AMD Zen CPUs
Images are sometimes not shown due to bandwidth/network limitations. Refreshing the page usually helps.
You are currently reading a thread in /g/ - Technology
You are currently reading a thread in /g/ - Technology
Who else is fucking excited for AMD Zen?
>Falling for AMD hype
prepare to be bulldozed
Only one more year to go.
I'm reserved about it.
You would be too if you remembered /g/ and bulldozer.
AMD Zen will do to Skylake what the Athlon 64 did for Pentium 4.
Absolute ass kicking.
Get hype.
>>52167756
I care more about Raven Ridge than Summit Ridge.
>>52167794
AMD has never been so desperate to release a product that can match Intel cpus. They are nearly bankrupt.
It is now or never. I dont want to be dramatic, but if Zen fails, AMD's game is over.
>>52167819
My thoughts exactly. I have to believe the hype, otherwise... Let's just say AMD is already circling the drain
>>52167834
I thought the same during bulldozer.
Sandy bridge was destroying the phenoms.
>>52167834
They have console market on lock, they're not going to "go bankrupt" for a little while longer.
>>52167838
At this point AMD is in the food disposer.
>>52167870
>They have console market on lock, they're not going to "go bankrupt" for a little while longer.
Kek
The consoles are not helping. 500 million consoles, but the profit margins are RAZOR thin. Think 500 million cents.
In real life though, AMD has been bleeding 125 million dollars A QUARTER for the last year and a half.
>>52167756
Intel will give OEMs "discounts" once again for using their CPU in laptops and business. AMD will still get cockblocked in sales which means no further R&D budget.
>>52167833
>more cores
yeah that has worked really well for AMD in the past.
They need a completely new design from the ground up, whatever they are using now can't compete with intel.
Unless you want to ramp up to 200W TDP
>>52167912
Your post is hilarious.
I'll let you try to figure out why you're a total shit eating retard.
if it is intended for passive cooling then excited
I would really love to see zen beat the shit out of memelake so that both companies will work harder and make better improvements per generation. Same with nvidia and radeon, i want them to go balls to the wall for improvement.
>>52168051
Its unlikely that the Zen core arch will be competitive against Skylake, regardless neither vendor will be releasing anything with significantly higher serial integer performance. Zen+ will be the last big hurrah from AMD then they'll be on the same 2-5% performance uplift cycle that intel has been facting.
High performance X86 cores are so well fleshed out that coming up with even a 1% performance uplift is a tremendous investment of resources and man hours. All of the low hanging fruit have long since been picked.
Its ultimately an issue with our approached to ICs themselves totally independent of the ISA. All Von Neumann architecture has finite throughput potential.
>people think Intel is intentionally gimping the performance of their latest CPU arch instead of this simply being the result of silicon reaching its zenith of diminishing returns
Most realistic good scenario for Zen is comparable IPC to sandy bridge.
Any higher than that is fantasy and highly unlikely (but not impossible).
More likely than it being better than ivy bridge is that it's worse than sandy bridge.
Let's just hope it's better than AMD's best processors; the Phenom IIs.
Did any details on the Zen core pipeline/frontend get released yet? All I heard was 3+3 ALU/AGU but I can't be certain
>>52168541
>Phenom IIs
>AMD's best processors
wahaha I responded
>>52168554
Its a 10 wide core
4ALU-2AGU-4FPU
http://dresdenboy.blogspot.de/2015/10/amds-zen-core-family-17h-to-have-ten.html
Fury X based gpu
Well, in the very least, cheaper DDR4 compatible chips might be nice. But I'm hoping Zen makes me proud. We'll see what comes of it.
>>52168580
>Its a 10 wide core
You fucking what, Intel can't even utilize it's 6 wide cores.
CAN'T END THE ZEN
>>52168590
>I have no idea what any of these words mean
Haswell has 4ALUs and 3 AGUs if you simply their load/store structure. Guess what if comes to if you add the FPU?
Waiting for that sexy new unified socket.
>>52168614
They actually removed 2 ALU pipelines and shifted the FPU
well done AMD
A 4-core Zen = 8-core FX
>>52167756
I am, but I don't plan on getting one myself.
I just bought a Skylake processor.
Tbh senpai I hope that Zen is a smash hit, I hope it makes me feel like I made a bad buying decision.
I just want AMD to live, to be the alternative to intel. I don't want to live in a world with just intel.
