>>21701088
Nope.avi

If this is true though, I will be in euphoria when I get my hands on such a card

>obligatory troll pic

I doubt performance is going to be as good as they predict.

This shit is so fake it's not even funny.

>Metro: Last Night beta
>Benchmarks at 1560 x 1600

Fuck.

sauce in case anyone is interested:

http://www.tweaktown.com/news/21936/leakedtt_nvidia_to_skip_600_series_jump_straight_to_geforce_gtx_
780_did_i_mention_it_is_nearly_twice_as_fast_as_the_gtx_580/index.html

>>21701088
Seems very faked actually.

Why would they skip an generation? I doubt they would do that.

Anyway, I hope it's going to be awesome and not be released a year after AMD.

>not properly labeling the axes

>>21701152

>Why would they skip an generation? I doubt they would do that.

They also skipped the 300 numbering when they jumped from the GTX 200 to the 400. No reason why they can't do it again.

Also, it makes better sense to keep the numbering similar to AMD's next generation of cards in the 7000 series. Less consumer confusion for the completely retarded.

>>21701177
That's because they got delayed big time and had fail yields. Also, they never skipped the 300-series, just the desktop one.

>>21701152

Are you new to computers? Nvidia went straight from 2xx to 4xx, and 3xx were rebrands of the previous generation. 5xx was also a filler gen and 7xx will be the real thing.

>>21701195
There was a very good reason why they skipped 3xx generation. They don't for 6xx.

Nvidia can pull retard stunts all they want, they still failed on this generation. I mean the 500 series are fine and really good compared to the HD6xx0 but they just got late.

>>21701209
6xx is already out and is intended for the mobile market and low end desktop htpc cards.

7xxx we'll be jumping from around 500 cuda cores to a 1000, and a smaller dye at that.

The performance jump will be huge.

>>21701233
>6xx is already out and is intended for the mobile market and low end desktop htpc cards.
Only rebranded old models, same as AMDs 6xxx series.

>>21701233

>1000 is twice as much as 500

There's your doubled performance folks. Show's over. Time to go home.

>>21701190
>>21701209
Model numbers are part of marketing. They wont' want to release a '6####' card while amd is releasing a '7####' card. Same reason why MS didn't call the 360 the xbox 2 when they were competing with the ps3.

>>21701257
And thats what I said.

Pic related, it's what Nvidia will have to work on after 79xx rapes whatever they had planned, like the 3xx series they mysteriously dropped, then postponed fermi.

>$1500-$2000 GPU's

No thanks, besides all games are going to be running at 1080p for the next 6 years because of next gen consoles. And face it, PC just gets shitty Xbox ports now a days.

>>21701282
Still, the higher segments in the mobile market will still be 6xx

>>21701281
So why didn't they do the same for the 5xx? or why jump from 2-4 instead of to 5?

>>21701289
>ati
>ever beating the competition

lol, enjoy your poorfag amd/ati

>>21701291
This. I have a 2-3 year old machine still capable of running Skyrim on ultra. There's no point upgrading until the next generation of consoles comes out.

>>21701289
Weren't you AMDfags saying the same thing about Intel before Bulldozer came out?

>>21701302
Even then, next gen consoles will run at 1080p at 45-50 FPS. While devs are just going to do half-assed ports like skyrim.

>>21701299
They do have the best performing video card.

And best performance/watt
And best performance/price

Looks legit.

>>21701291
>>21701302
What the fuck are you guys talking about? the flagships from these series aren't going to magically break away from the current paradigm and go to 2K or anything, you'll see just what we've seen the past 4 years with prices, and AMD has confirmed this by setting the 7970 at roughly $400.

http://www.overclock.net/t/1185339/wccf-leaked-nvidia-next-generation-performance-slide-pits-upcomin
g-flagship-gk-100-kepler-based-geforce-gtx-780-against-gtx-580

Speculation thread.

>>21701297
they didnt jump from the 200 series to the 400 series. the 300 series when straight to the mobile market, and was basically rebrands of the 200 series. the 400 series was the first fermi, and was relatively stunted at that. the 500s were what fermi was supposed to be. Now, with a new architecture, the 600 series will probably be 500 rebrands aimed at the mobile market, and the 700 will be the desktop market. Its all in the marketing.

Really tho, current 7000 series cards from AMD are just rebrands of the 6000 series(because of low yields), and the 600 and 700 Nvidia cards that will be out first will be A)40nm, and B) more than likley rebrands, because of again, low chip yields.

>>21701088
>'Leak' specs of new cards
>stock prices go up
That's about it.

>Metro: Last Night

Fake as fuck

Here we go

>>21701337
its not a true 7970 right now, its a rebranded 6970, and they are charging more for it.

>>21701363
Great, finally I can grill the pork chops as they should.

The benefits of releasing later than the competition allows Nvidia to drop housefire generators that barely beat AMD.

>>21701363
Clever.

