Editorial AMD Didn't Get the R9 Fury X Wrong, but NVIDIA Got its GTX 980 Ti Right

[Deleted member 158293]

Thought this article was balanced and well written, well done.

 

[Basard]

In other words, Give you a card with a full waterblock pre-installed like the EVGA's 'Hydro Copper' series of cards.

They could have but then again it will cost more and AMDs only real advantage at the moment is its price compared to the 980Ti.

Cards with full waterblocks pre-installed will be available shortly with the Fury X. I heard AMD didnt allow the other vendors to make their own custom cooling solutions so thats why a lot of the cards are physically exactly the same.

They already have a "full" water block preinstalled.... Just not a 'customizable' full waterblock....

I don't see how it would cost more, seeing as it wouldn't include a pump AND radiator AND fan.....

I would gladly work at any OEM for 16 bucks an hour.... telling them what water block to put on what..... I would definitely get a raise because my advise would be so godly.....

Most people have no clue how to set up their own water cooling setup in their case..... So therefore we get AOI solutions......

And here comes the "fury" cards, all set up for 'real' men to cool and power all on their own..... regular furies will be better, at least i hope so.... god only knows how epic they will be at 14nm!

(edit) ok, yeah, it will cost more to do what i was explaining, but shit, people with 650+ to spend would want the AIO setup to be removable.....

 

[newtekie1]

Yes, it affects single cards, and it happens more than once an hour. However, most systems that do have the problem frequently are OC'ed, and anyone doing testing on an OC'd system can go toss a bone, as that's stupid testing methods (control needs to eliminate OC as part cause).

Ah, I wasn't sure because I only had one driver crash on my GTX960 system, and I wasn't sure if it was from an unstable overclock or the bug. Either way, it's fixed now, and I'd still say it was a lot more minor than horrible performance at resolutions other than 4K.

 

[cadaveca]

Ah, I wasn't sure because I only had one driver crash on my GTX960 system, and I wasn't sure if it was from an unstable overclock or the bug. Either way, it's fixed now, and I'd still say it was a lot more minor than horrible performance at resolutions other than 4K.

One of the crashes completely hosed my OS on a stock system. That's hardly minor in my books. But that's one crash out of 100's I've had, and I do proper back-ups so that data loss wasn't that bad other than having to re-image the drive to get back up an running.

 

[RejZoR]

It was minor because the crashes were rare. It wasn't like it was a constant thing, making Chrome unusable, it happened maybe once an hour at the worst case. It affected SLI setups worst than single card setups(I don't even know if it actually even affected single GPU setups).

If it's constant, you at least know it'll happen. If it's not, it's even worse... Besides, a typical NVIDIA denial. Same as observed back in Windows Vista days when majority of BSOD's were caused by the NVIDIA driver and yet people were blaming Microsoft for it, you know, because it's BSOD, it must be a Microsoft thing. Except those BSOD's had NV*.sys mentioned on it...

 

[Yorgos]

Editorial / Opinion. Keep it civil.

go get yourself a techpowerup.blogspot.com


it's still under "news" though.

sorry dude that you have to hear this from me, but you have no idea, other than what the specs pdf says.
I'd like to hear an architecture break down from your "pen" and see there how much your knowledge upon the subject take effect.

'till that time comes, please stop posting your opinions, we can read the specification documents on our own and everybody can make similar assumptions.

 

[Frick]

go get yourself a techpowerup.blogspot.com


it's still under "news" though.
View attachment 66123

sorry dude that you have to hear this from me, but you have no idea, other than what the specs pdf says.
I'd like to hear an architecture break down from your "pen" and see there how much your knowledge upon the subject take effect.

'till that time comes, please stop posting your opinions, we can read the specification documents on our own and everybody can make similar assumptions.

Wow. The only thing I'll comment on is that the News section on the forums covers the front page, so it belongs there. Editorials are great when they are advertised (or something) as such.