That and when I'm planning a budget machine, an AMD processor is the first thing on my list, every damn time. I don't even bother considering budget intel processors any more.
>>52168614
Theoretically Excavator would do much better than Zen if devs weren't lazy and had languages with proper multithreading support that's not a cobbled mess of mutable memory black holes
>>52168633
haha, didn't know that /g/ swaps the urban shorthand for family to senpai.
Just had a polite chuckle at my folly.
>>52168580
> Excerpt:
4 wide decoders
4 integer ALUs
2 AGUs (for 2R 1W L1 cache according to a LinkedIn profile)
4 FP pipelines
That makes z ten pipelines with a general four wide design.
LOL even Apple has wider design.
AMD Zenshit confried for massive flop.
AMD is kill
>>52168594
CAN'T STOP THE AMD DROP
>>52168624
wut
>AMD will die in a few years
>mfw you'll have the pleasure of experienign it first hand
>>52168641
Theoretically, no it wouldn't. Moron.
Excavator still suffers from the Bulldozer family's cache structure and high latency. All of the incremental architectural changes from Bulldozer to Excavator only brought a 27%-30% IPC uplift. There isn't any left over performance hiding due to bad coding, the architecture itself is lacking.
One of the trade offs made in Steamroller was higher cache bandwidth over Piledriver, but this came at a cost of increasing latency to a significant degree, L2 took a really bad hit. Most of the IPC gains of the arch were lost due to this. Excavator put yet another bandaid on the issue and simply cut the size of the L2 in half to improve on latency. This is indicative of an issue so deep rooted that it could not be fixed. The entire core pipeline is fundamentally flawed and there is no way to rework it without gutting the entire design. You would essentially be creating a whole different core that only resembles a Bulldozer derivative, and the cost to that would be millions of dollars and years worth of work. For such little gains it simply would not be worth the effort.
Zen brings a 40% IPC uplift over Excavator.
A single Zen core has higher throughput than a complete Excavator module. AMD finally made a good choice by abandoning the Bulldozer family.
>>52168649
Good job outing yourself as a first time poster. Now go back to reddit.
>>52168742
Fuvk off pisskin
>>52168742
Lol look at him pretending like he knows anything
>>52168771
>be totally lost in a very basic technical discussion
>resort to shitposting because you're so insecure
Adorable.
>>52168742
>Zen brings a 40% IPC uplift over Excavator.
According to internal marketing slides.
Which have always been 100% accurate.
They're usually true on a technicality to avoid being sued, but they're never actually true.
>>52168799
>he thinks copypasting shit from anandtech forums is technical discussion
Precious.
>>52168802
It wasn't a marketing slide. It was part of an investor presentation. Look at the presentations they've given at ISSCC, and all the in depth detail they provide. Post restructure AMD is an entirely different company than they were a few years ago. Upper management was absolutely gutted. Su even trashed the head of their marketing department as a matter of fact.
When they first detailed Carrizo they touted a 5% IPC uplift over Kaveri, and they delivered that. 5% is only the average, and it can reach 9-13% uplift in certain workloads. Taking all that into account it would be totally baseless to doubt the 40% uplift figure for Zen, not that 40% would give you anything spectacular regardless.
Zen probably isn't going to compete against intel's latest offerings per clock in most things.
>>52168805
>insecure tech illiterate retard still just shitposting
You're not making yourself look any better.
>>52167911
I think part of AMD's strategy for remaining relevant will be to split funding to R&D and raw lawyer count to fight intel tooth and nail for all the bullshit they've pulled in the past.
>>52168862
>Taking all that into account it would be totally baseless to doubt the 40% uplift figure for Zen, not that 40% would give you anything spectacular regardless.
>Zen probably isn't going to compete against intel's latest offerings per clock in most things.
It should manage Haswell levels at least.
Skylake isn't that far ahead of haswell either. (and manages to regress in some loads)
>>52168862
Still, I think skeptical is the right approach to such a bold claim. Fool me once, shame on AMD, fool me twice, shame on me.
I remember the bulldozer hype too well. That said, Zen being good will be nice. I mean I've planned my next upgrade around Zen; if it's good I'm getting Zen.
If not; I'd better get a recent intel processor before intel prices shoot up.
>>52167756
Only excited because Intel will actually play some of their hand they've kept hidden for a while now.