And no, there's no way it's true. Remember those benches saying the 480 was going to RAPE the 5970?

>>21701372

Durr.
This is also completely legit!

Don't believe me? It has the black and red theme AMD slides have! FAGGOT

When are Nvidia's new cards supposed to be released? I know AMD's cards are supposed to launch on January 9th.

>>21701354

Can somebody please explain me "low chip yields"

>>21701388

7970 is actually being released on the 22nd of this month.

>>21701369
The cayman arch is getting rebranded as a 7850 IIRC, 7970 is 28nm GCN juicy goodness, or will be when it's released.

http://www.techpowerup.com/156876/AMD-Pulls-Radeon-HD-7970-Launch-to-December-22.html

Unlesss you have a source to refute this.

>>21701388
TBA 2012.

Meaning - We'll see what AMD does, then barely beat them in some benchmarks.

>>21701382
Do share.

>>21701388
High-end is coming in the 2nd half of the year, so NVIDIA is going to be late as all fuck, AGAIN.

>>21701388
April or May for GK104
Q3 or late Q2(probably July or June) for GK100 flagship

>>21701395
Paper launch != real launch.

I'm thinking of building a new rig, should I wait for the 7xxx series to pair it up with a 2500k? Are they going to be expensive?

>>21701394
Meaning how many "good" chips they get. When you're producing things at 40nm, you get lots of bad chips with defects.

>>21701394

Basically, from what I understand, the amount of success when producing hardware with new chips incorporating smaller transistors etc can be tricky and problematic, so that is why they are have a low yield. Only a handful are working properly, or even at all.

>>21701394
When a wafer is made for a GPU, it has upwards of 30 actual GPU chips on it, that are then broken apart and tested. Low chip yields are when not many of those 30 chips are functional, and have to be scrapped. AMD press release said something like 5% yields from TSMC, and Nvidia was around the same.

It's like I'm really looking at the back of a nvidia box!

Oh golly gee, I sure hope this one doesn't randomly combust on me though. That'd sure ruin my nvidia experience.

>>21701430
Housefire Generation is one of the selling points.

>>21701416
>When you're producing things at 40nm, you get lots of bad chips with defects.

you mean 28nm right? the 40nm process is so perfected now, that its something like 90% yields from it. the 32nm process is about 50% yields(which is just barely viable for having enough supply), and the 28nm yields are just terribad right now.

>>21701369
7970 and 7950 use a completely new architecture. 7970 is supposedly coming out in 4 days but it won't be widely available until January 9th. That's when the 7950 is coming out as well.

>>21701428

Fuck. I am fucking STARVING for some wafers now.

>>21701416
>>21701418
>>21701428


Thanks guys.

>>21701452
I meant both, they are both fuck all small.

You got sources on those yield number? I'm genuinely curious.

>>21701460
>7970 and 7950 use a completely new architecture. 7970 is supposedly coming out in 4 days but it won't be widely available until January 9th. That's when the 7950 is coming out as well.

as AMD said in a press release, they cant use the true 28nm tech yet, since the chip yields are way too low, and thus, are just rebranding the 6000 series until true 28nm stuff is useable. Cant find the actual AMD press release, I had it bookmarked, but AMD pulled it from their site. It stated that not only the low end, but also the flagship will be rebrands into february or march, since the chip yields are still around 10% on the 28nm process. It was an investor briefing.

>>21701452

Both AMD and nvidia has already said the yields are surprisingly low on 28nm, got anything to back up all the shit you said?

>>21701452
http://www.eetimes.com/electronics-news/4230286/Foundries-have-28-nm-yield-issues--say-execs-
1.7% yield wasn't even a problem at one time. I wonder what the yield is on 28nm dies.

>>21701523
Biggest load of FUD I've ever seen.

>>21701523
Are you sure you aren't confusing the lower end chips with the higher end?
Here have a read:
http://semiaccurate.com/2011/12/16/low-end-radeon-hd-7000-series-rebrands/

I cannot find ANYTHING that claims the 79xx series is not 28nm, please do share if you can.

>>21701483
see
>>21701523

I had an AMD investor briefing/press release bookmarked, and now its pulled from their site. More than likely, it was a mistake that it went up, but it gave yield numbers of 10% or less for the 28NM GPU wafers, and said that they will most likely be rebranding the 6000 series GPUs well into the first quarter of 2012. They already confirmed that the 7000m series is just a rebrand of the 6000m series, and the low end desktop GPUs will be rebrands of the 6000 series cards. Now, because of TSMC, they have to rebrand some higher end cards. They cant delay the 7000 series launch any more, since investors are already pissed, and early adaptors will get a shitty rebranded card.

OH FUCK YES, NOW INSTEAD OF GETTING 1FPS AT 2560x1600 I WILL GET 2 FPS.

>>21701548
no, it was in an investor briefing.

>>21701572
Since you can't even back that shit up with screenshots, it's nothing but FUD.