 

[john_]

It was minor because the crashes were rare. It wasn't like it was a constant thing, making Chrome unusable, it happened maybe once an hour at the worst case. It affected SLI setups worst than single card setups(I don't even know if it actually even affected single GPU setups).

Having crushes once an hour for something that you could be doing possibly for many hours a day it's not minor. Fire up Chrome, load TPU, write a big post at someone and have your browser crush just before posting and that's only one example.

 

[R-T-B]

go get yourself a techpowerup.blogspot.com


it's still under "news" though.
View attachment 66123

sorry dude that you have to hear this from me, but you have no idea, other than what the specs pdf says.
I'd like to hear an architecture break down from your "pen" and see there how much your knowledge upon the subject take effect.

'till that time comes, please stop posting your opinions, we can read the specification documents on our own and everybody can make similar assumptions.

Editorials should probably have their own section, but it's always been done this way. They ARE tagged "editorial" if you pay attention.

 

[john_]

data loss wasn't that bad other than having to re-image the drive to get back up an running.

LOL. OK.

It's really funny. I am sorry but it is. You people even when you agree that a specific Nvidia bug is not as minor as some want others to believe, even if it was involving data loss and re-imaging, even then "Well, OK it wasn't that bad, I had a back up".

If that was an AMD bug I would e reading all over the place
"NEVER AGAIN AMD" and 50 likes in every post of this kind.

 

[haswrong]

IMHO, people that buy the fury x need to have their brain checked out. It's same priced as the 980ti, it's hotter, slower, weaker, less powerful, consume more electricity, lack of driver support/update, takes more space on your system cause of the watercooler unit, and louder.
Why on earth people want to buy a fury x ? Other then they're dumb or a blind amd fanboy.

It's just an opinion, lets keep it civil

it has 4096 shaders. maybe its not so less powerful, only amd forgot to show how it can be used efectively.

...
EDIT: If AMD could have shown how great an overclocker it is - I'd be buying one. That for me was the only real weakness of it. It trades blows with 980ti at my current gaming resolution but the overclock killed it for me.

that too is probably one nail in the fury coffin, as amd promised it had been designed with overclockers in mind. a very unfortunate untrue hype gamble doomed from the beginning.

but do not despair, there will be the dualchip card, which will be the interesting thing, especially with the possibility of doubling the memory capacity for resources under dx12 / opengl review. the only drawback i fear is if you program a game yourself, you will have to ask amd to add a stupid profile for it in the drivers. shouldnt this work automatically?

 

[btarunr]

Don't be this guy:

go get yourself a techpowerup.blogspot.com .

I am a content head of this site. I set its news policy. I can decide to post hello-kitties on the front page, and it will be there. It will stay there. The "editorial" tag allows me to post, well, editorials. If you're good at reading specs sheets, then don't waste your time commenting on our, well, my content.

 

[RejZoR]

And so I have sold my soul to the green devil. Ordered the ASUS Strix GTX 980 OC. Looks like a good overall package with slightly higher price. The Ti version was a bit much, I'll rather replace it one more time instead when Pascal and Arctic Islands come out. If even needed...

 

[Vayra86]

I think the who won is consumers. Competitive gfx is good for everyone.

With $650 dollar price tags, I don't think so.

We haven't won anything. Look at the pricing of cards in both camps and you know why.

- We are still on 28nm
- We are paying more for a rebrand with a bit more VRAM, a full year after the original was released
- Nvidia prices are not dropping, while Maxwell is cheaper per % performance than Kepler. Look at the 960 and its price point, and compare it to the 660 on release. We get lesser 'metal' for a price premium
- AMD prices have gone up across the board even though it is an old product stack
- AMD is the only one bringing a new tech with limited availability, but has no added performance to maintain a higher price point, effectively negating the advantage of having HBM (mind you this is precisely why Fury X is seen as a 'failure')
- Nvidia/AMD are unbalanced in the market, 75 vs 25% share, a gross imbalance in R&D budgets and products perform relative to this difference in budget
- There is no real competition for Nvidia

Competitiveness on the GPU market is fictional at best at the moment.

