@DragonLord How much? Mine is Sager as well :D

Whoa 780M nice

Optimus sounds like Enduro

it is.

No :(

I seriously hope I can continue downloading after tomorrow

I bet Optimus is a better Enduro

@LewsTherin architecturally they're practically identical -- the only thing that could be different between them is performance, really

I think there was a bug with Enduro on HD 7970M. Seems like they fixed it with HD 8970M

and Enduro is generally seen as a good thing... the main gripe with Enduro is that, if you're using Enduro with an AMD APU, your CPU performance is going to be vastly inferior to what you could get with an Intel i7

I never choose AMD processors

it's pretty hard to find systems with AMD CPUs unless it's budget pricing, and then you probably aren't shopping for high performance anyway

Ah.. just finished BF 4. I couldn't get 60fps Ultra :( So disappointed. 16-40 fps

I would like to know how desktop users fare.

Tbh, if Nvidia GPUs were affordable I'd go for them.

haven't even bought BF4 yet, probably won't until it's very cheap... I loved the first Battlefield game, but ever since then, it seems they've focused more and more and more and more on the infantry combat, and infinitely less on vehicle combat

Well cough cough

I only buy games that are very good and for the multiplayer. But BF 4 for now has a lot of bugs.

When it gets fixed the will I

I much preferred ship vs ship or ship vs sub or aircraft dogfighting or bomber vs ground forces, than this "lol let's remake Call of Duty and call it Battlefield 4" thing

the only reason I'm even considering buying BF4 is to see how the Mantle technology works (or doesn't work)

since I have a GPU and driver capable of running it

GPUs supporting GCN architecture right?

Because the specs doesn't specify whether my card supports Mantle. It better because I want to test it to.

yeah... which means, not even any particular model (because they have been known to stick old VLIW5 stuff in HD7000s and even HD8000s at the low end) but GPUs with that specific architecture

APUs for the longest time were still VLIW5, but the very latest APUs are finally GCN... 8970M is definitely GCN (first iteration, which is like the same as 7970 but with minor bugfixes and clock tuning)

I imagine Mantle will probably work best with the second iteration GCN on the R9 290X and family

If it even works at all I will be happy

if the developers do a proper job, it should result in FPS increases... how much, I don't know

Yeah, that's the hope. But I also assume it depends on how much overhead already exists by using D3D or OpenGL

they're keeping tight-lipped about the status of Mantle support in BF4, but my gut feeling is that they probably got it mostly working to the point where it renders the game, but there are either rendering artifacts, or weird, unacceptable FPS drops, or stuttering, or stability issues, or actual performance regressions where it's better on D3D

they aren't using OpenGL at all right now, and the overhead in D3D is pretty damn high, so I would expect it to add 10 - 20 FPS to people who aren't yet getting 60 FPS

Lol, so you think Mantle isn't getting the desired effect?

cynically, knowing how these kinds of things go, that's probably what's happening -- they're putting the API to the test, and finding bugs in the driver, and having to find workarounds for performance issues

it's going to take a lot of back and forth between AMD and EA, and they'll probably do something silly like release a "Catalyst 14.1 BF4 Edition (Beta)" driver that has hotfix updates to support BF4's use of Mantle, and say something like "you MUST use this driver if you want to turn on Mantle support"

then we'll see it start to settle down after the next quarterly release of Catalyst contains not only those critical hotfixes but further improvements and it starts to stabilize

While there will be heavy collaboration between EA and AMD, I sure hope they manage to get Mantle independent of BF 4 or EA shizness.

I have no doubt that AMD rushed this Mantle API to production before it's really ready, just to get gamers excited about it and starting to demand it from game devs, which puts pressure on companies like EA to use it, but it's just not ready

Yeah, they should have had it ready first.

They wanted to sell their R9 cards I guess.

well, Mantle's API is structurally and architecturally very similar to the APIs used on the new Playstation and Xbox consoles, which both have GCN parts, and those are definitely stable APIs on those platforms because there are people playing production-quality games on them right now in their living rooms

but the "port" of the Console-esque API over to Windows and Linux seems to have been bungled

:S Wasn't Mantle written from scratch?