>>21701572
Alright, and like I shared in
>>21701548
we already know the lower/mid ends are just rebrands
but there is NOTHING on the 79xx not being 28nm GCN.

Discussing mid end GPUs is fine...but difficult as neither side has released benchmarks on them.

>2012
>AMD releases a new batmobile.
>The US Army buys all GeForces to use in the invasion of Iran.
Batmobile it is.

where's in elite when you need him ;_;

>>21701630
He was cremated.

WHEN IS IT COMING OUT

I'M DYING OVER HERE

>>21701630
We all told that troll to fuck off long ago.

>>21701630
I'm sure if I were on the other side of the fence I'd share your hopes, but as it is that guy is a douche, troll or not. I may be saying this because I like AMD and he spends 24/7 bashing them but still...asshat.

>>21701647
Should have gotten a 6850 or 6950 instead of that CPU you'll never use, fuckface.

>>21701612
im trying to find a mirror or cache of the investor briefing, but its not looking good

>>21701592
why the fuck would I take screenshots when Im reading something? Like i said, its been pulled from the AMD site, more than likely was not supposed to be up there in the first place.

>>21701326
>best performing card
Single 580 beats a 6970.
590 beats 6990, especially in the acoustics department.
MarsII is the fastest single card ever.

wut.jpg

Remember when Nvidia said the GTX300 series would slap everyone's shit?

Remember when AMD launched the HD5000 series which was 30-40% faster than Nvidia's cards at the time?

Remember when Nvidia said the GTX300 would outperform it?

Remember when the GTX300 didn't launch for months, and was eventually renamed to GTX400, launched many months late, and didn't slap AMD's shit at all?

Remember when AMD is launching their 28nm cards in two weeks and Nvidia's 28nm GPUs aren't coming for another 4-6 months?

Yeah, I'm just saying.

>>21701647
>81c
>70C
please tell me those aren't idle temps

>>21701672
That picture is wildly inaccurate. The AMD car has a driver.

>>21701647

DOSE TEMPS

>>21701681
but the driver is crashing

GTX600 series will be mobile and oem gpus

>Out in 2015

>>21701686

I see what you did there.

>>21701674
They arent, they're full load temps, well full load was around 84-85C, but they went down after I minimized to screenshot.

BUT

Those were the idle temps when I had my fans running on stock, I used rivatuner to 100% 24/7 so they would run at not insane temps.

>>21701682
It's like my single slot 4850's long lost older brother!

>Launch pulled to before Xmas from 22nd of January
>Nvidia shills all over /g/ next day

http://vr-zone.com/articles/amd-radeon-hd-7770-pictures-leaked/14278.html

>>21701686
The car is crashing.

Yeah, those Unigine demos? Well, on the 5000 series, they run at 1. Whereas on the 7000 series, they run at TWO. That means they run ONE better.

Day one buy for me.

>GTX 780 fermi

>>21701716
The driver caused it

>>21701710
>150mm^2

Price/performance and Performance/watt card right here.

>>21701719
Yeah man.

>>21701647
>>21701674
>>21701682
Uh, guys? You do know that temps like those are normal for video cards, right? 70c to 90c has been the standard for a while now. It's perfectly normal. At idle, newer cards will downclock, and the fan speed will decrease, but it will still be pretty hot.

>>21701721
Made my day.

>>21701752
my 8800gtx has idled at like 65c 24/7 for about ~5 years or so

>>21701752
My card never goes above 60s.

Wtf is wrong with everyone's temps? My radeon hd 4870 is 41-45 C on idle. maybe 65 C max on load.

>>21701088
This will be all fine if the card doesn't use three slots and 700W.

>>21701752
NIgger, don't be silly.
Newer cards run even lower on idle.

Yeah too bad AMD can never go out of business thanks to the fact that monopolies aren't allowed.

>>21701768
My fermi idles at 29c and loads at 70c

>That feel when you will soon have to let go of the 5770 from your old budget build when you upgrade.

I am actually pretty sad about this.

>>21701781
My 6870 idles at about 32c and get to 60 at load. Fucking nvidia house fires man.

>>21701752
>70c to 90c has been the standard for a while now. It's perfectly normal. At idle, newer cards will downclock, and the fan speed will decrease, but it will still be pretty hot.

How about no. Unless this is something Nvidia does to make their cards double as a space heater during winter.

>>21701781
Care to mention cooling system.
>>21701788
Is that 4890?
>>21701788

>>21701742
>150mm^2
>Price/performance and Performance/watt card right here.
I'm glad someone noticed. That sucker is tiny.

Any assuptions on when the 78xx will be coming out? And how much could 79xx models be?

>>21701832
>Is that 4890?
Yes.

>>21701814
Just put it back in the box. Or sell it to someone.

700 series will be a huge flop like Tesla 3

>>21701816
the 8800gtx (or ultra) was outperforming ati's best card like 2:1 at the time or something. nothing else really came close so no one gave a fuck about the heat.

