Vega Rumors

I don't think they need Imagination technologies. They just need more time to get better at what they have. Shader wise they are fine. Front end wise have to some more work to do, they just need more geometry units. There needs to be a better balance of shader units vs geometry. Back end wise they just have to keep going at what they have and innovate.

nV got Maxwell because they incorporated what they learned from Tegra. Tegra already incorporated nV's GPU shader units. Tegra looks like a fail on paper (for cellphones), it wasn't a bad chip it was actually pretty good ahead of its time specially the ones after the first Tegra, but the lessons learned from it gave nV a huge advantage in the GPU market.

Buying another company out for IP really only works in the long run . Short term, never works out. It takes time for corporate values and structure to integrate well. Just look at AMD and ATi the first 5 years were hell, neither of them worked well together. Different marketing messages, different scopes. At the end, AMD CPU division made ATi their red headed step child by cutting their R&D even though the GPU division was paying for every screw up the CPU division had. AMD's values were messed up 10 years ago, and if the values are bad doesn't matter who does what, its all down hill.

Ya know AMD had a chance of buying out nV instead of ATi right? The reason that didn't happen was because Jensen wanted to be the CEO of the new company. AMD investors should look back at it and know how screwed up AMD management was at the time lol.
 
Last edited:
razor1 said:
Ya know AMD had a chance of buying out nV instead of ATi right? The reason that didn't happen was because Jensen wanted to be the CEO of the new company. AMD investors should look back at it and know how screwed up AMD management was at the time lol.
Click to expand...

Given how much better Nvidia could have performed with x86 against Intel compared to AMD vs Intel perhaps it's NVidia stockholders who actually took the comparative bath because of Jensen. Sure Nvidia did well, but how much better could it have been?

Silly what if game.
 
well one thing that changed AMD's future after the buyout of ATi was the the r600 flop. And then the bad management discussions again with trying to cut down costs on making BD, then the even more idiotic move by cutting down ATi and going small die. Then the R&D budget cuts, that's when all of AMD's sr. graphics engineers jumped ship.

All of these things would never happened if it was nV instead of ATi....... Half the problems are right there. Then we have HBM lock in, then we have AMD's refusal of nV chips sets which were a hot commodity for them prior to their GPU business falling apart.

We can link everything back to the AMD's management discussions.

If they merged with nV instead of ATi, AMD would have been out of debt 2 years after the launch of the g80, if they so wished it. They never had that chance with the ATi buyout, not only did they not have that chance they went into more debt and with that debt didn't change their ways of doing business. If Jensen was there you think he would have done the same shit AMD's Roy Reed did? I see there would have been NO way he would have done that. Just for one reason, they would have had enough money not to do a BD ;)
 
Last edited:
Elios said:
so overclocked to hell and back with a almost 500W power draw it JUST about trades blows with a 1080... for 2x the price....
Click to expand...


Well it should be faster than the 1080 at that point, but still power draw more than double that of the 1080 (well 3 times).
 
razor1 said:
Well it should be faster than the 1080 at that point, but still power draw more than double that of the 1080.
Click to expand...

ok fair at that point it would fall between a 1080 and 1080ti but still your looking at 1k card + water cooling or the hybrid card at 1.5k... JUST to get a bit better then 1080 speeds and still need 490w to 500w JUST for the card
Pascal Titan X is a better buy at that point

hell i bet my 1070 OCed at 2.1Ghz is faster then the top of the line "Gaming" Vega at this point

what i dont get is why people thing a consumer card that will cost 1/2 as much is going to be faster the prosumer card

Vega FE is clearly the halo sku i dont see any of the consumer cards being any where near as fast
 
Last edited:
maybe equal... doesn't really matter, anyone doing that much modding, will understand its so much damn easier to do it with a Pascal card and get more out of them while not spending more money on their power supply or cooling.
 
razor1 said:
maybe equal... doesn't really matter, anyone doing that much modding, will understand its so much damn easier to do it with a Pascal card and get more out of them while not spending more money on their power supply or cooling.
Click to expand...

exactly.. it simply doesn't worth the hassle... it wasn't as the HD7900s or R9 290s where all the investment on liquid cooling, extra power, extra PSU was well worth rewarded with a lot of performace.. with VEGA it's lot of worth for no valid reason.
 
