Slashdot Mirror
AMD Continues To Pressure NVIDIA With Lower Cost Radeon R9 270 and BF4 Bundle
MojoKid writes "The seemingly never-ending onslaught of new graphics cards as of late continues today with the official release of the AMD Radeon R9 270. This mainstream graphics card actually leverages the same GPU that powered last-year's Radeon HD 7870 GHz Edition. AMD, however, has tweaked the clocks and through software and board level revisions updated the card to allow for more flexible use of its display outputs (using Eyefinity no longer requires the use of a DisplayPort). Versus the 1GHz (GPU) and 4.8Gbps (memory) of the Radeon HD 7870 GHz Edition, the Radeon R9 270 offers slightly lower compute performance (2.37 TFLOPS vs. 2.56 TFLOPS), but much more memory bandwidth--179.2GB/s vs. 153.6GB/s to be exact. AMD and its add in board partners are launching the Radeon R9 270 today, with prices starting at $179. The Radeon R9 270's starting price is somewhat aggressive and once again puts pressure on NVIDIA. GeForce GTX 660 cards, which typically performed lower than the Radeon R9 270 are priced right around the $190 mark. Along with this card, AMD is also announcing an update to its game bundle, and beginning November 13 Radeon R9 270 – R9 290X cards will include a free copy of Battlefield 4. NVIDIA, on the other hand, is offering Splinter Cell: Blacklist and Assassins Creed – Black Flag, plus $50 off a SHIELD portable gaming device with GTX 660 and 760 cards."
Time to get the shopping done.. These bundles are getting sweeter and sweeter.
Harrison's Postulate - "For every action there is an equal and opposite criticism"
Along with this card, AMD is also announcing an update to its game bundle, and beginning November 13 Radeon R9 270 – R9 290X cards will include a free copy of Battlefield 4.
Beginning November 13th, you say....
Maybe someone else has a decoder ring, but it's alphabet soup trying to figure out what video card one should get.
http://en.wikipedia.org/wiki/Comparison_of_Nvidia_graphics_processing_units#Comparison_tables:_Desktop_GPUs
http://en.wikipedia.org/wiki/Comparison_of_AMD_graphics_processing_units#Comparison_table:_Desktop_GPUs
If you stare at the article above, it's a blob of numbers worthy of A Simple Mind.
I left the PC gaming rat-race a while back, and I've never been happier -- the only real downside is that I can't possibly suggest to people what video card to buy beyond saying, "meh. Go spend $200 on Newegg."
How many 4k monitors can it simultaneously drive? If I want a 2x or 3x 4k setup, will it drive it? I see a Dual DVI, an HDMI and a DP (so max of 2 for Seiki 39" 4k unless the D-DVI and HDMI share a channel...but it doesn't say).
Is it just my observation, or are there way too many stupid people in the world?
Braille displays are comparatively low resolution, and can be driven in software over serial, USB, or Bluetooth without specialized hardware. (The exception might be older TTY/TDD systems, which use some...eccentric encoding schemes that are of very limited compatibility with many computer modems)
Oh really? I assume you've never had to write software for these electronic pintos
Mod me down, my New Earth Global Warmingist friends!
Best card for Final Fantasy XIV at 1920x1080 with maximum settings, taking into account they announced a DX11 update in 2014?
Get free satoshi (Bitcoin) and Dogecoins
Yeah and some poor bastards probably had to work 90 hour weeks to make the AMD GPU work enough to ship.
Mod me down, my New Earth Global Warmingist friends!
Nope.. unless of course you define 'shitty user' as someone who chooses to run something besides the latest 3 games the installed drivers were 'optimized' for, at the expense of compatibility with everything else.
Sorry, ladies and gentlemen. I was a longtime fan of Radeons, and I bought me a new shiny Radeon+Phenom notebook - just to find that the Radeon X-Windows drivers don't support FreeBSD anymore. They need a Kernel Mode Switch that is obviously absent. Now the FreeBSD team implements it while my book collects dust. The Nvidia drivers are closed-source and glitchy - but at least they exist and they work.
AMD drivers are shitty, and before that ATI drivers were shitty, even before ATI made 3d cards. I've been watching ATI drivers cause Windows to crash since Windows 3.1.