>>21701864
>Maxwell
Thats my cats name.

>>21701832

4870 owner here. I use standard cooling. My case air flow is kinda shitty too. I have an ASUS 1 GB brand card though. Maybe it has something to do with their HSF.

Pic related.

I'm stuck with an 8800GT that idles at 61 (and it's actually pretty chilly in here) and is ~82 C at load. Meanwhile, my E6600 is sitting at 33 idle and maybe 46 at load.

>>21701873
Cool story bro.

>>21701889
>dem low profile chokes

>>21701836
I'm lowing AMD's dedication to small dies for GPU compared to Nvidia is really something.

I was hoping they'd make GCN smaller than Cayman but I guess it's really difficult to compete with the flagship die that's twice smaller than the competition.

>>21701752
>>21701752

>70-90ºC
>standard

HAHAHAHAHA

my gtx 560 doesn't even reach 50ºC in full load.

>>21701889
I bet it's quite noisy under load.

>>21701889

Are your heatsink fins near the bottom bent a little? It's driving me crazy because I don't know if it's normal or not.

>>21701873
>the 8800gtx (or ultra)

Oldfag detected. No one can remember past Fermi or HD 4800.

The X1800 and 2900XT come to mind.

I've always felt that my reading were inaccurate. (they just take the readings from speedfan for temps)

They only reach 42c at load.

>>21701927
>2900XT
>Slow as and hot as fuck

>>21701927
should have realized half the people itt are 17 irl when they started freaking out about his 8800gt's temps

>>21701927
It was only 5 years ago. It was just retard tier for the price.

>>21701233
>dye

>>21701914

It does get pretty noisy :(.

>>21701922

I don't think mine are bent, but I honestly don't know. I can't see from the outside of my case and its too much trouble to open it up to look since I'd have to move a shitload of stuff around.

>>21701873
Even HD4870 outperforms 8800 ultra ya bastard. Nvidiafaggotry detected.

>>21701987
The HD4870 hadn't launched together with the 8800GT.

AMD was basically non-existent before the 4800 series, which was so good it forced Nvidia to slash its prices by like 70% overnight.

Also the 8800 Ultra is one of the worst, if not the worst GPU ever made. It literally self destructs. The 9800GX2 is equally bad, and the GTX590 comes pretty close.

It will be difficult separating from my 3y old friend.

>1020 core clock constant, fuck yeah

>>21701987
...yes

Isn't the 4870 from the next gen, though?

>>21701942
Yeah that's why I bought a HD 3850. I wish I had picked up an 8800GT in those days though.

>>21701956
>should have realized half the people itt are 17 irl when they started freaking out about his 8800gt's temps
>HD 3850
>93 C at load
>2007

Those were the days.

>>21702034
two generations. the 8 series came out in 2006, the 4800 series in 2008.

>>21701927
Is that a real motherboard?
If it is, what is it called?

>>21702031
Remember when Sapphire Toxic cards were good?

Is OP going to keep this same acting over the next 6 months?

He will just end up disappointing a lot of nvidia fans.

http://www.tweaktown.com/articles/1131/crossfire_hd_2900_xt_vs_8800_ultra/index.html
Fucking newfags everywhere.

>>21702017
>and the GTX590 comes pretty close.

In every dual GPU review I've seen that faces off the GTX 590 vs. the HD 6990, they pick the GTX 590 solely based on the fact that it is not as noisy.

I can deal with a hot video card. A noisy one is another level of pain.

>>21702042
my 3870 idled around 70 and went up to around 120 load

and my 4890 idled around 80. though it didnt get much over 100c under full load.

fucking case had only 8cm fan ports

>>21702065
ASUS ROG Extender

>>21702071
You mean before the 6xxx cards? Yeah, it wasn't that long ago.
Beefy PWM/VRM, high quality mosfets and silent as fuck fans.

Now they put on a toxic sticker with no circuitry improvements, fucking Sapphire.

>>21702094
now you honestly can't be telling me those who own $600 cards are going with aircooling.

but even when under water the 590 wins because nvidia always overclock better than ati.

>>21702128
Toxic cards are just factory OC'd with vapor-x HSF. Why would they add anything more than that?

>>21702152
maybe this gen.
but the 48xx and 58xx both overclocked better than the 2xx and 4xx series

>>21702152
>now you honestly can't be telling me those who own $600 cards are going with aircooling.

Mmm, quite true. I only look at the water cooled dual GPU cards.

>>21702173
Because a good HSF isn't the only thing needed for a good overclock.

>>21701291
Are they really going to be that much?

>>21702178
this is a joke right?

i can oc 200mhz on ati cards and i'll see 5% increase in performance, while overclocking by just 100mhz on nvidia cards i'll see around 15%

>>21702217
your not comparing the same thing that i was.

>this is a joke right?

no, your comparing the results of your overclock, not the ability to overclock.

