• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

Poll: ** The AMD VEGA Thread **

On or off the hype train?

  • (off) Train has derailed

    Votes: 207 39.2%
  • (on) Overcrowding, standing room only

    Votes: 100 18.9%
  • (never ever got on) Chinese escalator

    Votes: 221 41.9%

  • Total voters
    528
Status
Not open for further replies.
The right one definitely looks way different (worse) detail/colour wise, there's something off when the guy was rotating the view fast but that could be the guys camera causing that effect; wonder if they confirm which is which like he said!
 
How's it a fail? Clearly show previous nvidia cards struggling, not just AMD, bottom line is Gameworks is deisgned to get users to upgrade by hurting previous gen cards. It's not anything good, it's a marketing scam.
Its a fail due to the fact that Gameworks and Nvidia gimping old cards are 2 different things, for starters hes comparing charts from launch Fallout 4 to 1.3 Beta drivers which if you bothered to look at the actual 1.3 release driver benchmarks you would find that cards from both vendors showed an improvement opposite to what the graphs in the video show.

Gameworks is just optional eye candy with a performance hit if enabled if people are not happy with the performance hit they can simply turn it off its purely optional and quite simple to comprehend, the fact that he also used Batman Arkham Knight as an example of his ignorance even with Gameworks effects disabled the game still ran like a steaming pile of crap nothing to do with Gameworks. If he knew anything about game software development and optimisation its not like games are designed around Gameworks. Also read post
#6432 he sums it up..
 
Outrageous these companies not releasing products until they are released.

The problem is we hit peak hype quite some time ago, and now it's turned to frustration. It's generally not good to frustrate and annoy your customers, because then they jump ship to a competitor who hasn't annoyed them. Look at the number of people that jumped ship to green when FE was released and RX was no where to be seen. Or the T-shirt reveal. Or the reddit AMA that told us little to nothing. Or the "Poor Volta" ads.

There's only so long you can keep your loyal customers on the hook before they are no longer loyal or your customers, and I think AMD is well past that point now. They need to put up or shut up, and I think if Vega is below par or expensive, there will be a lot of AMD love that turns to hate. You let your customers down too often, and they lose any brand loyalty.
 
Only issue with that is how much bandwidth the highest-end Navi may need.

600mm2 (or 4 200mm2 dies if it's MCM) on 7nm could plausibly give over 40 Tflops, so 3xGTX 1080 Ti ballpark.

Even the max GDDR6, at 16000 MHz, and on a 512-bit bus would give 1TB/s. That's likely not enough for 40 Tflops.

I'd have thought 20-25 Tflops would be the limit of 1TB/s, and that would only be a ~350mm2 chip on 7nm. Also 512-bit buses are usually avoided, so realistically GDDR6 tops out at 768 GB/s.

The 1080 Ti and Titan Xp still make nice gains having their memory overclocked up to 600 GB/s, and they're only ~14 Tflops.

So potentially, unless there's some more memory efficiency gains, all chips over ~300-350mm2 on 7nm could ideally do with 4 stacks of the slowest HBM3, or 2 stacks of the fastest.

Whereas the smaller cards would prefer GDDR6 and 256-bit and 384-bit buses, because it's cheap and easy. So Probably both AMD and Nvidia will have to design for GDDR6 and HBM3 in mind, for their 7nm generations.


Possibly, after all as AMD have done with gddr5 for Polaris and HBM for Vega they may have a staggered release program with Navi and something else, I don't thimk we'll see anything near 40 Tflops though, That said it would be great if they found a way to make a threadripper type gpu.
 
The problem is we hit peak hype quite some time ago, and now it's turned to frustration. It's generally not good to frustrate and annoy your customers, because then they jump ship to a competitor who hasn't annoyed them. Look at the number of people that jumped ship to green when FE was released and RX was no where to be seen. Or the T-shirt reveal. Or the reddit AMA that told us little to nothing. Or the "Poor Volta" ads.

There's only so long you can keep your loyal customers on the hook before they are no longer loyal or your customers, and I think AMD is well past that point now. They need to put up or shut up, and I think if Vega is below par or expensive, there will be a lot of AMD love that turns to hate. You let your customers down too often, and they lose any brand loyalty.

I'm in the market for a GPU soon, and I know when Vega details are out so I'll wait til then to make a decision based on what the product actually is, not on some frankly slightly odd concept of loyalty to a multinational corporation.

Maybe I'm just getting too old for all this hype train business :P
 
I'm in the market for a GPU soon, and I know when Vega details are out so I'll wait til then to make a decision based on what the product actually is, not on some frankly slightly odd concept of loyalty to a multinational corporation.

Maybe I'm just getting too old for all this hype train business :p

The problem is that even those that had no brand loyalty but were simply waiting to see which company delivered them an upgrade stopped waiting and bought green. Nvidia had products to sell, AMD had lots of promises that led to disappointment when nothing (not even any info) was delivered. If you were an agnostic customer and needed a mid-high end card, the best AMD could offer was 480/580, while Nvidia had 1080/ti. You can't sell what you don't have.
 
Why would FE buyers care if RX is faster in gaming? FE is compute/workstation card.

AMD better make sure that the RX is gimped in professional workloads, or they will be in for a whole lot of pain lol ! Coz now every reviewer and their dog will be testing the RX in productivity benches.
 
The problem is that even those that had no brand loyalty but were simply waiting to see which company delivered them an upgrade stopped waiting and bought green. Nvidia had products to sell, AMD had lots of promises that led to disappointment when nothing (not even any info) was delivered. If you were an agnostic customer and needed a mid-high end card, the best AMD could offer was 480/580, while Nvidia had 1080/ti. You can't sell what you don't have.

I didn't. nVidia's current high end offerings aren't good value, I'm waiting to see what competition Vega offers.

I'm not particularly interested in raw numbers, more value. 1080ti performance at a sensible price? I'm interested...if not....well I'm not paying 700 EUR for a GPU so I guess I'm waiting a little longer :)
 
I didn't. nVidia's current high end offerings aren't good value, I'm waiting to see what competition Vega offers.

I'm not particularly interested in raw numbers, more value. 1080ti performance at a sensible price? I'm interested...if not....well I'm not paying 700 EUR for a GPU so I guess I'm waiting a little longer :)

RX Vega will probably be between 1080 and 1080Ti performance. It won't beat a Ti...apart from the price of it.
 
Possibly, after all as AMD have done with gddr5 for Polaris and HBM for Vega they may have a staggered release program with Navi and something else, I don't thimk we'll see anything near 40 Tflops though, That said it would be great if they found a way to make a threadripper type gpu.

The long term won't be a Threadripper type GPU in terms of monolithic packages on a substrate communicating but more like effectively dissecting the GPU into blocks of sub-systems and controllers in seperate packages using recent advances in substrate technology to be able to communicate fast enough between them to make it work with highly specialised and dedicated interconnects.

DRAM is about due a shrink to 1x nodes as well which should see a huge jump in capibilities.
 
Status
Not open for further replies.
Back
Top Bottom