 

[Vayra86]

Get ready for some hi res 3Dfx porn...

I think the exact phrase was " overclockers dream"

As an overclocker, I'd say an overclock of 75-100MHz (max) under water constitutes less a dream than a stupor.

That is usually the way of things when a company is playing catch up, although AMD in recent years seems to have developed a marketing insecurity. I think they attempt to portray themselves as the little engine that could, but their marketing tends to come across as hesitancy wrapped up in bluster ( The Roy Taylor Syndrome). The company almost always use the competitions products for reference. Nvidia tend to do the same thing with Intel, less so with AMD where they are more confident...while Intel? Well, when was the last time any of their PR material/PPS/review kit mentioned AMD or Nvidia at all? I can't think of a single instance in at least the last half dozen years.
When a company has confidence it does not need to reference the competition, but I don't think AMD as a company has ever embraced the "less is more" angle of the brand, which is a shame because ATI before AMD's swallowing of them, tended to carry the aura of "walk softly and carry a big stick" - the kind of company projected attitude that builds consumer confidence in a brand.

Very well put. AMD is really the only instigator of its own demise and of the high expectations people had of this release. I remember very well how the GTX 970 'just landed' and how people were blown away by Maxwell's efficiency. What marketing? Just business as usual, with Nvidia saying 'look, here is Maxwell, enjoy' instead of 'OMG 2x PERF/WATT OVER EVERYONE ELSE'. With AMD"s most recent slides, they did just that, and that's fine if you do deliver, but if you don't, well. We also remember the 970 as the 3.5+0.5Gb card very well. And I damn well know for sure that that is going to stick much longer. Why will it stick? Because Nvidia responded saying 'THIS IS 4GB AND IT FUCKING WORKS, deal with it'. Arrogance.

Arrogance always gets rewarded negatively, and in the past months both Nvidia and AMD have seen this happen right in front of them..

 

[moproblems99]

And so I have sold my soul to the green devil. Ordered the ASUS Strix GTX 980 OC. Looks like a good overall package with slightly higher price. The Ti version was a bit much, I'll rather replace it one more time instead when Pascal and Arctic Islands come out. If even needed...

I too am in that position, though not as dramatic. I am looking at the 970 vs 290/X vs 380. The 970 is close to the performance I want but the 380 is more the price I want since I am going to be upgrading sometime next year....oh the decisions. I guess I am going to hold out and see what Nano brings...although I expect to be in the same boat I am at now.

 

[rtwjunkie]

go get yourself a techpowerup.blogspot.com


it's still under "news" though.
View attachment 66123

sorry dude that you have to hear this from me, but you have no idea, other than what the specs pdf says.
I'd like to hear an architecture break down from your "pen" and see there how much your knowledge upon the subject take effect.

'till that time comes, please stop posting your opinions, we can read the specification documents on our own and everybody can make similar assumptions.

Sorry dude. You are making this complaint on a privately owned website, that you are allowed to be a member of. Since you've not been around here for a great deal of time yet let me explain. It has always been so that occasionally the News head writes editorials, just as you will find editorials in a hard-print newspaper. If you don't like them, don't read them. It's as simple simple as that.

Until then, W1zzard, who owns the website, allows btarunr to handle the news page in the manner he has been for a long while now.

 

[RejZoR]

I don't get it why AMD needs whole month to release all their Fury cards. So annoying. I was placing hopes in the vanilla Fury and partially R9 Nano, but I just got tired of waiting and decided for the GTX 980... It's their loss.

I'm giving NVIDIA a try this time because they support higher level of DX12, run quieter on air and seem to overclock quite nicely. I'm not gonna pay almost GTX 980 money for a rebranded old card (390X). I'm mad, but I'm not that mad. GTX 980 just seems like a better long term investment for gaming really... And partially because I really wonder if I could use HW PhysX to boost Natural Selection 2 performance. Not sure how this one works with it since they don't use PhysX to make things more fancy afaik. Should mean they use it to just accelerate existing basic physics everyone have on CPU. Hopefully. We'll see.

