From what I've heard, unless AMD crippled memory on Vega 56 in comparison to memory on Vega 64 (we'll know soon), it should turn out to be the card to get. If memory is ok, a simple OC to the same frequency Vega 64 runs at should give performance around 3% lower than Vega 64. All that for $100 less! Vega 56 has the potential to make a huge splash and AMD knows it - they told reviewers to focus on 56 instead of the 64! Watch out for this card, as it can turn into the best choice till Volta comes out. I'll be buying it, since I have a FreeSync ultrawide monitor and AMD makes most sense for me

Scisca said: From what I've heard, unless AMD crippled memory on Vega 56 in comparison to memory on Vega 64 (we'll know soon), it should turn out to be the card to get. If memory is ok, a simple OC to the same frequency Vega 64 runs at should give performance around 3% lower than Vega 64. All that for $100 less! Vega 56 has the potential to make a huge splash and AMD knows it - they told reviewers to focus on 56 instead of the 64! Watch out for this card, as it can turn into the best choice till Volta comes out. I'll be buying it, since I have a FreeSync ultrawide monitor and AMD makes most sense for me

Well. No.

Vega 56 has 14% less compute than Vega 64.
14% less fillrate.
17% less bandwidth.

So don't be surprised if it is about 15% slower than Vega 64 *when* the core is overclocked to the same clockrate at Vega 64.
Expect it to be around 20% when not overclocked.

Will it be value for money? Shit yes. nVidia is still better though, hard to pass up Pascals potency and efficiency...
Still I am of the firm believer that there is no such thing as a bad GPU, only a bad price, AMD has to price Vega competitively due to how power hungry it is.

No they aren't, not 100% in the objective sense, as everyone's setup is different, thus they won't always run exactly the same as those within the bench results.

You can't tell me my results are baseless and yet their benchmarks aren;'t and then call my hw busted because it doesn't fit with whart you consider an objectively good setup based on their specs to mine.

I will argue till the end of time that my previous setup, let alone my new current one isn't "broken".

I know you have wisdom and all that, but you aren't skynet, you aren't 100% objectively right all the time in every sense. I've seen this in your whenever anyone says anything differently. It's a bit obvious when even CGI has to don a shield (even if it's meant for fun, it still tells me about how you'll argue differently).

Pemalite said: Scisca said: From what I've heard, unless AMD crippled memory on Vega 56 in comparison to memory on Vega 64 (we'll know soon), it should turn out to be the card to get. If memory is ok, a simple OC to the same frequency Vega 64 runs at should give performance around 3% lower than Vega 64. All that for $100 less! Vega 56 has the potential to make a huge splash and AMD knows it - they told reviewers to focus on 56 instead of the 64! Watch out for this card, as it can turn into the best choice till Volta comes out. I'll be buying it, since I have a FreeSync ultrawide monitor and AMD makes most sense for me Well. No. Vega 56 has 14% less compute than Vega 64. 14% less fillrate. 17% less bandwidth. So don't be surprised if it is about 15% slower than Vega 64 *when* the core is overclocked to the same clockrate at Vega 64. Expect it to be around 20% when not overclocked. Will it be value for money? Shit yes. nVidia is still better though, hard to pass up Pascals potency and efficiency... Still I am of the firm believer that there is no such thing as a bad GPU, only a bad price, AMD has to price Vega competitively due to how power hungry it is.

Well. No.

That's not how AMD GPUs work. Compare Fury X and Fury, which are somewhat similar to Vega. 4096 CUs vs 3584 CUs (14% difference), 640 BG/s vs 512 GB/s bandwidth (25% difference), 5% less fillrate - yet after setting the same clocks, the difference in performance is about 3%. Compare the RX 470 and RX 480, or regular R9 cards and the "+" cards - same story. AMD GPUs don't scale so well when close to the top performance. Unless they intentionally held the memory back, expect Vega 56 to be only a hair behind Vega 64.

Well no. I don't think you comprehend how GPU's work.

If you are limited in another part of the pipeline, then of course the difference will be reduced as that would be your bottleneck.
The Fury and Fury X had the same ROP capabilities and had the same Geometry capabilities... And was extremely limited in VRAM.
Which means... If the game was ROP, Geometry or Memory intensive, then performance would equalise.

But throw a compute-only task that only stressed the shaders, then you will see performance between the two start to spread apart.

