http://wccftech.com/nvidia-pascal-3dmark-11-entries-spotted/
>The first entry we have is of an unidentified Nvidia GeForce graphics card with 7,680 MB of graphics memory, 512MB short of 8GB, very likely a test chip. More interestingly is that this GPU’s VRAM is clocked at 2000Mhz – 8000Mhz effective. It’s a good time to point out that Nvidia has no GPUs on the market configured with 8GB of 8000Mhz GDDR5 memory. And certainly no 8GB GPUs with performance that’s nearly identical to the GTX 980 Ti.
>However as mentioned earlier, neither does the GTX 980M, or GTX 980 notebook match the performance figures you see above. Which are inline with the GTX 980 Ti. So we might very well be looking at the GTX 1070 ,GTX 1080 or the mobility version of either. Also worthy of note is the 545mhz clock speed recorded above. After we’ve done a bit of digging around we found that in most cases when 3DMark does not fully identify the graphics card in question it will report a ~540mhz clock speed, despite whatever the actual clock speed of the card is.
>Now let’s move on the next two entries. Again we have two cards with 8GB of 8000Mhz GDDR5 memory, showing performance that’s very similar this time around to factory overclocked GTX 970 graphics cards. So we could potentially be looking at early mobile variants of Pascal.
>The last interesting entry is of an unidentified Nvidia 3GB 2500mhz – 10,000Mhz effective, graphics card. So we could potentially be looking at the very first GDDR5X pascal graphics card here.
>Interestingly this card is also showing performance very similar to factory overclocked GTX 970 graphics cards. Keeping in mind that GDDR5X memory modules come in 512MB, 768MB, 1GB, 1.5GB and 2GB configurations this card could plausibly have a 128bit or 192bit bus for a total of 160GB/s or 240GB/s of bandwidth. So we might be looking at a ~ GTX 1060 class graphics card here
>>53559720
It's an indian rumor site. Nothing there is fact. Remember when they said the entire 300 series would have HBM?
>>53559720
>512MB short of 8GB
>nVidia is STILL making cards that lack .5GB of VRAM
Top fucking kek nvidiots
can't wait to get back on the winning side. Release this shit soon nvidia please
Fuck nvidia
>>53559720
>7.5
tell me it's a meme, a joke or both, please
http://www.3dmark.com/3dm11/10969217
http://www.3dmark.com/3dm11/10985139
http://www.3dmark.com/3dm11/11061015
>512MB short of 8GB
Oh for fucks sake Nvidia, the joke wasn't funny the first time around.
I wonder if they're going to be honest about it on the box and market it as a 7.5GB card.
>>53559720
>7.5
top cuck
>>53559720
>512MB short of 8GB
7.5/8, laughed a good amount
amd poorfags won't make the .5 joke this time because Polaris will only have 4GB.
>>53559720
>7.5
REV UP THOSE MEME BOYS
the real meme should be expecting a whole number, i don't know when this shit in computing started but for the first few decades it didn't happen
>>53560414
Nobody cares if it's 4096MB of memory or 4GB, rounded numbers are just easier, you're not writing for a fucking compiler that you need to explicitly mention 4 096 000 000 bytes
![hqdefault[3].jpg](https://i.imgur.com/D4Xd20v.jpg "hqdefault[3].jpg")
>>53559720
>The first entry we have is of an unidentified Nvidia GeForce graphics card with 7,680 MB of graphics memory, 512MB short of 8GB
Meme magic is a powerful thing
>>53560437
Anon...
Pascal has dp 1.3 right?
So its freesync capable, but nvidia wont allow it?
>Also worthy of note is the 545mhz clock speed recorded above. After we’ve done a bit of digging around we found that in most cases when 3DMark does not fully identify the graphics card in question it will report a ~540mhz clock speed, despite whatever the actual clock speed of the card is.
>So in this case and indeed in all the following entries you will see 540mhz erroneously reported as the GPU frequency because 3DMark failed to read the actual clock speed of the card, possibly as the result of a BIOS misread. This can be confirmed by looking at some other misidentified GTX 970 and GTX 980 entries with a recorded clock speed of 540mhz. Which also happen to show performance figures that are on par with what’s normally expected of these cards at their boost frequencies. Really no where near what they would score at ~540mhz
Not fake, I think it's the prototypes
>>53559720
replacing GTXmeme with GTX1080 as soon as it launches
>Performance that’s very similar this time around to factory overclocked GTX 970 graphics cards. So we could potentially be looking at early mobile variants of Pascal.