Araxie said:
exactly.. it simply doesn't worth the hassle... it wasn't as the HD7900s or R9 290s where all the investment on liquid cooling, extra power, extra PSU was well worth rewarded with a lot of performace.. with VEGA it's lot of worth for no valid reason.
Click to expand...

Yeah.. only place high end cooling might be worth it anymore is on the real high core count CPUs. Too bad since custom water was kind of fun.
 
1 question: When Nvidia introduces their own shader intrinsics, how will AMD stay relevant? Vega will be relegated to competing with 1060, and won't even touch a 1070.
 
I suppose we are hoping that RTG driver team really messed up and now has their act together to get a decent set of drivers that will unlock the Vega Hardware for games in 1 month. Maybe Vega hardware in reality is sufficiently different, new flows, new processes that really does need significant time for the driver team to catch up. GCN has not changed that much since Tahiti, Polaris was the biggest change I could tell. The more likely scenario is the drivers are sufficient to run most of the hardware efficiently. We have seen running Vega cards last December if not November - that is a hell a lot of time to June launch of the FE.

Looks like ~ 1070, 1080 and above 1080 performance. Which can be sufficient to maintain a presence in the market for the next generation to hit as long as the price is maintained good for the performance (meaning reductions when Nvidia reduces prices if they do and when Volta hits). I hope RTG has full burners on now for Navi. For Navi they will need the HBM2+, interposer production capability and lower cost due to manufacturing experience and volume. RTG in on a long journey, while the cpu side looks good on recovery, RTG may now have to recover as well.
 
the one thing NV has proved is with GDDR5X and now 6 is you dont need HBM for consumer cards ... this is really i think going to hurt AMD in the long run on the cost side had Vega not used HBM they likely could cut costs down another 100 to 150 bucks which would made the cards and ok value
 
Well secondary thing too, AMD in Jan of 2019 or 60 days prior something like that, I think it was has to have 600 million in cash in a specified account to cover the first portion of the bonds it took out.

They can't increase R&D until they get vastly into the black, just not possible. I think what we are seeing is a combination of bad choices on the ram side and pretty much desperation on the business side for RTG. (also why we haven't see much increase in R&D even after Ryzen seems to be a success) Its not that AMD doesn't give a crap about RTG, its just they don't have a choice but to sacrifice RTG in the short term till either they get to above 2 billion gross profit per year. If they get to 2 billion in the black per yer (500 mil per quarter) they can sustain current cash balances while ensuring current operating expenses. In 2020 the 2nd batch of bonds come to maturity which is 500 million, and in 2021 the final batch of bonds must be repaid. So they aren't out the shit hole yet. So far the reason AMD has lasted this long they have sold parts of the themselves but really not much more left to AMD to sell off unless its core tech which is never a good idea.

So pretty much as these bonds mature, they need to have that cash in hand and have the money to keep current assets, and operating expenses sustained. Until that is done, R&D will not go up very much for RTG.
 
Last edited:
To me, the 400W draw from Gamers Nexus at 1700MHz with their hybrid cooling loop seems very close to what rumors have the RX Vega at. While these types of articles aren't always accurate, the leaks claim that the gaming RX version will be all in one cooled with a TDP of 375W which matches well. Gamer Nexus results with their hacked together liquid cooled FE means that I'm expecting the leaks to be accurate with the top card to be clocked at around 1700 MHz boost with the liquid AIO cooler much like a Fury X implementation.

https://videocardz.com/70890/amd-radeon-rx-vega-xtx-vega-xt-and-vega-xl
 
Elios said:
so overclocked to hell and back with a almost 500W power draw it JUST about trades blows with a 1080... for 2x the price....
Click to expand...