Are we talking 40% faster in synthetics or real workloads? Because Haswell is around 30% faster per core than Piledriver in real loads, but almost double in worst case synthetics(low thread, FP only)
>>52168929
Intel aren't hiding anything. Remember that they're competing against themselves. Their goal for new product launches is to get people who have sandy, ivy, haswell processors to buy the new ones.
They're not holding anything back- improving IPC on x86 around 14nm just really is that hard.
>>52168926
The biggest source of Bulldozer hype was a single AMD employee who was talking out of his ass on Overclock.net. He without authorization was making comments about the arch when he wasn't in any position to even be privy to that knowledge. He explicitly stated that IPC would increase, and it turned out to not be true.
He was fired.
The other half of the hype was just coming from fanboys blowing smoke up their own assholes. AMD didn't officially make any public statements about Bulldozer like they have with Zen.
>>52167756
Well, i used to be excited about Zen before it was known when both Skylake and Zen will be released.. But Intel was much faster, so i've ended buying 6700k. I would probably buy Zen, but AMD is few months behind.
AMD has never been competitive, why makes you think it'll be now? Get over it, AMD has always been a third rate company copying Intel designs and somehow not getting sued for it
>2016
>a thread about actual technology that is not graphic cards
>/g/
I'm impressed
>>52169077
This thread is shit m8, everyone falling for a crook company scam
>>52168958
>AMD didn't officially make any public statements about Bulldozer like they have with Zen.
Well I factor in AMD's financial reality.
I think they'd have no problem lying. Because what's the worst that'll happen? If Zen fails they're dead no matter what so having their bankrupt company sued doesn't mean anything.
They need the positive image to keep the investment coming to even finish Zen at all.
>>52169093
This isn't a Intel thread yet, son.
>>52168719
>mfw Qualcomm will become the AMD we need
>>52168742
desu senpai i wouldn't know how to gb2r desu senpai. Dont have reddit lol
A lot of people here are missing that amd wants to compete against (at the time) canonlake and now possibly kabylake. If they say that they're targeting the high performance market they must have more than just Sandy bridge IPC, else there's no way to be competitive against broadwell-e when it releases sometime this year.
Ofc remains to be seen if they're talking the truth.
Personally I can't wait to finally buy zen + Arctic islands(GPUOpen + freesync).
>>52169108
Intel aren't the crooks selling you 3 core CPU which are actually 4 cores
>>52169102
Companies don't just die. I have no idea why this idiotic mindset is so common on 4chan, but you should put it out of your head entirely.
When you pull in several billion dollars annually you don't abruptly close up shop. The course a struggling giant takes is to continually spin off divisions, sell IP, and restructure downwards.
AMD has roughly $2,000,000,000 in long term debt. They're making $5,000,000,000 annually in revenue during their worst years in recent history, previously a $10,000,000,000 year was the norm. They're literally nowhere near bankruptcy. For fuck sake, VIA is still in business.
>>52167912
>yeah that has worked really well for AMD in the past.
good thing the 'more cores' thing is more important than ever right now.
>>52169140
Every analyst predicted AMD is dead before 2018, stay delusional.
>>52169190
Shitposters on /g/ who proclaim "X IS FINISHED" are not market analysts, though I can see why someone with diminished faculties would be confused.
>>52168719
>AMD
>dying
That'll never happen, Intel needs them from being broken apart
>>52169140
Well okay, but if Zen fails they're not going to be making competitive x86 any more. The critical parts of their company that is required to produce desktop x86 processors is not going to exist if Zen fails.
They will turn to using other people's basic designs and might become an ARM company.
>>52169205
>abloo abloooo
>abo bloooo
>aboo ablooo
That's you.
>>52169138
AMD never sold quad cores as three core CPUs you are straight up lying. They did sell two and three core CPUs at discount prices that sometimes had functional third and fourth cores but there was never a guarantee of that because those cores were occasionally disabled due to defects.
>>52169210
Zen will bring a performance uplift over Excavator no matter what, and all the Zen based chips are 14nm FinFET instead of 28nm planar. They'll have a performance uplift and a tremendous boost to overall perf/watt.
Summit Ridge could be roughly equivalent to Sandy Bridge-E per clock and they'd still have something viable to sell.
No matter how it ultimately ends up they'll be a in better position than they stand currently. They'll keep selling chips, building a portfolio of semi-custom IP, and pay off their long term debt one way or another. Work on the Zen+ core arch started early 2015, and that investment isn't going to disappear into thin air. They'll keep making new generations of chips.
What's the point of only being 40% faster when Intel is over 80% faster?