I have a feeling that this is the mobile benchmark.
Last time the something scored same as the GTX 780 and it turned out to be the mobile card.
>7.5
Literally meme card
>>53560448
cant wait for the memevideo
>>53560448
>8GB
KEKED AGAIN
>>53560525
yep
amd cards would be G-sync compatible... if nvidia allowed it.
>>53561129
yeah because I want a $600 Acer monitor....
>>53561129
AMD allows nvidia to be freesync-capable. Intel already supports freesync. Nvidia is hitler.
I can't wait to spend $2000 on 3 of their new highest end GPUs and waterblocks.
>>53559720
>7,680 MB of graphics memory, 512MB short of 8GB
KEK, this has to be a troll
>>53560384
>7.5GB
>>53561160
But he did nothing wrong :^)
>novidia has a memory goblin that keeps stealing 0.5GB from their cards
DAMN NVIDIA BACK AT IT AGAIN WITH THE POINT FIVE RAM
>>53559720
>7.5gb
Nvidia does it again.
>7.5
>So we were working on a new line of GPU
>And nvidia calls and says :
>We don't need a new fiasco like the 3.5!
>So we tried
>we thought to up the amount of vram
>we tried GDDR5, HBM, everything !
>and what did we do ?
>we made a .5 gpu again!
>7.5 GB!
>And we even advertised it as a 8GB card !
>7.5 !
https://www.youtube.com/watch?v=BdwUsalwBJ8
Okay someone really needs to do it.
>7.5
AHAHAHAHAHHAHAHAHAHAHAHAHHAAAHHA
I really hope Nvidia keeps fucking up so AMD will hopefully be able to regain some market share
Daily reminder that you have a moral obligation to buy AMD since buy from Intel/Nvidia means you support a cancerous corrupt monopoly
>>53562387
I support this entirely.
The only Intel hardware I have is the CPU in my 5 year old ThinkPad.
The only nVidia hardware I have is a second hand 7100 GS that I got for $5 as a primary boot card for VGApassthrough because it was fanless.
>>53559842
>missing 5gb
literally what
>>53562387
It's going to take a much bigger fuck up than something like 3.5 to change anything. Nvidia brand loyalty is so strong that even if AMD wiped the floor on them they'll still buy Nvidia. It really is the Apple of GPUs where "it just werks" wins out over anything and they'll pay more for the convenience.
>>53562557
>i can't read
>>53562557
This is why high resolution monitors are better.
>7.5GB
it's like they're not even trying
>>53562387
I'll recommend amd but I'm going to buy Nvidia for myself.
>>53562310
I'll do it later using webm for bakas and post it on /gif/, link it here.
PS - later is a 3 or 4 hours
>>53562387
I have a GTX 970, but i'm buying a Polaris because i believe in Raja.
Raja engineers GPUs for greatness, not profit.
>>53559720
NVIDIA DOES IT AGAIN
>>53562662
He'll put the extra oomph in poolaris
>>53562662
I have a 970 too and I am leaning heavily towards Polaris, especially if Nvidia jews out and only offers HBM on the Titan equivalent
Fuck GDDR5X is a compromise of old technology in every respect. The future is HBM and I feel AMD is going to have more offerings using it since Fiji already has it
I won't buy another video card that doesn't have HBM after my 970
>>53562729
>3.5gb to 4gb
lol'ed irl
>>53562387
this card is just enabling HT on a pentium to turn it into an i3. they're basically the same exact chip just with less features.
>different products have different prices
people who would buy something like this are the problem. fucking gateway,
>>53562762
What the fuck is with you people and posting Unity benchmarks trying to prove some sort of point?
It's a broken, shitty game that runs like garbage on everything and should be considered invalid
And who's to say Polaris will cap out at 4gb?
>>53562768
There's more to an i3 that just HT.
>>53562817
Okay, I'll keep posting 4k benchmarks until you find a game you like.
:^)
>And who's to say Polaris will cap out at 4gb?
AMD.
>>53562937
>Polaris will cap out at 4gb
F I N I S H E D
I
N
I
S
H
E
D
Any hope for an updated version of the 750 ti?
Love the low power draw.
>>53562945
With DX12 multi-gpu support, high end games should be able to utilize two GPU's with 4GB each as a full 8GB of memory.