Okay I am not defending AMD here I think Vega will likely be a disappointment, but this is something I hate in forums. First of all its 400w, now thats not great but you kinda exaggerated. Second for goodness sake stop comparing this to gaming vega. AMD has said this wait for gaming vega for gaming. You can't bitch about price until you compare it to gaming vega price. Now those who bought this, bought this knowingly for whatever the fuck reason they bought it for. I for once will never touch this card at that price.
 
400W at the PCIE power cables + the 75w or so from the slot so yeah close to 500w
 
razor1 said:
Well secondary thing too, AMD in Jan of 2019 or 60 days prior something like that, I think it was has to have 600 million in cash in a specified account to cover the first portion of the bonds it took out.

They can't increase R&D until they get vastly into the black, just not possible. I think what we are seeing is a combination of bad choices on the ram side and pretty much desperation on the business side for RTG. (also why we haven't see much increase in R&D even after Ryzen seems to be a success) Its not that AMD doesn't give a crap about RTG, its just they don't have a choice but to sacrifice RTG in the short term till either they get to above 2 billion gross profit per year. If they get to 2 billion in the black per yer (500 mil per quarter) they can sustain current cash balances while ensuring current operating expenses. In 2020 the 2nd batch of bonds come to maturity which is 500 million, and in 2021 the final batch of bonds must be repaid. So they aren't out the shit hole yet. So far the reason AMD has lasted this long they have sold parts of the themselves but really not much more left to AMD to sell off unless its core tech which is never a good idea.

So pretty much as these bonds mature, they need to have that cash in hand and have the money to keep current assets, and operating expenses sustained. Until that is done, R&D will not go up very much for RTG.
Click to expand...

Absolutely agree. i think they might go gddr6 with navi hoping they learn from this shit and for sake of making more profit if anything.
 
Elios said:
got proof of that? 425 is still REALLY REALLY BAD
Click to expand...

dude they are pushing the card, and 425 in not 500. next google it and read up. They are pushing the cards beyond its specs you expect it to use less watts? Its like kicking a dead horse and hoping it wakes up. Power is not the point here, whole point of that article is get hits thats it. Also gaming vega will probably cost less then FE. So price shouldn't even be compared to 1080. AMD has been very clear to let gamers know don't buy this card for gaming. If you still buy it for gaming then its on the buyer not amd. Cuz I wouldn't touch that shit for that price.
 
the point is its pulling almost 500W of power JUST get slightly better then 1080 performance

i would not expect consumer Vega to come any where near 1080 numbers btw why sell a card for 1/2 the price thats just as fast? even if it has a bit less ram
 
Elios said:
the point is its pulling almost 500W of power JUST get slightly better then 1080 performance

i would not expect consumer Vega to come any where near 1080 numbers btw why sell a card for 1/2 the price thats just as fast? even if it has a bit less ram
Click to expand...

jesus christ. This card is not for gamers. I don't know how rx vega is and I don't care. LOL but I like to stay objective and go by what amd has stated and not judge until product is out. FE is a weird launch its not for gaming which amd made clear for gaming. I am sure they were expecting everyone to bench it. They also said rx vega will be much better optimized for popular gaming titles and faster then FE. So if its little faster then FE it will trade blows with 1080 likely. As to price, we don't know why don't you wait another 2 weeks and find out.

and you are exaggerating again. Who here said vega will be half the price of 1080. WTF? thats 250 lol. You expect vega to be 250? Then rx 580 will be free ROFL!
 
razor1 said:
Well secondary thing too, AMD in Jan of 2019 or 60 days prior something like that, I think it was has to have 600 million in cash in a specified account to cover the first portion of the bonds it took out.

