[WCCF] AMD Radeon R9 390X Pictured

Page 86 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
Status
Not open for further replies.

jamesgalb

Member
Sep 26, 2014
67
0
0
Here is a dead giveaway that R9 390X is R9 290X rebrand (as if we didnt have enough).

R9 390X board is marked as "LF R29FA"
Picture

They use this on PowerColour R9 290X cards
http://www.coolingconfigurator.com/step1_complist?gpu_gpus=1477

Anything verifying that new water cooled card as a 390x?

Very likely that Powercolors was looking to display their new cooler but was doing it on an old card, which is why they thought they could leak it and not violate the NDA...
 

EightySix Four

Diamond Member
Jul 17, 2004
5,122
52
91
Secondly i would asume gpu logic is more repeated so to speak so the benefits is less. It might be wrong whats your take on it?

It depends. Let's pretend an imaginary piece of logic and its connections are 1µm (this is all BS sizes btw) and there are 600 of them. The total size of that logic block is 600µm. Now we are able to reduce the size of that logic block to .9µm, the total size is now 540, allowing us to fit 667 of them in the same 600µm space that the 600 originally took up. That's an 11.2% increase in the imaginary blocks for the same size because we shrank each block by 10%.

It gets way more complicated when you start talking about routing and mixing different blocks in the same space, and hand optimizations... but you can see how decreasing the total size of a repeated element can drastically impact the amount of logic you can fit on a die.
 

chimaxi83

Diamond Member
May 18, 2003
5,457
63
101

MrTeal

Diamond Member
Dec 7, 2003
3,919
2,708
136
Interestingly enough, in that picture that's supposed to be the PowerColor 390X, the PCB has the 4GB box checked.
The original tweaktown link is a 404 now, where's any branding showing that card is actually a 390X?
 

krumme

Diamond Member
Oct 9, 2009
5,956
1,596
136
AMD is getting 33%+ on the CPU size so I took the 10% as a reasonable # to aim for in estimating the possible increase in shaders.

Kaveri has 2.41 billion
Carrizo has 3.1 billion

Overall CPU + GPU 28% increase in density.

I think 10-15% for GPU is a fair assumption.

You might be right. Looking at the first bd dieshots where the die is big but its like nothing is there- i asumed most of the benefit was from all the non core stuff - everything in between.
But anyway we get a s...load of transistors there so stellar performance is comming our way. Great.
 

at80eighty

Senior member
Jun 28, 2004
458
5
81
What exactly is some random picture supposed to say dude? Seriously? Just about all of your "leaks" and "info" is stuff like this, just randomness. Why?

its quite amusing to watch him go about doing his thing. just treat it as such.
 

Glo.

Diamond Member
Apr 25, 2015
5,930
4,991
136
What exactly is some random picture supposed to say dude? Seriously? Just about all of your "leaks" and "info" is stuff like this, just randomness. Why?

I suggest looking at notebookreview forum to see who leaked first information on big Nvidia Maxwell cards ;). And it was in last year ;)

But lets stay on topic ;).
 

Cloudfire777

Golden Member
Mar 24, 2013
1,787
95
91
I suggest looking at notebookreview forum to see who leaked first information on big Nvidia Maxwell cards ;). And it was in last year ;)

But lets stay on topic ;).
I have posted so many legit leaks to forums over the years, but people only try to focus on the negative ones. But I don`t mind. I`ve come to term that people will let the emotions get the best of them on forums like this and start getting all defensive.

I`m even in the market for both Fury X and 980Ti and havent decided yet what to buy lol. But it won`t stop me from reporting my finds, negative or positive.
 

maddie

Diamond Member
Jul 18, 2010
5,204
5,614
136
It depends. Let's pretend an imaginary piece of logic and its connections are 1µm (this is all BS sizes btw) and there are 600 of them. The total size of that logic block is 600µm. Now we are able to reduce the size of that logic block to .9µm, the total size is now 540, allowing us to fit 667 of them in the same 600µm space that the 600 originally took up. That's an 11.2% increase in the imaginary blocks for the same size because we shrank each block by 10%.

It gets way more complicated when you start talking about routing and mixing different blocks in the same space, and hand optimizations... but you can see how decreasing the total size of a repeated element can drastically impact the amount of logic you can fit on a die.

Now this is what I see as another big benefit of interposers.