>>21702194
As evidenced by the 590

>>21702217
I'm...really too lazy to look up benchmarks to see if you're right or talking out of your ass.

...so do the work for me and go find em.

Please?

>>21702217
you are doing it wrong

>>21702217
http://www.techpowerup.com/reviews/HIS/Radeon_HD_6950/32.html
>The actual 3D performance gained from overclocking is 11.3%.

http://www.techpowerup.com/reviews/Zotac/GeForce_GTX_560_Ti_448_Cores/30.html
>Actual 3D performance gained from overclocking is 10.1%.

How about no.

>>21702217
this is a joke right? Do you use Celeron?

>>21702238
well now you're just retarded, overclocking has always been about performance, and nvidia is just king at clock for clock.

since you're a retard and only care about clockspeed, nvidia holds the crown for overclocking records anyways. you show up at a benchoff with an ati card and you'll get fucking laughed at

>>21702251
eh, I kinda lied

http://www.guru3d.com/article/radeon-hd-6950-1gb-vs-geforce-gtx-560-ti-review/17

looks like the 6950 handles overclocking pretty well, at least on par with the 560 in gains from it, yes?

>>21702305

And for overclocking records, check the official source: http://hwbot.org/benchmark/3dmark2001_se/

intel/nvidia has always been the king.

It's funny how Nvidia fanboys think Nvidia is still good.

Nvidia was good when they made the 8000 series, then they stopped caring about desktop GPUs and now only care about ARM, Tegra and Project Denver.

That's why the 9000 series was all rebrands, that's why the GTX400 and 500 series have been the worst in the company's history, that's why AMD/ATi cards perform better, overclock better, and even have better drivers since the 4800 cards.

Nvidia has stopped competing, they don't even care about desktop GPUs anymore. Their products are all universally shit.

>>21702322
Might I direct you to 8.5GHz Bulldozer?

>>21702322
>>21702350
He's talking about synthetic benchmarks, not real world usage.

Nvidia always did better in synthetic benchmarks, it's part of their marketing strategy.
They do it again with the god-awful Tegra in ARM, where they increased the FPS cap just to score higher in benchmarks than the competition, meanwhile Tegra 2 was weaker than single-core SoC GPUs.

>>21702345
Oh hey look, it's InElite's alter ego!

http://vr-zone.com/articles/amd-radeon-hd-7970-listed-online/14277.html

>>21702322
>3dmark2001

>>21702350
I think what our nvidia fanboys are getting is that Bulldozer isn't so good when overclocked because they use a lot of power.

But at the same time, it's completely acceptable for nvidia to use more power than anything else in the world.

>>21702370
>over 500
ow, just ow.

hopefully that'll go down for the actual launch.

>>21702305
again, i said nothing about the performance. nor about the current gen. what i said was the the 48xx series, and the 58xx series, which also killed the 2xx and 4xx series in performance, overclocked to much higher clock speeds.

i didnt even mention that it yeilds better results on the 69xx series, nort the 5xx series.

again, your comparing something that i didnt even mention.

and you, have never once been to an "overclocking competion" your a little shit, know it all that sits on his ass arguing about shit that is completely irrelevant to the current conversation.

your wrong, grow up, let it go

>>21702413
That's nothing. Nvidia set the gts280 for 600-650$, 1 week before the 4870 was released.

>>21702396
nobody is being a fanboy here, bulldozer is just universally known as shit.

no, don't even put intel into the competition, bulldozer even fails to beat its predecessors.

>>21702393
http://hwbot.org/benchmark/3dmark11_-_performance/rankings?cores=4#start=0#interval=20

>>21702393
That pic is the first thing that made me laugh all day.

And the closest thing to a "lolcat" (god that hurt to type) that I have ever laughed at

>>21701291
>shitty xbox ports
it's the other way around most times. except for ubisoft games. but why the fuck are you playing ubisoft games?

>>21702413
Leak prices are always stupid, they will go down it's always like this.

But remember the price is based on PERFORMANCE, not on the 28nm yields and architecture.

Why do you think Intel chips that are so small cost so fucking much? They're good performers.

>>21702451
>3D Mark
>relevant
Stop reading shitty sites.

So how do these cards perform for games that actually matter, like skullgirls?

>>21702413
>euros
>$395 american dollars
>limited release
>dutch website
keep trying.

>>21702456
;_; and bethesda games

/sigh
getting 22 FPS on ultra in whirerun with a dual 6950 955BE setup makes me wanna cry.

>>21702413
TSMC's yields at 28nm are so bad that Nvidia's 28nm GPUs aren't coming for many months (couldn't afford) and Apple decided not to make the A6 at TSMC and instead went back to Samsung (the same Samsung they're suing all over the world).

I wouldn't be surprised if AMD's HD7000 series launch was only on paper like Nvidia usually does, they probably don't have many cards at all.

