• 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.

AMD VEGA confirmed for 2017 H1

Status
Not open for further replies.
Jesus, just how bad is the GloFo 14nm process??? Doesn't seem so great if this is true! Unless the NCU cores are true powerhouses (which is doubtful) :eek:
when asked AMD always say the process is great, but it's really not, they wouldn't admit it because , you know, partener and all.
but the main limit for Ryzen CPUs is the process, all cores SKUs 4-6-8 hit a hard wall at 4.1GHZ, and that's mainly due to glofo's 14nm, compared to 5Ghz of intel, or even 16nm's TSMC.
 
Last edited:
when asked AMD always say the process is great, but it's really not, they wouldn't admit it because , you know, partener and all.
but the main limit for Ryzen CPUs is the process, all cores SKUs 4-6-8 hit a hard wall at 4.1GHZ, and that's mainly due to glofo's 14nm, compared to 5Ghz of intel, or even 16nm's TSMC.
But when 3.8ghz of AMD matches and beats the 5ghz of intel what's the problem
 
4000 cores was a massive chip on 28nm. The Fury X was 596mm2. Amd's biggest ever chip I believe. On 14nm it would be somewhere around 300-350mm2 I think.

Something is up with the Vega NCU's:
All reports are saying that Vega is a 4096 NCU (CU) device, however the 14nm Vega die we have seen is bigger than expected at 475mm^2 to 525mm^2

https://www.reddit.com/r/Amd/comments/5m6uj3/vega_die_size_475mm2_xpost_rhardware/
https://www.computerbase.de/2017-01/amd-vega-preview/

Where did those extra transistors go?!?
 
when asked AMD always say the process is great, but it's really not, they wouldn't admit it because , you know, partener and all.
but the main limit for Ryzen CPUs is the process, all cores SKUs 4-6-8 hit a hard wall at 4.1GHZ, and that's mainly due to glofo's 14nm, compared to 5Ghz of intel, or even 16nm's TSMC.

But when 3.8ghz of AMD matches and beats the 5ghz of intel what's the problem

When does 3.8ghz of amd match intel at 5ghz core for core?
Yes i mean say 4core vs 4core or 8core vs 8core at those speed mentioned.

AlamoX as i see it wasnt knocking ryzen/vega just explaining how the difference Fabs/process's seem to be working out atm
 
but the main limit for Ryzen CPUs is the process, all cores SKUs 4-6-8 hit a hard wall at 4.1GHZ, and that's mainly due to glofo's 14nm, compared to 5Ghz of intel, or even 16nm's TSMC.

Well, the process and:
  • the architecture - physical layout of blocks and inter-connectivity
  • the detailed design (critical path length)
  • the fan-out on clock, data and signal paths
  • the clock structures - jitter, skew, etc
  • the power delivery and stability
  • power density/heat density
  • decoupling/ground bounce

These are important, but your architecture and detailed design are just as important. Issue is, if either one is lacking, you're stuffed.

I recon a better measure would be performance per transistor, although that is also difficult to compare due to the inherent differences between Intel and AMD. These include AMD being a SoC, Intel (Broadwell-E) having more PCIe and Memory controllers, Cache sizes etc.

Anyway, I think you're correct that GloFo process isn't great, however I doubt you can get more than 5% uplift by shifting to another process. There are just so many factors that determine your Fmax. What does "worry" me is that Ryzen 7 has 4.5B transistors vs. 3.2B for Broadwell-E (wikipedia :/)...
 
Last edited:

Something is up with the Vega NCU's:
All reports are saying that Vega is a 4096 NCU (CU) device, however the 14nm Vega die we have seen is bigger than expected at 475mm^2 to 525mm^2

https://www.reddit.com/r/Amd/comments/5m6uj3/vega_die_size_475mm2_xpost_rhardware/
https://www.computerbase.de/2017-01/amd-vega-preview/

Where did those extra transistors go?!?

It is useless for gaming. VEGA is more towards professional market and stuff they will reveled in there upcoming presentation. Like HBCC took quiet a die size of VEGA.
 
It is useless for gaming. VEGA is more towards professional market and stuff they will reveled upcoming presentation. Like HBCC took quiet a die size of VEGA.
5110101-ace+ventura+gif.gif
 
It is useless for gaming. VEGA is more towards professional market and stuff they will reveled upcoming presentation. Like HBCC took quiet a die size of VEGA.