You can access the main GPU die very close to where the signal is needed. The microbumps are ~ 380/mm2 and are all over the bottom. No great need to route multiple memory signals to the perimeter.

This alone will save some die space.
 

jackstar7

Lifer
Jun 26, 2009
11,679
1,944
126
I have posted so many legit leaks to forums over the years, but people only try to focus on the negative ones. But I don`t mind. I`ve come to term that people will let the emotions get the best of them on forums like this and start getting all defensive.

I`m even in the market for both Fury X and 980Ti and havent decided yet what to buy lol. But it won`t stop me from reporting my finds, negative or positive.

It's not about positive or negative, it's about what you actually know or don't know and what you pass off as facts, as I recall you being specifically warned against doing any further.


I posted the link to the TT pics that I saw at [H] before you, but I didn't say anything more than it was being claimed to be a 390x.

I understand you're not a journalist, but you're doing a disservice to the boards by just posting every thing you see and then passing rumors off as facts.
 

chimaxi83

Diamond Member
May 18, 2003
5,457
63
101
I have posted so many legit leaks to forums over the years, but people only try to focus on the negative ones. But I don`t mind. I`ve come to term that people will let the emotions get the best of them on forums like this and start getting all defensive.

I`m even in the market for both Fury X and 980Ti and havent decided yet what to buy lol. But it won`t stop me from reporting my finds, negative or positive.

I was asking a legitimate question. You post rumors, leaks, slides, and all that is fine, but almost every time, you post them as facts. Why?
 

Creig

Diamond Member
Oct 9, 1999
5,170
13
81

iiiankiii

Senior member
Apr 4, 2008
759
47
91
As .vodka stated, the 290X PCB might be compatible with the new 390X GPU. A PCB is not a GPU. You keep assuming things to fit your theories. And it's starting to get old.

Kinda like how the 980ti uses the exact same PCB as the titan x. Yet, they're different cards.
 

exar333

Diamond Member
Feb 7, 2004
8,518
8
91
As .vodka stated, the 290X PCB might be compatible with the new 390X GPU. A PCB is not a GPU. You keep assuming things to fit your theories. And it's starting to get old.

This is true...

I bought a 670 waterblock for my 970 because the PCB layout was the same. Not even the same generation. :)

Edit: These pictures don't tell a complete story.
 

ozzy702

Golden Member
Nov 1, 2011
1,151
530
136
It's not about positive or negative, it's about what you actually know or don't know and what you pass off as facts, as I recall you being specifically warned against doing any further.


I posted the link to the TT pics that I saw at [H] before you, but I didn't say anything more than it was being claimed to be a 390x.

I understand you're not a journalist, but you're doing a disservice to the boards by just posting every thing you see and then passing rumors off as facts.

I was asking a legitimate question. You post rumors, leaks, slides, and all that is fine, but almost every time, you post them as facts. Why?

I agree with the above.
 

Vesku

Diamond Member
Aug 25, 2005
3,743
28
86
Only actual confirmed facts we have are from the physical display of Fiji: quite large and 4 HBM stacks.

It would be quite a change for AMD if they priced something so it was actually more $ for less performance than an existing Nvidia SKU. So I am skeptical about them trying to sell a mildly OCed 290X with 8GB as the 390X at the same price as a GTX 980. Unless the intent is to capture sales arising from the sudden "4GB is not enough" internet PR blitz, which seems odd considering there is nothing indicating AMD is behind that especially given the apparent 4GB limitation of their first generation HBM.
 

maddie

Diamond Member
Jul 18, 2010
5,204
5,614
136
Only actual confirmed facts we have are from the physical display of Fiji: quite large and 4 HBM stacks.

It would be quite a change for AMD if they priced something so it was actually more $ for less performance than an existing Nvidia SKU. So I am skeptical about them trying to sell a mildly OCed 290X with 8GB as the 390X at the same price as a GTX 980. Unless the intent is to capture sales arising from the sudden "4GB is not enough" internet PR blitz, which seems odd considering there is nothing indicating AMD is behind that especially given the apparent 4GB limitation of their first generation HBM.

Isn't it more accurate to say is that:

Only actual confirmed facts we have are from the physical display of one Fiji model: quite large and 4 HBM stacks
 

tential

Diamond Member
May 13, 2008
7,348
642
121
Only actual confirmed facts we have are from the physical display of Fiji: quite large and 4 HBM stacks.