It will be interesting to see how this pans out, but it would be disappointing to see AMD cap out at 4GB, although I wouldn't be in the least bit surprised if they manage to get up to 8GB on a single GPU.
>>53562984
Yeah it's called the 950
>>53562729
>The future is HBM
meh, I will let benchmarks speak for themselves.
you know, when they are eventually happen.
AMD is on life support, they really need a good home run with Zen and whatever else is going on, otherwise it is all over.
>>53562937
>4K
1440p and high refresh rates are far more interesting to me than moar pixels.
3440x1440 ultrawides with 100hz refresh rates are the shit too
>>53563017
Intel and Nvidia will just price their shit higher. Everybody wins.
>>53563017
>AMD is on life support
This is true. They're betting the house on Zen and Polaris, and I really hope they can pull it off
Pic related
>>53563023
AMD disagrees.
>>53563002
I'd love to see a 950Ti made from GM206-300 dies (from the 960) binned specifically for power consumption.
>>53563060
AMD caught lying again.
>"smooth gameplay"
>150ms when the ideal is 19ms
>>53559720
>wccftech
>>53563041
200*14 is not 3800
>>53563060
I support AMD, but FPS count is misleading.
What you need is maximum (or 1% worst average) frame delivery TIME, or the largest time taken to draw a single frame.
>>53563060
4K is just easier to market because it's a buzzword
Ultrawides and and other less well known resolutions are still sort of a niche
>>53559720
>they will release another 0.5 GB card
maxwell rebrands confirmed
>>53561642
did 4chan just discover H3H3 and filthy frank?
>>53562937
Looks like the fury x gets fucking creamed
>>53562570
My loyalty goes to the company that has the highest performance at the price i want when i buy, nothing more.
>>53562991
who will get a CF...
There is literally nothing wrong with 7.5GB.
The GTX 970 was only terrible because the additional 512MB on top of the 3.5GB was slower, which actually makes it worse than simply being 3.5GB, and constitutes false advertising.
>7,680 MB of graphics memory, 512MB short of 8GB
There's no way. It couldn't happen again, could it?
>>53562310
>970
>the most popular gpu ever made
>fiasco
>>53563727
This.
The 970 performed just fine as long as you weren't retarded with your graphical settings.
The 390 with 8gb of vram is stupid because it will almost never use that much and if it is then the actual core is already too slow to keep up.
Even if the 970 had the full 4gb it would never make a real difference in performance.
>>53563735
>popular=not fiasco
Yeah, that's definitively why Intel graphics are the greatest ever
>>53563728
Read thread nigga
there's nothing wrong with it having that specific number of memory provided it's a unified, non-segmented architecture with none of the L2 cache disabled.
Which, at the moment, there's no reason to believe.
The GTX570 had 1280MB of memory, there's no specific reason either NVIDIA or AMD has to stick to an even VRAM number.
>>53563845
Do they not pay you enought to actually read the posts you are replying to, Rajesh?
>>53563728
The only reason the 970 memory is a meme is because they didn't disclose it at launch for whatever reason
If they explained it correctly people wouldn't have cared
The 660ti had a similar memory configuration
>>53563972
They didn't disclose it because then reviewers would have tested it and found the stutter.
The card would have been declared problematic on day1 and only sold half as much.
>>53563186
No, but the userbase on 4chan is welcoming a new generation of youngsters.
>>53564009
Just like how the 660ti was such a failure?
>>53564054
I think the issue was the fact that the 980 seemed extortionate for what was a slight upgrade from the 970, so when people found this Vram issue, it was a bit of a piss take.
>>53561129
G-sync is better anyway.
>>53562570
What's wrong with being equated to a company which consistently makes better products?
>>53563076
Yes but that AMD chart is bigger than yours, I'm going to have to go with AMD.
>>53559720
Nvidia users said a game does not need more than 3.5GB why are they pushing for 7.5?
What ever happened to that class action lawsuit for false advertising 4gb of vram with the 970?
>>53562387
I have intel cpu and a amd gpu. Fuck nvidia
>>53559720
>7.5 gbs
at least they're being honest this time
>on 980ti
>going to hold out this weak die shrink introduction and buy the generation after where they start cranking it up to its full potential
>>53562729
Polaris 10 and 11 have already been confirmed to only use GDDR5, and not even the X version...
>>53564616
I mean, the XX80 will probably be like 4% faster than your 980ti, and the XX80ti will probably be like 10% faster. If you're gaming at 1440p, you will be happy with that card for a long damn time.