Why'll there be a port?

it's sort of both -- from a game dev's perspective, the function interfaces being used to write a Mantle engine on Windows are quite similar to the interfaces used to write an engine for a PS4 or Xbox One game

but under the hood, in terms of the implementation, the support in the driver for Mantle had to be re-done completely because of the need to implement Mantle in terms of the Windows HAL and kernel infrastructure

Argh. I hate ports. Reuse is a bane most of the time.

and a lot of the code they had built up to support Direct3D can't be re-used for Mantle

Well knowing AMD I'm sure they will fuck up initially

there are many, many layers to a graphics driver... at the lowest layers, you're interacting with Windows kernel APIs that aren't quite the same as PS4's low-level APIs or Xbox One's APIs, so that layer has to be redone... at the hardware interface level, you can actually re-use all of the code for interacting with the GCN GPU between consoles and Windows, and use your low-level abstraction layer as a platform-independent buffer, which is what they do

and then there's an even higher layer, still within the kernel driver, that exposes interfaces for the userland API that the game devs use; that layer is well-suited to D3D, but the one that existed on Windows before isn't well-suited to Mantle

so they tossed the lowest layer, kept the middle layer, and tossed the upper layer... lol

That sounds uber fucked

Do you think I can continue download L4D2 after tomorrow?

I don't want to leave my laptop on overnight

it's either that, or write a completely separate driver for every platform, which they don't have the resources to do; there are a few articles on reputable news sites from a year or two ago saying that AMD laid off 50% of their software development team (the driver devs)

@LewsTherin once it's added to your library, you have it -- you won't lose it even if you stop the download altogether

@allquixotic :O Wow that sucks. Are they back to hiring?

@allquixotic GTK thanks

you basically just have to click the Play Game button while logged in, and not get an error message, and you're done... L4D2 is permanently yours

Kwl.. no errors :D

I don't know if they're hiring again, but they completely closed their Starnberg, Germany driver development house, now they're all in Canada

Not even CS/IT is safe :(

well, they appointed a new idiot CEO when AMD was losing money (because of dumping tons of money into R&D for processors and watching them not sell because Intel is so much better) and his logic was "We're losing money! Let's stop spending so much!"

which is completely non-sensical for a company whose revenue comes from R&D

it wasn't because of the recession or anything like that. it was just that AMD's CPU sales tanked for a couple of years after the failure of the Phenom processors et al (historically, AMD has kept pace reasonably well with Intel, until 2008 or so), so they lost money, while still spending tons on research to make better CPUs

I see.

fortunately, the graphics division is a profit center for them, and always has been, because ATI graphics cards have always fared well against Nvidia, and in recent years I'd say they're downright superior for gaming

but the CEO wasn't intelligent enough to understand what an uber mistake it would be to reduce staff in the graphics division, which is their only chance of making profit at all

mainly because you can get comparable performance with an AMD GPU than an Nvidia GPU that costs $100+ more

Yeah their price performance GPU is what makes them floating.

I wouldn't they say they are superior yet though

if AMD hadn't purchased ATI when they did, I think the company AMD (independent of ATI, in a hypothetical world) would have gone bankrupt by now

meanwhile ATI, focused entirely on graphics, would probably be at least as good today as AMD's Graphics Product Group is doing

Is he still CEO of AMD?

they found a pretty cool market for their APU parts though, which consist of an inferior AMD CPU with a poor performance-per-watt factor, coupled with an extremely economical to produce GCN core... not only are these good for small HTPC boxes, but they are perfect for this year's consoles

@LewsTherin he is indeed

I just got a kick out of the idea of, what if INTEL had purchased ATI O_o!

they could let ATI fabricate their GPUs on Intel's 22nm process that they're using for their CPUs now, and Intel's processors would have a little GCN in them O_O

Same league as Nvidia I guess

But expensive

What like a better iGPU?

@LewsTherin yeah -- exactly... the coupling of Intel's superior CPUs, with Intel's superior fabrication technology, with ATI's superior graphics technology research, would yield basically the ultimate CPU

you could realistically have an Intel CPU that could game on its own, without needing a discrete card

and the CPU performance would still be as good as you expect

intel.com/content/www/us/en/high-performance-computing/…

serversdirect.com/Components/GPU_Cards/id-CO0004/…

this is a hint at what the future may hold

right now that's a server-only GPGPU part, but they could probably turn that into a gaming GPU without too much effort

Isn't Xeon what they are using in consoles?

I've seen it somewhere... vaguely

@LewsTherin no, and don't get confused by the branding... Xeon is a server-grade CPU brand usually, but "Xeon Phi" is a special brand all of its own

that Xeon Phi is basically a 57-core i7, except that each core is split into a bunch of small compute units, which makes for a perfect parallel processor, very much in the spirit of Nvidia's Tesla or AMD GCN

I wouldn't mind testing it out

it's only $1600 :P

Lol, "only" :L

Xeon Phi is Intel's first attempt at turning their hardware resources toward a truly massive parallel processor, but the work they did on Xeon Phi is clearly leading them in the direction of making an Intel discrete GPU that can compete with something like a GTX TITAN