It would be quite a change for AMD if they priced something so it was actually more $ for less performance than an existing Nvidia SKU. So I am skeptical about them trying to sell a mildly OCed 290X with 8GB as the 390X at the same price as a GTX 980. Unless the intent is to capture sales arising from the sudden "4GB is not enough" internet PR blitz, which seems odd considering there is nothing indicating AMD is behind that especially given the apparent 4GB limitation of their first generation HBM.
It's such a change that people suggesting it at this point who are seasoned gpu users clearly have an agenda at this point.
 

exar333

Diamond Member
Feb 7, 2004
8,518
8
91
Only actual confirmed facts we have are from the physical display of Fiji: quite large and 4 HBM stacks.

It would be quite a change for AMD if they priced something so it was actually more $ for less performance than an existing Nvidia SKU. So I am skeptical about them trying to sell a mildly OCed 290X with 8GB as the 390X at the same price as a GTX 980. Unless the intent is to capture sales arising from the sudden "4GB is not enough" internet PR blitz, which seems odd considering there is nothing indicating AMD is behind that especially given the seeming 4GB limitation of their first generation HBM.

Well said.

I also think AMD just has too many products. Its confusing to many novice users and too many, too close in price. Keeping the $200+ product stack consistent, and uncluttered helps with marketing and word of mouth. The below doesn't even include all the R7 models either, which all pretty much line-up against just the 750/750Ti on the NV side.

AMD
270
270x
280
285
280x
290
290x
295x2

NV
960
970
980
980Ti
Titan X

Keep it simple! :)
 

tential

Diamond Member
May 13, 2008
7,348
642
121
Well said.

I also think AMD just has too many products. Its confusing to many novice users and too many, too close in price. Keeping the $200+ product stack consistent, and uncluttered helps with marketing and word of mouth. The below doesn't even include all the R7 models either, which all pretty much line-up against just the 750/750Ti on the NV side.

AMD
270
270x
280
285
280x
290
290x
295x2

NV
960
970
980
980Ti
Titan X

Keep it simple! :)
Nvidia has tons of products too. Are you seriously suggesting it's normal for nvidia to have a product lineup of
960
970
980
980ti
Titan x

? We all know they'd have refreshed the whole gpu stack but 0 competition to do so. Don't act like a 960ti 950 950ti and 940 aren't the norm....

People seriously are in these threads rewriting history right now.
 

geoxile

Senior member
Sep 23, 2014
327
25
91
Isn't it more accurate to say is that:

Only actual confirmed facts we have are from the physical display of one Fiji model: quite large and 4 HBM stacks

Huh? You mean one die. That's like saying there only one Hawaii model because both the Pro and XT share the same die.
 

maddie

Diamond Member
Jul 18, 2010
5,204
5,614
136
Huh? You mean one die. That's like saying there only one Hawaii model because both the Pro and XT share the same die.

definition
Die = GPU die


I mean one Fiji model because the Die will be the same, but the interposer size will be different if more than 4 stacks HBM.
Its speculation, I know, but we know that there can be more or less than 4 stacks of HBM memory. Confirmed by AMD.

I have trouble with accepting that AMD will make a flagship possibly as fast as 295X2, and only put 4GB HBM maximum. It makes no difference if they can optimize the 4GB so there is no appreciable performance shortfall. They must know what Nvidia marketing will do in that situation.

I also see AMD as having a huge window of opportunity in professional/scientific markets as Nvidia dropped DP performance in Maxwell. 4GB is a huge problem in those markets.

They must have higher than 4GB models.
 

jamesgalb

Member
Sep 26, 2014
67
0
0
Well said.

I also think AMD just has too many products. Its confusing to many novice users and too many, too close in price. Keeping the $200+ product stack consistent, and uncluttered helps with marketing and word of mouth. The below doesn't even include all the R7 models either, which all pretty much line-up against just the 750/750Ti on the NV side.

AMD
270
270x
280
285
280x
290
290x
295x2

NV
960
970
980
980Ti
Titan X

Keep it simple! :)

Like NVidia?

GT 705
GT 710
GT 720
GT 730
GT 740
GTX 750
GTX 750 Ti
GTX 760
GTX 760 Ti
GTX 770
GTX 780
GTX 780 Ti
GTX Titan
GTX Titan Black
GTX Titan Z
 
Status
Not open for further replies.