Vega however has the same Shader count as Fury, but dials up the clockrate, has overhauled ROPS, Geometry capabilities, doubling of RAM and the introduction of new techniques which further increases memory bandwidth and Geometry/ROP throughput. It's not going to be apples to apples.

In short. It's a difference beast and we should wait on benchmarks from the likes of AMD and then possibly revisit this discussion if you so desire.

You still don't get it. - I don't care about Ancedotals.

If you think I will blindly believe someones claim over the internet. Think again. - That would be moronic for me or anyone to do.

And I am more than happy to change my view and even apologise if I am incorrect, if something is substantiated.

Tell it to not just him but to everyone in the world! 

I get so unbelievably angry that people will pass off anecdotes as facts ... 

There wouldn't be so many lies in history and theology on earth that we would have to endure ...

That being said it'll be interesting to see how Vega will perform in Wolfenstein 2 and Far Cry 5 as it might improve RTG's prognosis in the future ... 

Cool story bro.

Fury X vs Fury - 3%
HD 7970 vs HD 7950 - 3-5%
R9 380 vs R9 380X - 1-3%
R9 390 vs R9 390X - 3-5%
RX 470 vs RX 480 - 3-5%

It's been a thing for AMD for generations that cutting down CUs doesn't impact performance as much as one would think. I think it's legit to expect the same story repeating with Vega, especially since despite the fact Vega 64 launches today and Vega 56 in almost 2 weeks, AMD told reviewers to focus on 56 - they know what people are gonna buy. Obviously, nothing is granted before we see the benchmarks, but Vega 56 card has the potential to be a Vega 64, 1070 and 1080 killer. IF! If it doesn't powerthrottle and run lower HBM voltage. The much lower TDP of Vega 56 may indicate that the voltage is lower and it would be a bummer if that happened, since for Vega cards memory clock matters a lot. HBM clock is more important for performance than core clock (for Vega FE at least). The main issue will thus be, whether we have enough voltage to push memory back to full speed of Vega 64. If that's the case - the cards' performance in games should be basically indistinguishable.

Obviously, if AMD messed with the memory, like they did with RX 470 having worse memory, the gap will be more significant. All we can do is wait now, but I hope Vega 56 turns out to be epic

I don't need to think again, because I don't need to fully believe in the benches on those sites, since we're all sporting different hw.

I find it rather short sighted to take the benches as fact across the board for many sets of hw and to then turn around and go "well yours is busted or something", when not even knowing how said hw is broken.

So that's to say that has happened a minute number of times on here then. You've been factually right 97% of the time?. I haven't really seen you change to what others have posted nearly as often as some have on here. I've seen you counterpoint to other's counters to your original points, just like you're doing with mine (like my previous rig being broken, instead of questioning why the different results. I look as benches and I don't take them as gospel either, I'll always question them).

Of course. Because you introduce a bottleneck into the pipeline which is the point I am trying to convey.

With that. Vega has arrived. And looking at Anandtechs benchmarks, the performance delta is a good 10-20% different between Vega 56 and Vega 64.
http://www.anandtech.com/show/11717/the-amd-radeon-rx-vega-64-and-56-review

The benchmarks are accurate, we know they are accurate because there are dozens of tech-sites on the internet which have gathered empirical evidence and compiled them into appropriate charts for you to peruse. - And they all tell the same story.

You are literally trying to justify your ancedotal point of view as something that is more accurate than some of the most reliable sources of information on the internet, that's not factual and not something I can adhere to. - And you will *never* change my mind on this... As I apply this standard of thinking to everything.

Instead of conveying some message, focus more on reading. I've made the remark about crippled memory numerous times, including in the very first sentence I wrote in this thread.

From what I've learned thus far, most probably memory itself is fine, but AMD has bloked voltage in the BIOS, thus artificially pulling the card back. Even flashing BIOS isn't working as of yet This is exactly in line with what I've written. There's no real bottleneck, just an artificial, arbitrary muzzle on the card. I hope non-reference cards will ship with modified BIOS, that will allow to overclock it to the card's fullest. Should manufacturers do that, we'll see performance I was talking about. I mean, in the link you've provided the difference ranges from 3% to 15%, depending on game and resolution. All that without overclocking to match the clocks (I assume). This is very promising, but we need to get by that stupid BIOS.