Jim Keller. Enough said.. he's destroyed Intel twice he'll do it again
>>52169356
>>52167912
>>52167991
Oh can I guess why hee a retard? Is it because Zen is his exact post...
>rework from ground up
>>52169380
He gone. Zen's the last thing he did for AMD, after that they're fucked
>>52167770
nice meme
>>52169356
what's the point of buying a new intel cpu when it's only 5% faster than SB
>>52169185
lol wat
>implying devs will spend precious jew time to optimise their programs for 8 threads
AMD needs to make an arch which has a strong ratio of instructions per cycle, otherwise the new Zen processors will fail in the same way the FX series did. There's a reason a quad core i5 could beat out the fx 8xxx
>>52169434
>implying
SKL is 20-30% faster, enjoy your memebridge
>>52169444
>memelake
get cuck'd, goy, enjoy your 5% ipc increase
>>52169444
>paying full price for a fractional improvement in synthetic benchmarked performance.
>>52169443
it's like you don't even know what this thread is about
>>52169444
>lake of memes
intel cucked you, faggot
>>52169468
So basically you're saying that "moar corez = faster proces"
For a cpu to be competitive, it's single core performance must match or outperform it's competition at that price point.
Well designed arch > More cores
>>52169500
>So basically you're saying that "moar corez = faster proces"
no? do you have any reading comprehension or has being an intelcuck killed off the last few brain cells you had left? seriously i can't even comprehend how you came to that conclusion.
>>52169511
I replied to a post saying that more cores was important, if it was sarcastic I did not pick it up
>>52169468
i noe rite.
zen is 14 nm . all the criticism amd gets needs to be put into perspective. they have been making 32 - 24 nm chips for a while and intell has had 14 nm chips for years.
skylake is still 14 nm no 10 shit like we got promised. intel made excuses about stability and scrapped the 12 nm line
amd doesnt have to beat intel out right in performance they just have to have a better price point. but i would expect good performance from a amd 14 nm chip
Even if AMD Zen is faster and uses less power nobody will buy them.
Because AMD are a noname brand linked to unreliability and problems.
>>52169543
more cores being much more important these days than it was when amd was pushing for more is entirely valid and more cores being important is not mutually exclusive to single core performance still being important as well.
>>52169553
nvidia isn't making zen.
>>52169557
I wasnt saying they are mutually exclusive, just that after the FXmeme shitshow AMD needs to focus on efficiency and not brute force to get their shit done, because devs arent going to go out of their way to optimise their processors for (realistically) more than 2-4 cores
>>52169553
>amd
>noname brand
I guess you didn't exist in the late 90's and 00's
>>52169618
>oldfags like us know it rite??
>>52169398
The image clearly states more CU, referring to the IGP.
The shit eating retard anon thinks this means more CPU cores.
>>52169402
Yea but Athlon is still bought so lol
>>52167870
>>52167898
If things get truly dire then Apple, Microsoft or Samsung would almost certainly be willing to buy out AMD for a bargain price
>>52170175
And even if they don't, Sony and/or Microsoft would almost surely bail them out, so they can keep making chips for their consoles.
I'm really tempted to swap my Phenom II out for a Skylake i5 but I feel like waiting for Zen is a much wiser thing to do...
What should I do /g/?
>>52170251
Flip a coin.
First Zen based part coming out is Summit Ridge, and won't be available til end of the year.
>>52170251
personally, i'm waiting for AM4 to drop then picking up a cheap carrizo chip and 16 or 32GB of DDR4.
AMD is fucking garbage. Everything they make is absolutely, objectively crappy as shit and only a fat neckbeard would buy this shit-tier crappy shit because he's afraid of running out of autismbux (because God forbid he spend less money on Taco Bell and Mountain Dew.) I tried building an AMD gaming rig once and it was the worst experience of my life. First off, that CPU clip is fucking impossible to use. It looks just like an old Pentium CPU clip, but it's autistically retarded and I couldn't deal with it. And the graphics card was so fucking huge, I had to cut a hole all the way around my case to fit it in. Fucking AMD.