They can't increase R&D until they get vastly into the black, just not possible. I think what we are seeing is a combination of bad choices on the ram side and pretty much desperation on the business side for RTG. (also why we haven't see much increase in R&D even after Ryzen seems to be a success) Its not that AMD doesn't give a crap about RTG, its just they don't have a choice but to sacrifice RTG in the short term till either they get to above 2 billion gross profit per year. If they get to 2 billion in the black per yer (500 mil per quarter) they can sustain current cash balances while ensuring current operating expenses. In 2020 the 2nd batch of bonds come to maturity which is 500 million, and in 2021 the final batch of bonds must be repaid. So they aren't out the shit hole yet. So far the reason AMD has lasted this long they have sold parts of the themselves but really not much more left to AMD to sell off unless its core tech which is never a good idea.

So pretty much as these bonds mature, they need to have that cash in hand and have the money to keep current assets, and operating expenses sustained. Until that is done, R&D will not go up very much for RTG.
Click to expand...

All images taken from CFO Commentary from on AMD's Quarterly Earnings page. Choose the PPT, or Powerpoint file for the pretty slides.

j31hrNG.png

Album: http://imgur.com/a/X5or3

Includes the Debt Maturity Profile slide for all quarters of 2016 and Q1 2017, along with the Q3 slide describing AMD's stock market moves to restructure/pay their debt.

Edit: I'm sorry, I just realized that what I posted didn't really say much. I've changed the images above to something easier to understand. The stock maneuvers they performed in Q3 2016 allowed AMD to basically reduce their 2019 debt repayment to $196 million, and push back and reduce their long-term debt amounts by a good bit. They obviously need to be making cash, but 2019 won't be the cliff edge.
 
Last edited:
ah ok cool thats how they reduced the debt, couldn't find those files lol, was going based off of old info.
 
NKD said:
jesus christ. This card is not for gamers. I don't know how rx vega is and I don't care. LOL but I like to stay objective and go by what amd has stated and not judge until product is out. FE is a weird launch its not for gaming which amd made clear for gaming. I am sure they were expecting everyone to bench it. They also said rx vega will be much better optimized for popular gaming titles and faster then FE. So if its little faster then FE it will trade blows with 1080 likely. As to price, we don't know why don't you wait another 2 weeks and find out.

and you are exaggerating again. Who here said vega will be half the price of 1080. WTF? thats 250 lol. You expect vega to be 250? Then rx 580 will be free ROFL!
Click to expand...

It said right now AMD's website that Radeon Vega Frontier Edition is for gaming.

  • I Just Want to Game
    If you want to game on a Windows® platform, get your driver here and switch to “Gaming Mode.”
    Click to expand...
 
Elios said:
got proof of that? when overclocked? 425 is still REALLY REALLY BAD
Click to expand...

Not to defend Vega but you do realize 2 8 pins connection is suppose to draw according to specification is 300 watts + 75 watts from PCIE. So I am really confuse how you gotten 500 watts,
 
_mockingbird said:
It said right now AMD's website that Radeon Vega Frontier Edition is for gaming.
Click to expand...

Did I argue FE didn't have a gaming mode? I am not saying vega might be super good. Ofcourse it has gaming mode, no shit! everyone has said that! There is absolutely no difference when it comes to gaming mode or pro mode. Gaming mode turns on more settings in control panel. Dude I am not arguing the fact there might be no difference.

I am going by repeated AMD statements that RX vega will have better optimizations. Who the fuck in their right mind would buy FE for gaming when RX vega will be half the price. if AMD is telling you to wait for rx vega for gamin is it too hard to do? Regardless of performance to continuously hammer the fuckin point that amd is asking you to pay 1000 dollars for a product that is not recommended for gaming. When there will another one launched in 2 damn weeks that might be half the price. Yea twist the statement about turn on gaming mode to amd saying FE is specifically for gaming. Ignorance has no cure, sorry. If you don't understand me I apologize.
 
razor1 said:
ah ok cool thats how they reduced the debt, couldn't find those files lol, was going based off of old info.
Click to expand...

