• 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.

Now's your time AMD

Meh, them grapes they are sour. I was under the impression that Navi is supposed to be a Polaris replacement where Vega is the Fury replacement. Navi going into the PS5 along with 7nm Ryzen cores will be a beastly console.

Bottom line I am in 1060/580 territory and holding out for the next generation whoever brings it first. I could see it being a 7nm Vega or possibly a 7nm Navi or a 2060.
 
Reading around it seems a tweaked 7nm Vega sounds likely early 2019 with perhaps 7nm Navi at the back end of 2019? The multicore approach appears to be a long way off still but would be a game changer if they could get it to work.

I think I've been persuaded to sit on my 1070 until the next round of AMD cards appear.
 
Look at the Vega 20 for example. She presented the chip in May, and the only thing we know is an almost 21tflop 5nm Vega tiny behemoth already getting sampled. It's drivers are in Linux since early July. Yet nobody knows anything about it. Nada.

I'm hoping that is a typo and you meant 7nm. TSMC Fab 18 is still in final construction phase, Globalfoundries isn't even close to having 5nm production capability, Samsung will be sampling early next year and first 6 months is exclusive to in-house products. The only company AFAIK that can produce a 5nm sample for anyone at all right now is IBM who is only working with Samsung and GF to bring the node upto speed and not sampling for clients.

Reading around it seems a tweaked 7nm Vega sounds likely early 2019 with perhaps 7nm Navi at the back end of 2019? The multicore approach appears to be a long way off still but would be a game changer if they could get it to work.

I think I've been persuaded to sit on my 1070 until the next round of AMD cards appear.

Any kind of true MCM design that moves away from reliance on some kind of SLI/Crossfire type technology will need second generation 7nm at the very least to be remotely feasible and I can't see either company producing anything before they rework their architecture to accommodate that which is almost certainly going to be a two stage process with the first use of that architecture still monolithic cores and not implemented in any kind of multi-core way just laying the ground work.
 
Last edited:
Reading around it seems a tweaked 7nm Vega sounds likely early 2019 with perhaps 7nm Navi at the back end of 2019? The multicore approach appears to be a long way off still but would be a game changer if they could get it to work.

I think I've been persuaded to sit on my 1070 until the next round of AMD cards appear.

AMD just presented a 7nm card for professional users.
Navi or whatever 7nm is on the way asap.
Not unresonable they have a nice pipecleaner that runs 1080 etc speed at a normal level for price
 
I am interested to see what the next gen AMD APU brings to the table. These could be fab chips for low power mini pc projects


The current Apu is not *quite* there yet for me but if they can up the graphics performance by say 30% I think it would be.
 
I'm hoping that is a typo and you meant 7nm. TSMC Fab 18 is still in final construction phase, Globalfoundries isn't even close to having 5nm production capability, Samsung will be sampling early next year and first 6 months is exclusive to in-house products. The only company AFAIK that can produce a 5nm sample for anyone at all right now is IBM who is only working with Samsung and GF to bring the node upto speed and not sampling for clients.
.

My mistake. Zen 5 is going to be at 3nm, for around 2023
https://segmentnext.com/2018/05/15/globalfoundries-skip-5-nm-amd-zen-5/

And here is the latest on 7nm and AMD products
https://segmentnext.com/2018/05/15/7nm-amd-zen-2-vega-navi/

So on track for Navi & Zen 2 for 7nm next year, and 7nm euv in 2020 with Zen 3 and next gen GPU.
 
I'm looking forward to 2020. Should hopefully have 3 players in the market. I'm sure my 1080 will see me well until then.

+3


Hopefully that'll be when things become more competitive.

I wouldn't be surprised if that has had an impact on Nvidia pricing today. They're overpricing it while they can, The sad bit is that as I said when Pascal released, The competition are going to be watching the prices people are willing to pay & they'll want a piece of the pie, Price wars will happen but not straight away so early adopters of new ranges will be milked by everyone from now on.
 
+3


Hopefully that'll be when things become more competitive.

I wouldn't be surprised if that has had an impact on Nvidia pricing today. They're overpricing it while they can, The sad bit is that as I said when Pascal released, The competition are going to be watching the prices people are willing to pay & they'll want a piece of the pie, Price wars will happen but not straight away so early adopters of new ranges will be milked by everyone from now on.
Yes totally agree. AMD will likely come in £50-£80 cheaper when they release. But if nvidia has successfully been selling at £1300. Well... The AMD cards will likely still be £1250.

Honestly we only have ourselves to blame. Preordering cards at over a grand with no performance figures just confirms to nvidia that we are morons who will pay anything for anything.
 
Here is what's up with AMD.
What's also rumored is this is why we aren't seeing anything from AMD until 2020 time frame or thereabouts. Kinda hard to believe that to be honest. But once the news broke out that Navi was designed for consoles and why Raja left AMD it's starting to add up IMO.
Roughly 1.5 year hiatus from AMD on new graphics cards could indeed explain Raja Koduri's departure from AMD into Intel's loving arms.

Why Koduri left is simple enough, He was fired, He knew he would be which is why he was job hunting beforehand, The biggest surprise is that he's been able to make his way into Intel's employment after what he did at RTG. He couldn't deliver & instead of focusing on the right things he spent the last year or so of his time as head of RTG diverting r&d funds into irrelevant projects like the one back in his home country (Bollywood style VR Movie), pulling the wool over his superiors eyes, constantly diverting consumers focus away from Vega whenever it was meant to be about Vega (the holo-cube made several appearances, never to be heard from again), using any & every opportunity to obscure the situation they were in with Vega & he eroded the consumer trust AMD's GPU section had earned over the years. There's an endless list of things he did wrong & a list of things he did to deliberately obscure the facts regarding Vega. Why do you think it was that Vega's performance on release was around the same as it was on the Doom Vulcan demo from a year or so earlier? He left AMD high & dry blaming a lack of funding when it was his own inability to do the job they'd trusted him with. That's on Lisa Su and her reluctance to remove him long before she did. It was a train wreck long in the making, anyone with an objective view could see what was coming.

Cant wait for AMD's :
Trailers
Posters
HYPE Train

Just to underdeliver like they did EVERY SIGNE TIME pass Hawaii gpu....

Koduri's era is over, Things should slowly get better from now on.
 
I'm really hoping AMD hey some great cards out in the £250-£450 range. That would really be good for the enthusiast/keen gamer that can't justify £500+ for a GPU. Halo 4K cards are nice but there must be a huge amount more sales to be had at 1080p and 1440p.
 
GloFo & AMD decided that 5nm is not big improvement over 7nm+ so they will it jump.
Apparently going smaller than what Intel calls "5nm" may be physically impossible, according to some industry sources.

So whatever GloFo's "3nm" is, it's going to be >5nm in reality.

Which kind of makes all this "we're skipping 5nm" a bit farcical really.
 
Apparently going smaller than what Intel calls "5nm" may be physically impossible, according to some industry sources.

So whatever GloFo's "3nm" is, it's going to be >5nm in reality.

Which kind of makes all this "we're skipping 5nm" a bit farcical really.

So whatever we call it there isn't much room left before we run out of process shrinks to increase performance? Are we looking at mid 2020's then new ideas will be required?
 
Back
Top Bottom