It's broadly believed that ATI's hardware is as good as or better than nVidia, but their drivers hold them back.
I'm still glad ATI is around, just to keep nVidia scared
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
Hmm lets see..
1. demoscene - breakage everywhere
2. game titles a few years old
3. anything that's not a game, even things like game map editors, tends to break, especially in opengl land. nvidia handles them fine, on geforce or quadro.
4. video playback causing bsods..
5. shitty linux support, though it's nice they're opening up the specs.
For what I have experienced, that perception is a bit outdated. AMD's drivers have been as good as NVIDIA's for about one or two years. On Windows. On Linux, NVIDIA is still way, way better. For newer cards. If you can use the proprietary driver at all.
AMD doesn't support BSD. Regardless of how much they beat NVIDIA, I'm not going to buy one of their cards.
Don't worry BSD is dead. No one gives a flying fuck mate.
Cost is only unimportant when you're spending someone else's money.
Is it just my observation, or are there way too many stupid people in the world?
Nvidia drivers are shittier than AMD. End of story. AMD drivers implement the graphics API's to the letter. Nvidia lets any old crap through. The result, is that apps developed on Nvidia GPUs, rarely conform to the target graphics API, and as a result end up failing on AMD/Intel hardware (although admittedly Intel can also fail due to either a lack of resources, or the occasional bug). If an app fails on AMD, blame the shitty software you're running, not the drivers. AMD releasing mantle would appear to me to be nothing more than a way of forcing Nvidia to adhere to an API spec for a change, rather than routinely ignoring it (as they do now)
I have written the 3D renderers for a number of commercial software packages for the film VFX and games industries. The parent is right. PEBKAC. The drivers are fine. You are a shitty developer (who needs to make sure your code conforms to the GL/D3D spec properly).
That's complete bullshit. Nvidia does indeed make their drivers very flexible, but it's trivial to force AMD's products into software mode (or black screen, blue screen of death/kernel panic, system freeze, trippy display noise, etc.) with totally valid state configuration.
Google it you fuck up.
Mod me down, my New Earth Global Warmingist friends!
I write a lot of OpenCL code at my job. I also mess around with OpenGL stuff in my free time, usually just screwing around with procedurally generated scenes and shaders (so no where near a full game engine, but hitting a lot of aspects of the modern pipeline). I haven't had any problems programming for the AMD card in my home desktop, whether for OpenGL, or for OpenCL when I was too lazy to log into a workstation in my office to run tests on small datasets. I've just gone off the specs as far as documentation.
I'm guessing you can just move the goal posts and say "I assume anyone not having issue with AMD never had to write software... that uses specific function XYZ." Of course, you could have said that in the first place and have been much more informative instead of making a post indistinguishable from a content-less fanboy rant.
1. The Catalyst control panel's dependency on .Net that made it bloated and slow to load.
The dependency on .Net doesn't make it "bloated", in fact if anything it offloads more functionality to the installed .Net libraries making it less "bloated". Also there is nothing inherent in .Net (or Java for that matter) that would make it take significantly longer to load, if it is taking a long time to load then writing it in a native language won't change that.
The dependency on .Net doesn't make it "bloated"
it might make the installation package larger if it comes with an offline installation of the .net runtime. but most use a web install for the few cases where the system doesnt already have it installed anyway.
if it is taking a long time to load then writing it in a native language won't change that.
the problem is more likely to be loading external resources than loading the runtime and yeah writing a native version wont help you.
Seriously? I've had to fix two "bugs" on NVidia hardwre in the middleware I wirk on as a job. Both were caused by AMD not enforcing the proverbial "letter of the law" vis-a-vis Direct3D and NVidia drivers correctly rejecting the relevant calls.
The driver on the 1650 was rock solid. It was like having an Nvidia card but without the crummy color quality (which even I, with my lousy color vision can see).
That said, the 4350 I tried to replace it with was junk. Nice card, good performance, dirt cheap, drivers crashed on everything but Call of Duty. I've heard that if you spend the big bucks ($400) you do alright, but I'm pretty sure I wouldn't trust even $190 ranged cards. Which is sad, because $190 for what the R9 270 does is ridiculous...