>>21702437
>your wrong

nope

you dont even know what this fucking thread is about

filtered your ip and called the cops, dont bother responding

Nvidia/Intelfags have no idea how competition improves the quality of the product. If AMD gets out of the market all we'll have will be crap for the next decade and beyound.

>>21702473
I'm sorrry, what the fuck are you talking about?

>>21702042
WTF you talking about? I have a 3870 just coming off load, bro.

>check dem temps

560ti idle

>>21702501
Why would AMD go out of the market?

A few months back they had 90% of the GPU market, and last year they passed Nvidia in marketshare, they are now number two, and Nvidia is number three.

Furthermore, Apple dumped Nvidia for AMD GPUs, and even the upcoming WiiU is using AMD GPUs.

AMD's GPU department is more healthy than ever.
In fact, if it wasn't for Nvidia getting regular payments from Intel for the next 6 years, they'd be in deep trouble right now.

Nvidia is kept alive on Intel cash.

>>21702499
>mentions that nvidia overclock better.
>gets told.
>refutes with something completly off topic
>gets told again
>filtered ip
nope
>called the cops
are you retarted


yea,
>your a little shit, know it all that sits on his ass arguing about shit that is completely irrelevant to the current conversation.
sounds about right

>>21702529
90% of the DirectX 11 GPU market* (i.e. current generation)

>>21702481
parsley, I get 60+ on ultra at 1920, u jelly parsley? And I don't even play the game correctly.

Remember when Jen said Fermi would buttdemolish AMD and when he said Tegra 3 would be amazing?

>>21702542
sure do. but it didnt stop me from buying a gf110 when they came out.

>>21702542
Tegra 3 is amazing

>>21702501

and thanks to bulldozer failing so hard we're already looking at $600 hex cores

>>21702529

amd is actually giving up on the desktop market and going into arm

i dont blame them. they have the r&d the size of intel's bathroom

http://www.xbitlabs.com/news/graphics/display/20111214212532_AMD_We_Are_Shipping_Our_28nm_GPUs_Now.h
tml

>AMD has begun shipping 28nm GPU's for profit
Meanwhile at Nvidia:
>We will begin production "soon"!

Stay classy, Huang.

>>21702568
>and thanks to bulldozer failing so hard we're already looking at $600 hex cores

You don't need a hexcore unless you need a workstation. It was the same situation under Nehalem.

>>21702565
>Tegra 3 is amazing

/g/...

>>21702561
So you waited 14 months after 5870's release.

>>21702565
Sure thing bro. Enjoy your moar cores on a tablet

>>21702583
ITT: same guys that said a dual cores were enough back in 2008 /g/

>>21702501

Brings back memories. Buggy Pentium FPU bug anyone? Pentium 4?

If not for AMD you would be eating overpriced shit today.

>>21702619
You mean the guys who were right?

>>21702601
no, i went from a dual 4890's to a 580, so it was almost 3 years.

I bought a 6850 before I had my entire build together. Fuck.

>>21702540
Yeah, a little.
I can 60 fine in most areas, but certain views in a city will kill me, fucking sucks man.

>pic related, it's the worst FPS I get
shadows are on medium I think, as they tank perf pretty hard for some reason.

>>21702631

i can't hear you over my q6600, my 930, and soon to be octo ivy

>>21702619
The Core i7 920 was $300 at launch. The Q6600 could be picked up for $266 for an older LGA 775 based system.

The problem with LGA 1366 at the time was the $250-300 boards and triple channel DDR3 RAM. All that changed when Lynnfield came out in September 2009. $100 boards and 4 GB of DDR3-1600 for $80. An acquaintance of mine has a good article on this.

http://www.anandtech.com/show/5174/why-ivy-bridge-is-still-quad-core

great, a picture promising insane performance came up, while the newest amd cards are a reality (testers have them, you can actually see some bench results, photos etc), no thx ill stick to amd, i might be biased coz before the amd card i have atm i had 2 nvidias, both burned (7800gt burned, on warranty i got a 7950gt which also burned, so i got a hd 3850, 40C idle, 75at load overclocked, runs fine), fuck nvidia.

>>21702651
why can't i run,
all these single threaded apps

>>21702651
>octo ivy

Intel doesn't compete anymore, they have no reason to improve things.

Ivy Bridge doesn't have 8-cores, it's just more quad cores for the consumer market, with no CPU performance improvements, the only things they're adding are (finally) native USB 3.0, PCIe 3.0, lower power consumption and better GPU.

>>21702685
x86 iS DEAD


DEAAAD

>>21702658
Oh god the RAM prices, I remember paying 160 eur for a 6GB kit. I also had the cheapest board on the market - more than half the price was for the X58 chipset

>>21702685
>no CPU performance improvements
15% better at the same clocks =/= no performance improvements.

>>21702658
nobody said it would be cheap to enjoy the enthusiast line for the next 3~4 years

and there WILL be octo ivy, just not at release.

thanks, amd.