>>53559720
>7.5 GB
7.5 GB
>7.5 GB
>>53564267
Because it's a more powerful than the 970 card
>>53564722
>confirmed
Where? AMD was positioned to get first pick of HBM
>>53564784
Did you not watch AMD's Capsaicin livestream?
>>53564722
We don't know if it's GDDR5 or 4GB of HBM1
>>53564807
The number codes on the shipping manifests that best matches Polaris 10 and 11 indicates that the parts are regular GPUs without an interposer, which means HBM is off the table for them.
And there are only TWO GPUs coming out this year for AMD.
>>53564797
They nerver talk about polaris, it's was about VR and the Fury pro duo.
>>53564837
>They nerver talk about polaris
Were we watching the same stream then?
>>53564756
Im going to be VR gaming so I have to own the best possible card for when they start introducing high fidelity games and upgrading hardware to 4k per eye.
But that should align with the generation after Pascal/Polaris.
>>53564855
I was here and they never talk about GDDR5, they only said HBM2 was for 2017
>>53492345
>>53564918
I didn't say they talked about GDDR5 in the event, I said that the letter codes for interposer-mounted GPUs from AMD in the most recent shipping manifests weren't present. When Fiji was being tested, those cards had a unique letter code to indicate that they were on interposers, which haven't been seen since.
Conference comes up April 6th so we will know what is legit.
Would be interesting if this leak turned out not to be a fake.
http://techfrag.com/2016/03/18/nvidia-pascal-geforce-x80-x80ti-gp104-gpu-supports-only-gddr5-memory/
>>53559720
>7.5
>>53559842
The 970 has 4gb VRAM, the .5 has just been locked so you can't use it because it actually runs faster on the 3.5 (this has been provin in tests where people unlock the extra .5gb. It reaches it's max processing power around 3.4gb so anything above that isn't really even needed.
>>53560367
Polaris10 already confirmed for 8GB m8.
>>53565067
Of GDDR5?
>>53565039
Wrong, but you're close.
It has a full 4GB but due to various wizardry within how it's accessed, it shits the bed after addressing more than 3.5~.
The whole 4GB is 'unlocked' from the box, but you're misleading others.
>>53565067
[citation]
>>53565085
No confirmation of memory type yet, which means it's likely GDDR5 because if it was HBM they'd be shouting that shit from the rooftops.
>>53565111
It can't be HMB1 because that caps out at 4GB, and it also can't be HBM2 because that hasn't even started mass production yet, which means that it won't be ready in time to be installed on graphics card before Q4 2016. Polaris will launch in Q3 2016.
>>53565111
I need the citation of the >8GB
>>53560448
The Universe is conscious through humans and other sapient creatures. We are all literally the Universe, thinking, speaking, learning.
Quantum mechanics is tied to conciousness, and the Universe operates through quantum mechanics. All our thoughts are the thoughts of the Universe. When a lot of us actually think of something, and our conciousness is tied to an idea, the Universe is thinking about it too, and the more of us think together, the quantum mechanics that is operating the Universe will also have that idea and it's events in it's pool of all possible random events, and the chances of it actually happening greatly increases, and becomes much larger than all the other random events that can happen. When we all think of something together, at the same time, and constantly, we are actually affecting the quantum mechanics of the Universe, and the chances that the Universe bends to what we thought of increases and the events we are thinking about actually happening increases, because the Universe is thinking about it, because we are thinking about it.
Meme magic is real, and it is more powerful than any of us can ever imagine.
>>53562557
>literally can't read
>>53559720
>7.5GB
I believe it.
>>53564040
>1980 was 60 years ago
>>53562768
L3 cache too idiot
>>53562595
How? Making a 4 pixel big dot to a 1 pixel big dot? Fail.
>>53566945
kek or vise versa with same dpi monitor
>nvidia pre release benchmarks
>from currytech
You literally can't get more unreliable
>Nvidia does it again
Does what again? Theres no reason to have a even number of memory if the chip design does not benefit from it. Not a fanboy, using Radeon's are my primary cards right now, vendor loyaliti is Jack shit.
>>53567097
having cards with N+0.5 GB VRAM is fine.
What is not fine is advertising these cards as having N+1 GB of VRAM
>>53567029
http://www.3dmark.com/3dm11/10969217
http://www.3dmark.com/3dm11/10985139
http://www.3dmark.com/3dm11/11061015
what is that?