But wait, that's not even the best part! When I tried to start up my shitty autismbox (which I sold to the Association of Retarded Citizens BTW) the absolutely massive power draw dimmed every light in my house and throughout the entire neighborhood like I was firing up fucking ENIAC in the 1920s or some bullshit. I had plenty of time to worry about my power bill too, because this shitty thing couldn't figure out which of its 32 CPU cores to use and it took half a day to boot. As soon as I tried to play a game (fucking Starcraft from 1998), it lagged so hard it went backwards. I thought I should try downloading a driver, but there aren't any. There are no fucking drivers. What the shit, AMD? No wonder Apple doesn't use your shit. It wasn't long after that that I realized this thing was putting out more heat than a goddamn jet engine and it burned my house down.
Two weeks later, when I sold that piece of shit, I built an Intel/Nvidia rig! Holy shit, it's like jacking into the Matrix! I built it 5 years ago and it still runs every game in 4K at 240 FPS while using no power and it actually cools my house during the summer. What a beautiful machine, I seriously hope none of you are poor and stupid enough to buy AMD.
>>52169121
This post gave me cancer
>>52170331
>5 years ago and it still runs every game in 4K at 240 FPS
Counter Strike?
>>52167911
Isn't that already happening?
I have yet to see a good carrizo laptop.
Intel might just be more subtle this time.
>>52168051
Well to be fair rebrandeon is giving geforce a run for its money.
Only issue is marketing.
>>52170526
With the abysmal battery life of all laptops with APUs I can't blame the manufacturers for avoiding them. Promising "8 hours of idle use" in their marketing slides doesn't really instill much confidence.
>>52170331
>AMD graphics card being big
this pasta is outdated
>>52170251
Have a phenom 2 as well and I'm waiting for zen. Why? Because I wouldn't buy a graphics card right now anyways due to hopefully a big jump with 14/16nm gpus.
I couldn't do shit with skylake due to my old gpu, so I'll wait at least until next year's gpus and see if amd dropped some news on zen and decide from there.
Never had an AMD CPU before. I'm definitely switching to AMD for my GPU next time I buy one, but I'm not so sure about the CPU.
>>52167756
literally just bought a 6600k to replace my old Phenom II X4 965, because I can't hold out for Zen anymore
If it's fantastic when it comes out, I'll be selling the 6600k
AMD will never go under because if they do then Intel will bail them out.
Worst case scenario, AMD stops making CPUs and turns into ATi once again.
>>52170695
I dont see a Gpu, also
>value ram
Anon...
>>52170695
why do british people have to buy a vat with everything? Don't you have enough containers by now?
>>52170755
Value added tax you mongoloid
>>52168510
Well, they sure as hell aren't TRYING to make anything revolutionary. Maybe this really is the best that can be done with current tech. Or maybe there's a different paradigm that would produce significantly better results if Intel were to sink as much resources into it as they have with their pentium-III derived architectures.
>>52170755
>enough containers
If that was true all the sand people would be out of the UK by now
>>52168719
HAHA like AMD isnt already undead
>>52170817
Thats the kicker, anon.
We're at this point because years and years of painstaking R&D has already been done. X86 has had billions of dollars poured into it.
There are a lot of potentially revolutionary designs out there, but they haven't had the same benefit. Because of that they can't compete with what we have today even though they may be able to surpass it and continue healthy scaling.
You can't just pull millions of man hours out of a hat, and diverting such a significant volume of human resources to one project would cause everything else to grind to a halt. Semiconductors aint easy.
>>52169123
Fury X was meant to kill everything too senpai
>>52170848
>AMD funneled most of their R&D into necromancy
>>52170695
6600k is a serious upgrade from a 965. I went from a OC [email protected] 2800nb 1333mhz 7-7-7-21-1t RAM to a 2500k at 4.5 and it was a small upgrade. Youre 6600kis legit upgrade I honestly regretted my upgrade. Now looking back it was enough to justify but until later this year theres not anything worth it.
>>52170853
I'm not even talking about moving away from x86-64. Prescott and the longer pipeline, bulldozer and CMT... I'm sure there are lots of other things that could be tried.
>>52169138
AMD gave me a quad core that was fully unlockable to a 6 core
For 90e
>>52170874
Dumb twitter ebonics posters should learn to segment AMD's GPU division and CPU division, like how you'd segment Samsung's phone and refrigerator divisions.
moutharfucking AMD ZEN KELLER "INTELKILLER". Will trash my intelshill processor for masterrace ZEN
Correct me if i'm wrong, but is the 40% uplift figure the slide quotes regardless of node zen will be on?
>>52167911
>Intel will give OEMs "discounts" once again for using their CPU in laptops and busines
The EU has been raiding Intel offices on random for some years now.
