]

Choo Choo!!

 

[Troezar]

Choo Choo!!

Hi Paul is that you?