Fuck all the numbers. To me the decision is between "Do i want six monitors, or do i want 3 3D monitors?"

Because I actually get work done. And play video games. Sometimes at the same time.

>>21702738
>15% better at the same clocks
Who feeds you this shit?

>>21702738
>trusting company info when there's no architecture change at all on the CPU department

See me after class.

>>21702727
i paid $170 for cl7 1866 xms 3

i sitll use them, though i can not for the life of my find cl7 of any brand any speed.

That's nice, but how's the power consumption and temps? And the size of these new cards?

And what games or applications are there that can take full advantage of this card?

>>21702738
Don't forget at 77W vs. 95W.

>>21702747
Sandy Bridge-E has octocore processors. They are just limited to Xeon and DP systems until LGA 1356 is out.

>>21702756
There are a few leaks floating around besides the Intel slides.

Also...
>not having a quad core
>2007

>>21702756
http://www.techpowerup.com/155994/Ivy-Bridge-Official-Benchmarks-%E2%80%93-Markedly-Better-Performan
ce-Than-Sandy-Bridge.html

+7% higher overall SYSmark 2012 score
+14% higher overall HDXPRT 2011 score
+15% higher Cinebench 11.5 score
+13% better ProShow Gold 4.5 result
+25% faster performance in Excel 2010
+56% faster performance in ArcSoft Media Expresso
+192% higher overall 3DMark Vantage score
+17% faster performance in 3DMark Vantage CPU benchmark
+199% faster performance in 3DMark Vantage GPU benchmark

>>21702782
Unfortunately its the future.

>>21702780
That kit I got there was 1600 cl 9, clocked perfectly fine without tweaking timings at 2000 though.

Not that it matters for x58 where you have plenty of bandwidth anyway

ivy bridge should see 5ghz on air pretty easily too, maybe 5.5

>>21702756
>>21702780

not that shithead, but realistically we're looking at a minimum of 30% performance increase for ivy when compared to sandy.

intel will be shipping those babies at the stock clock of 4.6ghz according to that spec sheet, and when turbo hits it'll be 5ghz.

on the stock cooler.

when pared up with watercooling, 6ghz is not too far off with ivy.

and to think you're lucky if you can hit 5ghz on sandy without using insane voltages.

>>21702821
These are Intel's "official" benchmarks where they try to use benchmarks that showcase the GPU improvements to pass it off as a general performance increase.

Unless you're using an integrated GPU, Ivy Bridge is like you're upgrading nothing at all.

>>21702685
also officially supported ram higher than 1333mhz

>>21702838
yes, well. it is 7-8-7-24
and i ran it around 2200, though it allowed for me to keep my 930 at 4.4 at 1.32v

i am still kicking myself in the ass for getting a 2500k. fuck what benchmarks say, that thing ran so much smoother for everything.

>>21702847
>stock clock of 4.6ghz according to that spec sheet
what

>>21702847
>intel will be shipping those babies at the stock clock of 4.6ghz according to that spec sheet, and when turbo hits it'll be 5ghz.

wut

http://www.cpu-world.com/news_2011/2011112701_Ivy_Bridge_desktop_CPU_lineup_details.html

http://www.cpu-world.com/news_2011/2011112702_Ivy_Bridge_desktop_CPU_lineup_details_part_II.html

>>21702855
Who cares, RAM performance hasn't mattered in ages, you're looking at less than 2% overall performance difference between DDR2 and DDR3.

>trolling

facepalm

>>21702847
...What the fuck are you smoking?


Oh man this I gotta screencap, oh holy fuck.

>>21702852
>benchmarks that showcase the GPU improvements
except you're forgetting about all the CPU specific benchmarks there?

>>21702888
dated as hell, look at the new spec sheet, it was posted on ocn just a week ago

>thread derailed into ivybridge
/g/ is funny.

>>21702847
>>21702499
>>21702217

Samefag.

>>21702847
6 GHz? I'm no computer engineer but I'm pretty sure that's physically impossible.

>>21702847
MAXIMUM HERPAGE

>>21702847
Here >>21702940?watch#21702940

I made a thread dedicated to your stupid.

>>21702935
Well you'd make a pretty terrible 'computer engineer.'

Record's up to 8.5ghz now.

>benchmarks optimized to hype next gen product and make it seem better than it actually is
nothing new here, move along. people will still buy AMD.

>>21702847

You want warp Drive, Phasers, and antimatter reactors for your impossible 6Ghz processor.

>>21702917
>dated as hell, look at the new spec sheet, it was posted on ocn just a week ago

A link would be nice.

>>21702847
>>21702842
Silicon-based chips have a hard limit at around 5 GHz due to exponential heat emission. I would be suprised if they could push much further than that under normal circumstances.

Our new Graphene overlords can't come soon enough. Too bad it will still take (many) years until those fuckers market-ready.

>>21702959
I love how you put that in quotes even though it's an actual professional title.