HBCC sounds more complex than a standard HBM controller, but that is a lot of transistors! Billions!

Not convinced it is targeted for computer or gaming markets, where HBCC should be useful for both. Have to wait and see I guess.
 
Pricing of RX 580 in my country compare to GTX 1070.

RX 580 pricing in my country compare to GTX 1070
SAPPHIRE PULSE RX580 8GB RM1799
SAPPHIRE PULSE RX580 4GB RM1439
SAPPHIRE PULSE RX570 4GB RM1269
SAPPHIRE PULSE RX550 4GB RM569
SAPPHIRE PULSE RX550 2GB RM479

GTX MSI 1070 gaming X 8GB RM 1799.

In shot, it is DOA.
 
Pricing of RX 580 in my country compare to GTX 1070.

RX 580 pricing in my country compare to GTX 1070
SAPPHIRE PULSE RX580 8GB RM1799
SAPPHIRE PULSE RX580 4GB RM1439
SAPPHIRE PULSE RX570 4GB RM1269
SAPPHIRE PULSE RX550 4GB RM569
SAPPHIRE PULSE RX550 2GB RM479

GTX MSI 1070 gaming X 8GB RM 1799.

In shot, it is DOA.

Nah your country is just overpriced.
 
What Standards? you say that and yet history clearly proves you wrong, the Fury-X was still a pretty big chip, actually a little bigger than big Kepler as i have been corrected ^^^^

The 7970 was also much larger than the GTX 680. the 290X was a pretty big chip too.....

AMD are quite capable of making chips as big or even bigger than those nVidia make, when they want or need to, and they have.

Big Pascal is 600mm2^, a 14nm Fury-X isn't going to be anywhere near that big, AMD have plenty of room to grow into, another 50% probably for a 6000 core GPU.
I feel like you're deliberately misunderstanding what I'm saying.

I said 'in a given range'. How does Fury X prove me wrong? It's exactly what I'm talking about. 4000 core GPU and the big chip of the lineup. Same with the 7970.

It's not about who makes bigger chips - Nvidia or AMD. Like, you're bringing that comparison up out of nowhere and it's not at all relevant to anything I'm saying. I'm not saying Vega will be 600mm². I'm saying it will be the 'big chip' of AMD's current lineup.
 
Well from the possibility of the leaks being true, I really do hope AMD have a bigger chip coming and these die sizes and power requirements are wrong. Mind you, if Vega does give roughly the same performance as the 1080 in DX11 and a bit better in DX12 and is cheaper, that's a decent GPU and will cope well with 1080P and 1440P max settings gaming. AMD have the upper hand with DX12 from what we know and a few more DX12 titles will do them well in the long run.

Fingers crossed they price it well and it does give better performance than expected overall and they will have a winner.
 
What ever form Vega comes out in, its going up against an overclocked 1080Ti in my eyes. If AMD want to promise more later on account of optimizations, partnerships and API preference great but I'm interested in the here and now that IS the bottom line... no smoke with mirrors is going to deviate me from that. Water cool Vega and you're up against a water cooled 1080Ti end of story.

That doesn't matter to me, just the out of box experience and performance. I have no intention off pulling of coolers and setting up a custom loop or 3rd party hybrid cooler.

I just want to install it and get on with it; I also won't be overclocking.
So it's all down to price and performance out of the box for myself, and many.

Same here, Later optimizations are a bonus but I buy a gpu based on how it performs on the day I buy it and I prefer to buy an oc edition if it's available as I only ever put an overclock on my card temporarily if I'm benching something which is next to never.
 
Well from the possibility of the leaks being true, I really do hope AMD have a bigger chip coming and these die sizes and power requirements are wrong. Mind you, if Vega does give roughly the same performance as the 1080 in DX11 and a bit better in DX12 and is cheaper, that's a decent GPU and will cope well with 1080P and 1440P max settings gaming. AMD have the upper hand with DX12 from what we know and a few more DX12 titles will do them well in the long run.

Fingers crossed they price it well and it does give better performance than expected overall and they will have a winner.
im still expecting 1080 performance but would not be surprised by the time they sort the drivers out it will sit between a 1080 and 1080ti which could be good for everyone with price on amd and nvidia
 
Status
Not open for further replies.
Back
Top Bottom