 

[moproblems99]

I am not even sure the Nano is going to be launched in July. I don't think they gave a release date for Nano. They just said sometime this summer, which could be up until September 21. Hard to tell with AMD.

 

[64K]

And so I have sold my soul to the green devil. Ordered the ASUS Strix GTX 980 OC. Looks like a good overall package with slightly higher price. The Ti version was a bit much, I'll rather replace it one more time instead when Pascal and Arctic Islands come out. If even needed...

It's not so bad no longer having a soul. Babies start crying when you walk by and dogs bark at you a lot but hey, you get a shiny new 980 Strix. So it's all good.

 

[RejZoR]

I am not even sure the Nano is going to be launched in July. I don't think they gave a release date for Nano. They just said sometime this summer, which could be up until September 21. Hard to tell with AMD.

No, there was no release date for Nano, but I sure in hell won't wait till Autumn 2015...

 

[moproblems99]

No, there was no release date for Nano, but I sure in hell won't wait till Autumn 2015...

Me either, July is my cutoff. Tired of waiting and being disappointed. Then I'll decide, unless a good deal pops up somewhere. About $275 is the most I am willing to pay for a 290X (BEFORE any rebates). Anything more than that and I will pay for the power efficiency of the 970.

 

[RejZoR]

At first I wanted to spend only 400€ on a graphic card, but since I won't be going anywhere for the holidays this summer, I've decided to treat myself with a 600€ graphic card instead. Coz why not

Besides, I'm planning to tweak my HD7950 to some proper clocks, test it out thoroughly, bake the settings to BIOS and then sell it with a bit upmarket price since it'll probably be among the fastest HD7950 available. With clocks of 1200/7000, that seems quite a possibility... So in the end I'll be spending a lot less for the GTX 980...

 

[the54thvoid]

Don't be this guy:



I am a content head of this site. I set its news policy. I can decide to post hello-kitties on the front page, and it will be there. It will stay there. The "editorial" tag allows me to post, well, editorials. If you're good at reading specs sheets, then don't waste your time commenting on our, well, my content.

Lol, smackdown.

I too am amazed by the bizarre sense of ownership peeps have.
TPU is not a democracy! It's a techitarian autocracy.

 

[newtekie1]

One of the crashes completely hosed my OS on a stock system. That's hardly minor in my books. But that's one crash out of 100's I've had, and I do proper back-ups so that data loss wasn't that bad other than having to re-image the drive to get back up an running.

How the heck did that happen? Was the OS installing updates when the TDR happened and you didn't give it time to recover and turned the system off?

If it's constant, you at least know it'll happen. If it's not, it's even worse... Besides, a typical NVIDIA denial. Same as observed back in Windows Vista days when majority of BSOD's were caused by the NVIDIA driver and yet people were blaming Microsoft for it, you know, because it's BSOD, it must be a Microsoft thing. Except those BSOD's had NV*.sys mentioned on it...

In the early Vista days nVidia wasn't the only one with poor drivers, pretty much all the hardware manufacturers got caught with the pants down. Even Intel's drivers caused shitloads of BSODs back in the day, their storage drivers for Vista were pretty bad in the beginning.

Having crushes once an hour for something that you could be doing possibly for many hours a day it's not minor. Fire up Chrome, load TPU, write a big post at someone and have your browser crush just before posting and that's only one example.

It is always funny when someone is commenting on something, trying to make a big deal about a problem they've never even experienced. When the crash happens you don't loose anything. The browser doesn't crash, the computer doesn't need to be rebooted. You can't be right in the middle of writing a long post, chrome just freezes for a second, then the monitor goes black, and 10-15 seconds later it comes back on and your right back where you left off with a message saying the nVidia driver crashed and recovered. The crash is a driver crash, not a browser crash or a system crash. The driver recovers and you keep on typing.

I've been dealing with it on my main system for about a month since they put out the first driver with the problem, it has happened 100s of times, and not once has it failed to recover.