Pretty ashamed to admit it, I was wondering how much the stock dilution helped their financials, scoured news on the quarterly reports, couldn't find it, finally started through the reports themselves, damn lines were staring at me the entire time.

During Q3 2016, we raised approximately $1.4 billion in cash, before issuance costs, as a result of issuing $690 million of Common Stock and $700 million of Convertible Notes due 2026. The $690 million includes $90 million from the underwriters’ option to purchase 15% or 15 million additional shares of common stock which was exercised in Q3 2016. The net proceeds from these capital markets transactions were used to retire approximately $796 million in principal of outstanding term debt and pay off the outstanding ABL balance of $226 million.
Click to expand...

AMD sold the $690 million of Common Stock at a stock price of $6 and sold their Convertible Notes at $8. Needless to say, if their stock had not rose, I'm sure we would be hearing the impending cries of "bankruptcy" for AMD again. Your entirely correct with your previous comment, AMD managing the hype into these gains was a smart move.
 
Chimpee said:
Not to defend Vega but you do realize 2 8 pins connection is suppose to draw according to specification is 300 watts + 75 watts from PCIE. So I am really confuse how you gotten 500 watts,
Click to expand...


GN got readings of ~400w at the PCIE cables when oc'ed to 1700mhz add the 75w its likely pulling from the slot and your getting close to 500w
 
Elios said:
GN got readings of ~400w at the PCIE cables when oc'ed to 1700mhz add the 75w its likely pulling from the slot and your getting close to 500w
Click to expand...

How many times do I have to say it is not pulling 75w from slot. Since you keep ignoring the fact and ignore to look it up. Here straight from PC perspective. Let me help you so you can stop ignoring it and know the truth.

"In general, the power draw of the Vega FE is under the 300-watt level, which is good. After the issues that haunted the Radeon RX 480 at its launch (drawing well over the 150 watts rated TDP), AMD has learned its lesson it appears. Power draw from the PCIe connection on the motherboard stayed very low actually (near 25 watts) with the vast majority of the power coming from the dual 8-pin power connections on the card. For reference, that power draw of 300 watts is 50 watts higher than the GeForce GTX 1080 Ti and the Titan Xp but 120 watts higher than the GeForce GTX 1080"

https://www.pcper.com/reviews/Graph...B-Air-Cooled-Review/Clock-Speeds-and-Power-Co

Reading is fundamental!
 
Last edited:
NKD said:
Absolutely agree. i think they might go gddr6 with navi hoping they learn from this shit and for sake of making more profit if anything.
Click to expand...
not going to happen. AMD has stated already they have plans with the Cache streaming thing and SSD. HBM Bandwidth is the whole foundation of this idea. I am sure they will still have the GDDR-X on some cards but HBM is too much a part of the design to scrap at this point.
 
NKD said:
How many times do I have to say it is not pulling 75w from slot. Since you keep ignoring the fact and ignore to look it up. Here straight from PC perspective. Let me help you so you can stop ignoring it and know the truth.

"In general, the power draw of the Vega FE is under the 300-watt level, which is good. After the issues that haunted the Radeon RX 480 at its launch (drawing well over the 150 watts rated TDP), AMD has learned its lesson it appears. Power draw from the PCIe connection on the motherboard stayed very low actually (near 25 watts) with the vast majority of the power coming from the dual 8-pin power connections on the card. For reference, that power draw of 300 watts is 50 watts higher than the GeForce GTX 1080 Ti and the Titan Xp but 120 watts higher than the GeForce GTX 1080"

https://www.pcper.com/reviews/Graph...B-Air-Cooled-Review/Clock-Speeds-and-Power-Co

Reading is fundamental!
Click to expand...
Look like another member campaigning for an IGNORE addition. Seriously he keeps spouting the exact same thing over and over despite any response to the contrary. HARD does not care about POWER. His 4790K@ 4.8ghz does not = saving the planet. My OCed 8350 can pull upwards of 400W, I don't care. If pulling 400W doesn't do it for you, easy solution, DONT BUY IT.