>>53562662
in order to engineer that gpu you're going to have to send him a prepaid visa for $4000 and give him access to your pc and social security number. this is absolutely vital, indians in tech what a laugh.
Why the fuck are they using windows 32 bit?
>>53565187
I like it, stalker
>>53559720
>>/v/
>>53559720
>7.5gb of GDDR5
>>53559720
>7,680 MB of graphics memory, 512MB short of 8GB
AMD WON
>>53559720
>The first entry we have is of an unidentified Nvidia GeForce graphics card with 7,680 MB of graphics memory, 512MB short of 8GB
NVIKEKS ON THE SUICIDE WATCH
>7.5gb of GDDR5
It's over
AMD WON
>could plausibly have a 128bit or 192bit bus for a total of 160GB/s or 240GB/s of bandwidth
the 390 has stock 384GB/s bandwidth, why is nvidia always so cheap.
>No HBM2
>GTX 980TI will still be a good as fuck card
waitfags BTFO
Am I missing something here? All the benchmarks are worse than the Oculus' recommended specs (~gtx 970). How can the 1080 be worse than the 970?
>>53570284
Nvidia finds a way.
>>53570284
prob the CPU?
>>53570325
The Way it's Meant To Be Played.™
>>53563671
was planning on it but all my plans are now on hold till more news about vega 10 comes out. it's going to suck watching everyone upgrade to pascal and polaris but it'll be worth the wait in the end for HBM2 cards.
>>53559720
>7.5
>GDDR
>not HBM
>literally screwing you out of half a gig AGAIN
they can't keep getting away with it!
the absolute madmen!
>Nvidia cucks everyone because AMD can't even compete with them
YOU COULD HAVE STOPPED THIS
>>53570492
HBM1 is shit
>7.5 GB
hahahahaahha
>>53562729
AMD will only bring HBM to higher end cards too familia
>>53562762
goddamn. amd still hasn't fixed frame jitter.
>>53559720
>512MB short of 8GB
>>53562570
>Nvidia brand loyalty is so strong
I love how you idiots say shit like this while unironically spouting garbage such as "you have a moral obligation to buy AMD!!!!"
This absolute lack of any self-awareness whatsoever is simply outstanding.
>muh 0.5
Fuck off guys
![[areyouevenbritish].jpg](https://i.imgur.com/rphx2J1.jpg "[areyouevenbritish].jpg")
7.5
>7.5
7.5
>7.5
7.5
NICE GPU NVIDIKEKS
>>53570362
lol
>>53562991
>With DX12 multi-gpu support, high end games should be able to utilize two GPU's with 4GB each as a full 8GB of memory.
that's barely a rumour...
>>53563186
I honestly did just like 3 weeks ago
>>53567234
But no one ever did that anon. 970 has 4GB, it just sucks at accessing the last 0.5GB. Anyway, the whole thing is a bit stupid given people were able to test the ACTUAL performance of the card all along, and performance numbers never changed.
>>53559720
Bump for
>7.5gb
>GTX 1080
Does this mean nVidia will finally be able to play at 1080p?
>>53559720
>card with 7,680 MB of graphics memory, 512MB short of 8GB
Oh man I laughed so hard I woke up my neighbors.
Does any videogame even make use of 8gb vram?
I mean we saw some games having problems with the 3.5 of the 970 (Shadow of mordor and one AC if I remember correctly), but even those were fucking minimal.
>>53578683
Most vram intensive games top out at around 6gb at 4K, I can't think of any that use a full 8gb currently, but I'm sure there will be some this year
>>53559720
>8GB vram
AMD DID IT FIRST
As long as there are no single GPUs that can drive 4k ultra at 60 fps nothing over 6gb matters
>>53578890
Hopefully a die shrink can put us over that bump finally
>>53559720
>7.5 GB
>7.5GB
Eh, it will still outperform any NVidia 8gb card
>7.5GB
>>53578822
They'll be a ton when NVidia pays them to upscale their textures until they use all 8gb, even if it provides no visual benefit.
>7.5GB
>>53578683
There are plenty new games that can easily get to 4GB. Rainbow Six Siege actually gets to ~3.8GB with ultra settings @1920x1080. Depending on how hard you mod it, TES:Skyrim can get pretty damn close to 8GB.
>wccftech
I bet they've been loving all the traffic coming from /g/ over the last week.
It's all made up, but you idiots keep lapping it up.
>>53559720
inb4 1024MB meme