Hi! I make Firefox Plug-ins. Check 'em out @ https://addons.mozilla.org/en-US/firefox/addon/youtube-mp3-podcaster/
Despite the massive amount of bashing going back and forth here, I feel compelled to point out that I've swapped back and forth between both AMD/ATI and NVidia over the years and I've run into problems with brand new games having glitches with one or the other on both sides. Even having said that, I'm talking two or three times in over a decade. Aside from that I've had fans go out on one card, and it still lasted long enough after that that I didn't feel bad when it came time to buy a new one.
For most people it really doesn't matter what card you get as long as it isn't ancient. For enthusiasts, compare specs and get what you need. If the specs look like they're in Klingon to you, take the time to learn what's what. If you can't be arsed to do that, then you aren't an enthusiast in the first place.
This isn't like rooting for your home sports team. There is no justifiable reason to give complete loyalty to any company when weighing your purchases.
I'm a converted NVidia user that would only recommend AMD right now, but even I will readily admit that "Trueaudio" is stupid marketing crap that isnt even close to passing the bullshit test.
...)
There is a reason that hardware accelerated audio died 15 years ago. Its because CPUs from that era were so powerful that many-channel realtime software mixing wouldnt use even a fraction of a percent of CPU time.
The best way to keep this in perspective is to realize that the number of samples per second for audio is less than the number of pixels wolfenstein3d, released in 1992, rendered per frame. 20 year old hardware could do realtime many-channel CD quality audio software mixing, and in fact there were more than a few programs that did exactly that (scream tracker, impulse tracker,
Hardware audio acceleration is was on its last legs 20 years ago.
"His name was James Damore."
At least AMD didnt have several years where solder reflowing cards and laptops was standard troubleshooting practice.
-
The problem is when do you give them the benefit of the doubt?
I've used AMD cards on and off over the years for well over a decade and the problem has always been that each time I've heard someone say what you just did and tried them it's simply not been true.
Maybe you're right this time, but given how many times I've been bitten it's hard to have any faith in such statements.
My friend uses AMD and is always whining about problems with his cards, especially when it comes to Eyefinity stuff so I'm loathe to believe what you say is true even now, especially when we've also had numerous reports and news stories of frame dropping issues also.
FWIW I first came to realise there was a general and wide ranging issue with AMD drivers back in about 2003 when I was still doing some tech support on a network of over 5000 systems with a wide ranging set of graphics cards between Intel, AMD, and nVidia. That's a pretty large sample size and AMD graphics issues were a couple of orders of magnitude more prominent than any of the others so there has definitely historically been an issue. When I've always had more issues with AMD cards at home (i.e. when laptops I've bought have come with AMD cards) I've just found it difficult to lose that perception - my personal sample sizes haven't of course been as large, so maybe me and my friends have just been insanely unlucky in picking the odd cards with bad drivers but I'm more prone to believing that it's just that AMD has never really got to grips with it's drivers, even now.
But even if they have sorted their fundamental issues with their driver development regime, it's going to take quite some time for them to regain the trust of many people like me - at least a few years without any reports of widespread driver issues.
when they release a fully enabled GPL driver, i'll be happy to rip out my Nvidia card and buy AMD's card. in the meantime, i'll stick with my Nvidia card and the ever improving Nouveau driver.
Anons need not reply. Questions end with a question mark.
Nvidia drivers are shittier than AMD. End of story. AMD drivers implement the graphics API's to the letter.
Yeah, my personal experience is that that's completely bullshit you're slewing there. AMD drivers fail to implement graphics API's properly, and thus are more fragile when something unexpected happens, like a call to a deprecated OpenGL function, and when you point out to AMD that their driver is breaking, they point out that the function is deprecated, failing utterly to grasp the meaning of the term -- yes, new software is not supposed to call it anymore, but it's supposed to continue working anyway until it's dropped from the spec completely. "Your software is old" is not an excuse for violating the API.
"Convictions are more dangerous enemies of truth than lies."