Good luck with your future response to said individual.
 
JustReason said:
not going to happen. AMD has stated already they have plans with the Cache streaming thing and SSD. HBM Bandwidth is the whole foundation of this idea. I am sure they will still have the GDDR-X on some cards but HBM is too much a part of the design to scrap at this point.
Click to expand...

Doesn't seem likely since Vega's actual bandwidth is lower than that of Fury X and 1080Ti...
 
Elios said:
what i dont get is why people thing a consumer card that will cost 1/2 as much is going to be faster the prosumer card
Click to expand...
Because 4-Hi stacks should be able to clock higher due to thermal limitations. At stock settings consumer cards can use more power as well. With really good cooling, FE should be the fastest, but not with stock settings.

noko said:
Maybe Vega hardware in reality is sufficiently different, new flows, new processes that really does need significant time for the driver team to catch up. GCN has not changed that much since Tahiti, Polaris was the biggest change I could tell. The more likely scenario is the drivers are sufficient to run most of the hardware efficiently.
Click to expand...
That's a given with the driver and LLVM changes AMD has been making. Many comments of "Vega does x instead". Not to mention all the tests showing measurable differences from Polaris and Fiji.

noko said:
RTG in on a long journey, while the cpu side looks good on recovery, RTG may now have to recover as well.
Click to expand...
That was expected as AMD indicated they shifted resources from RTG to CPUs as Ryzen/Epyc cover a larger, high margin market.

Probably the right move, but came at the cost of no high end Polaris, low end Vega(excluding APUs), and engineers for drivers. Vega also seems designed around all of Zen's IO so they needed that. Both designs seem complementary.

NKD said:
Absolutely agree. i think they might go gddr6 with navi hoping they learn from this shit and for sake of making more profit if anything.
Click to expand...
GDDR6 won't happen. It takes too much space, more power, less speed, and the only market it might have made sense on is low to mid. At which point GDDR5 or lower cost D4 would be used if a sufficiently large APU doesn't exist.

With nonvolatile memory starting to arrive, having HBM(or something similar) on the CPU makes a lot of sense. So that same low to mid market where they may have used GDDR6 is now HBM with APUs. Along with cheap Polaris rebrands for cost sensitive OEMs. That's if HBM based APUs aren't cheaper than the discrete option.
 
I don't know anything about Vega. I don't know anything about coding or DSBR or LLAPI or thread branching or GDDR5x or GDDR6x or HBM2 or AMD stock swaps.

I do know that "your" and "you're" are two different words. One is a contraction of "you are" and the other one is a possessive indicating that something belongs to, or is associated with, you. 'You're wrong in your use of "your" instead of "you're".'

(If this tip was helpful, please feel free to send me a new Vega card. ;) )
 
c3k said:
I don't know anything about Vega. I don't know anything about coding or DSBR or LLAPI or thread branching or GDDR5x or GDDR6x or HBM2 or AMD stock swaps.

I do know that "your" and "you're" are two different words. One is a contraction of "you are" and the other one is a possessive indicating that something belongs to, or is associated with, you. 'You're wrong in your use of "your" instead of "you're".'

(If this tip was helpful, please feel free to send me a new Vega card. ;) )
Click to expand...

No it does not.

AMD fanboys Hate Nvidia Fanboys.

Nvidia Fanboys Hate AMD Fanboys

But, everyone hates people who correct grammar on forums.
 
Don't really care too much about power when gaming, mining yes, gaming no. Cost and performance will be what drives my decision most of the time when it comes to gaming. I can see why some people would care, as they are running in mAtx, or silent builds, or their PSU is not beefy enough to handle it and they don't want to upgrade(cost again!)
 
You must log in or register to reply here.
Back
Top