>>21703031
you do realize sandy will break 5ghz and will go as far as 5.6 if you give it around 1.5v?

ITT: Uninsured homeowners.

>>21702727
DEM FLOODS
THEY WILL COME TO RAM TOO

>>21703059
Do you realize you're retarded and smaller process nodes have less of a voltage limit threshold than their predecessors?

>>21702847
I heard these new CPUs will come without a cooler.

>>21701672
Well, to be fair, the nvidia car should be on fire.

>>21703095
nope.

sandy actually has a higher voltage threshold than its older brothers.

1.55 will bring you beyond 5.3, slightly below 1.6 and you're looking at 5.5 if your processor isn't shit.

that is all of course, under a full sized loop.

>>21702639
Don't blame your graphics cards it is your cpu that is the problem. Now in 90 percent of games you wouldn't need an intel cpu, but skyrim is not one of those games. You see bethesda decides to be fucking retarded (there is no other explanation for it) and only use 2 cpu cores and then when only using 2 cores they decided to render shadows on the cpu, that coupled with poor optimization leads to the mess we call skyrim. I can almost guarantee you would be able to max it out on an i3 2100 because it has 2 more powerful cores.

>>21703216
>Don't blame your graphics cards it is your cpu that is the problem
Yeah I know, when did I blame my graphics card?

>Now in 90 percent of games you wouldn't need an intel cpu, but skyrim is not one of those games.
All of my hate, that's OC'd to 4GHz mind you.

>You see bethesda decides to be fucking retarded (there is no other explanation for it) and only use 2 cpu cores and then when only using 2 cores they decided to render shadows on the cpu

Not entirely true, in task manager I can clearly see all 4 cores being used extensively, but I believe you with this bit:

>poor optimization

>I can almost guarantee you would be able to max it out on an i3 2100 because it has 2 more powerful cores.

Nope, there are people complaining about this kinda shit with a 2500K at stock, a 2100 would have it worse.

>>21703216
http://www.tomshardware.com/reviews/skyrim-performance-benchmark,3074-9.html

>>21703059
Hence
> around 5 GHz

There's only that much leeway with Silicon left.

AMD pulling out the desktop sector kinda eliminated Intel's need to further Graphene chip research. No 500-1000 GHz CPUs anythime soon.

>>21703254
Just reread my original post, and more importantly, who I was responding to, I can see how it looks like I was blaming my GPUs. I apologize, I did not mean to do that.

>>21703283
Also as seems to be shown in
>>21703261
While it does USE 4 cores, it sees very little benefit over 2. That's just so sad.

Is there any word on bethesda fixing the shadows, or are they leaving it like that?

>>21703267

but that clock speed is far from extraordinary.

ofc you aren't going to reach that speed safely with your $20 212, but given a full loop, sandy will go beyond the standard 4.8 and well into 5.5, and ivy will reach even further.

>>21703321
Knowing bethesda they probably aren't gonna fix it. And to make it worse when you set the shadows to a lower setting you aren't actually lowering them, you are decreasing their detail but increasing their range. You can't change the values independently. So while the shadows do look worse you aren't getting as much of a performance boost as you should.

>>21703321
>bethesda
>fixing

Pick one.

>>21702513
Hows that 3870 handling games nowadays?

>>21703375
...great.

great.

is there any way to do it in the skyirm/skyrimprefs.ini files? I might read up on them.

>>21702568
They arent giving up on the desktop market, its not just their #1 priority.

>>21703261
Well I was wrong about an i3 being able to max it, I was just making a guess there as I know the two perform similarly in games with the i3 usually doing a lot better in dual threading games so I assumed it would be enough my mistake.

The thing about it using more cores still stands though as you can see you get a benefit from a third core but not a lot, and then almost no benefit from using a four core. Yeah that is technically using more than 2, but I wouldn't exactly say it is quad threaded.

>>21703392
You can edit the shadow values this link has all the values you need to edit. I suggest playing with them until you get something that is more playable, and hopefully looks about as good (maybe a little worse) than what you have now.

http://www.skyrimnexus.com/downloads/file.php?id=148

>>21703461
Huh, I'm not sure what kind of black magic this is but I'm getting better (32) FPS in that whiterun spot where my SS was, and shadows look a helluvalot better.

Thanks!

>>21703866
Wait wait no it had set some other options on "high" from "ultra". Just about the same FPS.

shadows still look better, though.

>>21703216

This rumour spread quickly without being verfied. Skyrim uses more than 2 cores, it uses all 4 physical cores on my 2600k

>>21702651
ivy is QUAD, and maybe dual tops, intel have already said this.

if you want more than 4 cores, you go SB-E now or wait til Q4 2012 for IB-E

All joking aside, Nvidia completely redeemed themselves from the 480 disappointment. My 560ti is a motherfucker of a card for the money

>>21702437
>Can't use contractions
>Tells other people to 'grow up.'