If Intel were caught bribing OEMs again, then I promise you, it will not take 8 years to reach a settlement. It will be far more than the meagre $1,000,000,000 that it was a few years back.
Intel, and more importantly their PR, cannot afford to take that risk.
>>52171096
IPC has literally nothing to do with process node.
Are sub $100 Athlons without iGPU going to emerge as a result of 14nm Zen being put into market? Or is it only reserved for high-end chips?
I'm going to build a new rig in late 2016 and would want something along of what X4 860k is today.
>>52168614
Nice job posting the fake slides.
Seriously. We've had the REAL slides from the actual FAD for month now. And retards still believe those fakes every time they come up.
>>52171111
>>52171096
Node could potentially impact IPC if certain components can't maintain the proper throughput rates.
Not everything runs at the same clock rate inside a CPU, mate.
>>52171120
mind linking the real slides then?
>>52171111
I never said it did, but if AMD was targetting a larger node suddenly having more space available (assuming its 14nm) would allow for design tweaks to get even more IPC from the chip. Thus its theoretically possible to see over 40% increase which would get Zen to skylake territory.
>>52171097
>their PR
I can fucking guarantee you that people will not be upset that their laptop was cheap. Nobody that knows anything about laptops buys AMD.
>>52171097
You're right, they won't try that trick again.
They'll use a new trick. Like spinning.
Okay but seriously intel is actually really good at less anti-trust methods, mostly marketing.
If Zen is competitive, expect retail workers to get more discounts for following 'intel product education'. Expect more TV ads that basically show how great and non-chinese the company is.
It's not a computer without intel, etc.
The way EU laws work is there's no punitive damages. Intel doing something on purpose has the same punishment as doing it accidentally, so you might as well do it on purpose and gain a massive advantage.
>>52171113
Summit Ridge is an enthusiast CPU with no IGP. Just 8 cores/16 threads to compete with the i7E line from intel.
The Zen based APU is Raven Ridge, and its not launching til 2017, though there will undoubtedly be binned chips with the IGP fused off you'd have to wait a very long time to see them.
Both Summit Ridge and Raven Ridge are 14nm FinFET parts.
>>52171125
You clearly have no idea what IPC even means, and you don't even understand the words you're putting together in that post.
IPC is a function of design and literally nothing else. A design has the exact same IPC no matter how its implemented.
Eversince I defended AMD.
But yesterday:
>want to view a movie via hdmi sound-outout to TV
>niice, there is an update for my amd Gpu
>do eet
>no Sound after it via hdmi whatsoever
>googled it - amd dun goof'd it
>Set my PC back to yesterday
>never gonna install a new driver for amd anymore
>never buy amd again
>>52171163
I'm assuming the last 3 lines are a joke- but AMD's HDMI audio driver is ass for some reason.
You'll need to download and install a seperate one. Maybe rolling back to window's own HDMI audio driver.
>>52169553
>Because AMD are a noname brand linked to unreliability and problems.
In my area, they're a name linked to Laptop performance.
The amount of Intel laptops I get returned because "It can't run DOTA and CS:GO good" is astonishing.
Then they go ahead and pick up an A10 powered laptop like their friend got, and what do you know. I never see it come back in the door.
AMD may not have much of a name in the enthusiast market. But just what 'enthusiasts'. People who spends thousands on their desktop, just to play a handful of MMOs.
Opterons are still a good option for companies who might not be able to afford Xeon based servers and workstations.
And despite 3 of their current GPUs essentially being Trinity rebrands, they still trade punches with what nVidia has to offer in their comparative price range. Despite being on 28nm, and 4 years old (in essence).
AMD are also the only ones who have actually innovated in the x86 market in the last decade. When was the last time Intel invented something?
Hell, AMD even innovate in the GPU sector. GDDR5 was AMD. HBM is AMD. Mantle pushed Microsoft to send DX12 out the door. And Vulkan with it.
Everything AMD do makes big waves in the market.
>>52167756
I hope I have cause to be excited. AMD really NEEDS to up their game, and Intel really needs the competition because they're starting to get lazy again, and is raising their prices. They CAN do it, because they've more or less caught up in nodes and because of Intel's Kaby Lake delay, and Jim Keller is a legend who can actually design CPUs well, not like the bozos who did bulldozer.
If it's better, it'll be remarkable and I'll be very happy. If it at least catches up, I'll be very happy because it'll drive Intel into lowering their prices drastically to compete and actually try again. We really need two players in the game or x86 is finished.