TrueAudio is an interesting technology IMO. While graphics improved every year the sound is actually worse now than in the XP era. Im kinda optimistic that games will use this technology, since its pretty much the same chip as in PS4. If this was an AMD only technology then i would expect it to be used in as many titles as GPU physx is used in. Pretty much useless. But since its also in PS4 im looking forward to games with better sound.
If you compare the quality of 3d audio before vista and compare it with current games its obvious something went wrong. Sure we have a lot more CPU power but games dont use it for sound.
Yep. Those Nvidia drivers are really nice since they caused a Low Profile GT210 to overheat, forcing a replacement by them under warranty. Sure I'll continue using the replacement where it is but I've restricted the drivers to the 296 series as this isn't a gaming build so there's no need for all the god damn updates to benefit games.
What I'd like to see both AMD and Nvidia tell the game devs is "Fuck Off" and use the god damn windows DX as you're supposed to and bitch to Microsoft about crap performance from theirDirectX system.
If they both did that and worked on getting power consumption down while keeping the performance - I'd live to see an R9-290X with the same performance that runs strictly off bus power (75w max). It's doable but they need to take the fucking time to refine the chips and get their power demands down far enough to not require an external connector unless runing multiple cards.
Mod me up/Mod me down: I wont frown as I've no crown
I was just getting ready to say that AMD's drivers are terrible.
But I'm primarily a Linux user. Nvidia binary blob drivers are on-par with (better than?) Windows drivers, while AMD binary and open drivers are both 3-10 times slower in games.
As for Windows, all I know is that AMD seems to release game-breaking updates from time to time. Remember when Diablo III came out? Blizzard was warning everyone not to update their drivers.
Finally, while I think OpenCL is the future, CUDA is the only game in town for deployed GPU compute at the moment. And coming from a hardware-oriented background, CUDA is actually a pleasure to work with.
On Linux, NVIDIA is still way, way better. For newer cards. If you can use the proprietary driver at all.
Although Catalyst is subpar(2) compared to nvidia's proprietary driver, AMD's opensource driver is quite good and its performance has nicely catched up for all the previous generation hardware(1). Of course that comes from the fact that AMD has actively helping the development almost since the days of the ATI acquisition, releasing docs, source code, and even having developpers on their own payroll. Meanwhile Nvidia only recently started announcing that they will be open to answer specific questions to help developers (until then, they were completely incomunicando with Nouveau opensource project which was 100% reverse engineering).
Opensource drivers are so good, that AMD has phased out their older drivers and r300 is their official driver for Radeon X cards (Their are good, stable and fast enough and relying on them relieves AMD from the burden of maintaining an older branch of catalyst for the older drivers).
(1): the r600 drivers, for VLIWx architecture, up to HD 6x00 serie. The HD 7x00 and R9 2x0 use GCN architecture and are covered by RadeonSI, which is still lagging behind for now, though progress is happenning there too, and AMD is publishing docs and code.
(2): on the other hand, Catalyst tend to be a much better Linux citizen and play nicely with Linux standard and implement them early, instead of going the "Fuck, this, we'll just recompile the same stuff as under Windows and be damned if it doesn't work the same way under Linux" (cf. all the "missing" features like optimus, and the whole Linus' "Fuck you, Nvidia!" Finger-flip).
"Sufficiently advanced satire is indistinguishable from reality." - [Tips: 1DrYakQDKCQ6y52z6QbnkxHXAocMZJE61o ]
While I don't disagree with your post, and you made me more than a little nostalgic at the mention of Scream Tracker, I feel I should point out that CPUs in that era were running ONLY your code, and in a deterministic manner. It was comparatively easy to ensure your mixed sound made it to the DMA buffer before it emptied.
In this era of pre-emptive multi-tasking operating systems, unless you're running a realtime kernel there is no longer a guarantee that your multi-channel rendered audio will be ready in time before the DMA buffer starves. Multi-core CPUs have helped this a bit, so you can have a separate audio thread that stands a better chance of getting more CPU time if not its own core.
I have yet to see a program made in the last ten years run with as much real-time, stutter-free simultaneous sound and visuals as the likes of Scream Tracker on a 486.
Further, the A3d hardware spatial audio acceleration from Aureal (before they were bought out by Creative) was most impressive.
"Nine times out of ten, starting a fire is not the best way to solve the problem." - my wife