Overclock.net banner

[Videocardz] AMD Vega: 16GB HBM2 and 1600 MHz clock

28K views 354 replies 92 participants last post by  KarathKasun 
#1 ·
Quote:
The new ID just popped up on CompuBench database. It is undoubtedly Vega: it was detected as gfx900 architecture, has 64 Compute Units and has new IDs confirmed by Linux patch to be Vega.

The reason why I didn't name this post "Radeon RX Vega" is because we don't know if this chip is Radeon RX, Radeon Pro or Radeon Instinct. It would work as a clickbait, but it might be misleading. AMD only confirmed two stack HBM2 for Radeon RX Vega and showcased a sample with 8GB on board. We simply don't know if the 16GB HBM2 chip is planned for Radeon RX series.
https://videocardz.com/69475/amd-radeon-vega-spotted-with-16gb-memory-and-1600-mhz-clock
 
#3 ·
Quote:
Originally Posted by TheReciever View Post

When do we get official news?
Computex Press Conference

16 days 04 hrs 43 min 28 sec

maybe
 
#4 ·
It's quite possibly the dual chip model...and if so, that's very promising, because historically they have always been clocked a bit lower than their single card brethren (Can't think off the top of my head if this trend had continued with the Pro Duo, I believe that was essentially two nano's on the one PCB?).
 
#6 ·
Quote:
Originally Posted by GHADthc View Post

It's quite possibly the dual chip model...and if so, that's very promising, because historically they have always been clocked a bit lower than their single card brethren (Can't think off the top of my head if this trend had continued with the Pro Duo, I believe that was essentially two nano's on the one PCB?).
No, the max value on ram would only be 8gb on the settings because a dual chip would be dual 8gb ram.

We had news that AMD was making a twin-chip version of the Vega PRO at 50tflop as an accelerator device - so it goes to be expected that we will see a dual chip vega, but whether that's actually a graphics card is unknown.

What will be interesting is how well the multiple op-per-clock on the shaders will work. At 1600 with reliable two-per-clock having an effective 8192 shader count will be astounding.
 
#7 ·
Quote:
Originally Posted by GHADthc View Post

It's quite possibly the dual chip model...and if so, that's very promising, because historically they have always been clocked a bit lower than their single card brethren (Can't think off the top of my head if this trend had continued with the Pro Duo, I believe that was essentially two nano's on the one PCB?).
Yes, Pro Duo had a clock of 1000Mhz and Fury X 1050Mhz. So 1650Mhz for single Vega?
 
#10 ·
This leak is fake. This is a perfect example of how fake news work. One random guy posts a completly fake RX Vega leak on reddit. One website called digiworthy takes the bait and writes a newsarticle/clickbait about it, without mentioning that the "source" is a random reddit comment.... Then tweaktown copies this article and they list as source digiworthy, and the cycle continues, no one will know who the original source was anymore, just that it is supposed to be a leak ... if people would at least properly write their sources, would anyone actually believe this "leak" if they mentioned that their source is a random reddit comment....
biggrin.gif
 
#14 ·
It's starting to feel like Vega is gonna release bundled with Half-Life 3.
 
  • Rep+
Reactions: LancerVI
#17 ·
If it is 1600MHz then yeah... it will beat a GTX 1080 Ti. Period. May even beat a TitanXp. Looks like there will be a Dual Vega (hence 16GB HBM2) and a single high performance Vega as well as a GTX 1070 and GTX 1080 competitor.

That single Vega, if 1600MHz with 8GB of HBM2, will most likely beat a TitanXp. Those clocks are very high, much higher than I expected.
 
#19 ·
#20 ·
Quote:
Originally Posted by Mahigan View Post

If it is 1600MHz then yeah... it will beat a GTX 1080 Ti. Period. May even beat a TitanXp. Looks like there will be a Dual Vega (hence 16GB HBM2) and a single high performance Vega as well as a GTX 1070 and GTX 1080 competitor.

That single Vega, if 1600MHz with 8GB of HBM2, will most likely beat a TitanXp. Those clocks are very high, much higher than I expected.
1.2GHz same speed as 1070 but beat 1080 Ti at 1.6Ghz, nah. Still not enough. Go check out how much faster is 1080 Ti compared to 1070 then use the factor to scale the clock speed. It will be close to 2GHz.

Unless we are talking about cut down Vega clocked at 1.2GHz which is comparable to 1070. Then yes, A full blown Vega with 1.6GHz can stand a good chance against 1080 Ti.
 
#21 ·
Quote:
Originally Posted by guttheslayer View Post

1.2GHz same speed as 1070 but beat 1080 Ti at 1.6Ghz, nah. Still not enough. Go check out how much faster is 1080 Ti compared to 1070 then use the factor to scale the clock speed. It will be close to 2GHz.

Unless we are talking about cut down Vega clocked at 1.2GHz which is comparable to 1070. Then yes, A full blown Vega with 1.6GHz can stand a good chance against 1080 Ti.
Well hell even if you did the math with Polaris scaling it up, and tack on a 1.6ghz clock speed. You would wind up with something faster than a 1080ti.

Which is why no one expected 1.6ghz.

Also don't forget AMD's Arch Scales better with clock speed than Nvidia does.
 
#22 ·
Quote:
Originally Posted by guttheslayer View Post

1.2GHz same speed as 1070 but beat 1080 Ti at 1.6Ghz, nah. Still not enough. Go check out how much faster is 1080 Ti compared to 1070 then use the factor to scale the clock speed. It will be close to 2GHz.

Unless we are talking about cut down Vega clocked at 1.2GHz which is comparable to 1070. Then yes, A full blown Vega with 1.6GHz can stand a good chance against 1080 Ti.

http://www.tweaktown.com/news/57544/amd-radeon-rx-vega-nova-599-beats-gtx-1080-ti/index.html

Take it with a grain of salt. Hence my underlined IF
 
#24 ·
Quote:
Originally Posted by aliquis View Post

This leak is fake. This is a perfect example of how fake news work. One random guy posts a completly fake RX Vega leak on reddit. One website called digiworthy takes the bait and writes a newsarticle/clickbait about it, without mentioning that the "source" is a random reddit comment.... Then tweaktown copies this article and they list as source digiworthy, and the cycle continues, no one will know who the original source was anymore, just that it is supposed to be a leak ... if people would at least properly write their sources, would anyone actually believe this "leak" if they mentioned that their source is a random reddit comment....
biggrin.gif
Pretty much the only good post in this thread
thumb.gif
 
#25 ·
Quote:
Originally Posted by aliquis View Post

This leak is fake. This is a perfect example of how fake news work. One random guy posts a completly fake RX Vega leak on reddit. One website called digiworthy takes the bait and writes a newsarticle/clickbait about it, without mentioning that the "source" is a random reddit comment.... Then tweaktown copies this article and they list as source digiworthy, and the cycle continues, no one will know who the original source was anymore, just that it is supposed to be a leak ... if people would at least properly write their sources, would anyone actually believe this "leak" if they mentioned that their source is a random reddit comment....
biggrin.gif
Except this isn't from a Reddit sub-board thing, it's a database leak and shows true, valid info. If you had read the article instead of just posting what you THINK you know, then perhaps you wouldn't look like such a fool.
tongue.gif
 
  • Rep+
Reactions: LAKEINTEL
#26 ·
Financal analysis day is tomorrow and should shed some light into product positioning for this quarter and year.
 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top