>>52170752
>>value ram
RAM is RAM.
Only reason not to cheap out would be for quality ECC for a workstation or server.
>>52171242
> it'll drive Intel into lowering their prices
PFFFFFFFFFFFFTTTT HAAHAHAAAAAAAAA
>>52171135
>>their PR
Having "X company bribed Y, sued for billions" over the headlines, isn't exactly good for your PR image. Investors especially. Even more so the companies that got bribed.
Stocks take a dive. Investors bail out.
And more importantly. It sends the message to consumers that they feel threatened by the competition. That the competition is better, and they don't want people to buy from them.
Intel would be stupid to make a move.
>>52171223
Its almost criminal (well, was) how AMD's apus have been effectively locked out of the laptop market - there is no way to get better performance in the same pwoer envelope for general usage than what AMD offers. Sure the cpu side still lags behind intel a decent amount but the gpu grunt is enormous.
>>52171301
There are a number of Carrizo based laptops for sale in the US. But I've yet to see any in stores here in the UK. It's kind of sad. Wouldn't mind one myself.
>>52171274
Investors care about making a profit. No lawsuit is ever going to affect intel's stranglehold on the market, nothing is going to affect their volume. They do $50,000,000,000 worth of business per year. Sometimes even more.
AMD could win a settlement against them for $20,000,000,000 and it wouldn't do much to intel's stock long term because they're guaranteed to bounce back from it.
am i the only one that loves monopoly?
now i can play latest games on max with 5yo cpu
>>52171274
People still buy Intel over AMD even with press about Intel #winning. Not everyone reads that they are bribing OEMs and gets upset, as far as I am concerned they still produce a highly competitive product and are simply better than the competition so its ok.
>>52171327
>$20,000,000,000 and it wouldn't do much to intel's stock long term because they're guaranteed to bounce back from it.
Actually with that much AMD could actually have an R&D department. But knowing AMD, a settlement of that magnitude means CEO bonuses.
>>52171150
And I suppose that's why the FX-4350 and A10-6800K have different IPCs, right?
I mean, they're on the same node, have the exact same core design. Should be pretty much identical, right?
I mean shit, they even have almost identical binnings!
6800K (4.1Ghz) on top, 4350 (4.2Ghz) on bottom.
Rough calculations puts the 6800K at 3.04 instructions per clock and the 4350 at 2.91. Normalizing the numbers for clock rate (using the 6800K's clock rate as the base) we arrive at 4350 at 2.84.
That puts the 6800K at 7% faster per clock than the 4350.
Node matters, even if on the surface it's "identical".
>>52171268
The reason it's making you laugh is exactly why it needs to happen!
AMD being actually as good as Intel for much cheaper will scare the pants off them, investor-wise. It happened last time. It'll make them try harder. They'll lower the prices to try to flood the market with surplus.
Also, they simply can't pull the same antitrust tricks they tried last time: EU is breathing down their neck and absolutely will have them for it.
>>52171364
>But knowing AMD, a settlement of that magnitude means CEO bonuses.
If the asshat Rory Reed was still in charge, no doubt.
But I get the feeling Su cares more about AMD as a company than her personal gain. After all, if she didn't care about the company, then she would never have fired half of the Marketing department. She would have just taken her millions and run.
No doubt that everyone in the company would see some kind of 1-time bonus as a "Congratulations, have a cookie" sort of deal. But I doubt Su would take the lion's share.
If anything, she's probably try to get Keller back for a few years with that kind of money to throw to staffing.
>>52171370
Amazing. You really are dedicated to making a complete ass of yourself.
Vishera and Trinity/Richland are made on literally the exact same 32nm PD-SOI node. There is no variation whatsoever, but that doesn't even have anything to do with your point. You have to have a brain tumor to make such a disjointed argument and think its sound.
The process has literally nothing to do with IPC, and the 7zip bench is doing nothing but showing margin of error.
The Richland chip has a slightly higher turbo, and the Vishera part has an L3. Apart from these two differences the CPUs are exactly the same. They have the same IPC. They matter of factly are the same CPU, and I mean that in the most literal sense. Their transistor populations are 1:1 identical, they share identical lithmasks for CPU logic.
>>52171216
No, i restored my PC to a safe back-up-point.
Amd Drivers are shit tier.
Wanted to watch my movie, don't wanted to do hours of trouble shooting and Drivers search
