Intel Abandons Discrete Graphics

Groan by Winckle · 2010-05-26 06:23 · Score: 3, Insightful

I hope they at least manage to incorporate some of what they've learnt into their integrated chips.

Intel's integrated chips have been appallingly bad in the past, some incapable of decoding HD video with reasonable performance. Manufacturers using those intel integrated chips in their consumer level computers did a great deal of harm to the computer games industry.

Re:Groan by Anonymous Coward · 2010-05-26 06:33 · Score: 0

You mean like Apple using the GMA950 and X3100 in their early intel Mac minis and MacBooks?
Starcraft II, Diablo 3, Steam on Mac OS X.... all great news except we can't use any of it because the intel integrated GPUs SUCK!
Re:Groan by Anonymous Coward · 2010-05-26 06:44 · Score: 0

Why not refer to something that's less than 5 years old? "Intel HD Graphics" on the I5 661's @ 900MHz. runs general games very well, let alone what MOST people need.
Re:Groan by Pojut · 2010-05-26 06:44 · Score: 5, Informative

For anyone stuck with an Intel GMA chipset: GMA Booster may help solve some of your problems. Just make sure you have a decent cooling solution, as it can ramp up the heat output of your system considerably. Still, if you're stuck with GMA, it can make the difference between a game being unplayable and being smooth.

--
Living With a Nerd
Re:Groan by toastar · 2010-05-26 06:49 · Score: 1

You mean like Apple using the GMA950 and X3100 in their early intel Mac minis and MacBooks?
Starcraft II, Diablo 3, Steam on Mac OS X.... all great news except we can't use any of it because the intel integrated GPUs SUCK!
Starcraft 2 chokes on my radeon 5430
Re:Groan by Anonymous Coward · 2010-05-26 07:08 · Score: 1, Insightful

That isn't saying much considering that the 5430 is an extremely low end GPU.
Re:Groan by Jeng · 2010-05-26 07:21 · Score: 1

Video cards for gaming.
Memory bandwidth within the video card is one of the best ways to grade a video card.
It doesn't matter much how much video Ram it has, its the speed.
Having a newer video card helps a little, but the big thing is the video memory bandwidth.
Video processing is secondary for gaming because the primary bottleneck is the memory bandwidth.
If you have a low end, but new video card it will perform comparably to the older version of the same card.
If you have an older premium card it will still perform better than a newer budget card primarily because of the memory bandwidth.

--
Don't know something? Look it up. Still don't know? Then ask.
Re:Groan by Anonymous Coward · 2010-05-26 07:34 · Score: 0

I don't know about that. It reminds of those "386to486 converter" programs that used to float around the old BBSs. Anything that claims to magically transforms hardware via software should be viewed with extreme skepticism.
Re:Groan by Pojut · 2010-05-26 07:41 · Score: 1

I've tried it, it definitely works. It "transforms" the hardware by greatly overclocking it :-)

--
Living With a Nerd
Re:Groan by FreonTrip · 2010-05-26 07:43 · Score: 1

True, but on the bright side you can now host a dedicated server on those systems without resorting to running Windows-in-a-box or rebooting.
Re:Groan by TheRaven64 · 2010-05-26 07:58 · Score: 1

Those 386to486 converter programs installed a hander for the illegal instruction trap, caught it, and emulated the dozen or so that were introduced with the 486. They let you run programs that required a 486 (rather than having them crash with an illegal instruction error), they just ran a lot slower than on a real 486.
This is completely different; it just tweaks the clock speed of the hardware. Given that most people who have integrated graphics and can't upgrade are laptop users, who have limited and non-expandable cooling, it's a terrible idea.

--
I am TheRaven on Soylent News
Re:Groan by Creepy · 2010-05-26 08:49 · Score: 1

That isn't really true - memory bandwidths have a fairly minor impact, similar to CPU/memory - maybe 5% of the speed of a graphics card is from latencies due to memory (either bus or strobe latencies). The rest of the time is spent doing transforms, running shaders, doing depth testing, etc, and those processes depend on the GPU clock. The old AGP shared memory model is actually creeping back in (look at the G100s, for example), especially in the mobile processing area because memory bandwidth matters so little. There once was a time when main memory to GPU memory was a serious throttle, but those days died around AGP 4x and haven't returned.
Memory itself is a pain to compare, as it may be your GDDR5 is faster clock-wise than GDDR4, but the latencies negate any speed gain, same as for processor memory. To further complicate things, depending on burst size and how many times it has to call the strobes (e.g. RAS to CAS latency), you may get mixed results.
Re:Groan by kimvette · 2010-05-26 11:20 · Score: 1

Back in the days of the 8086/88 and 80386 I used to run a math coprocessor emulator and it worked amazingly well, much to my surprise. I never ran 386to486 but it would not surprise me if it did a good job of running programs requiring a 486.

--
The Christian Right is Neither (Christian nor right). See: Matthew 23, Matthew 25, Ezekiel 16:48-50
Re:Groan by Anonymous Coward · 2010-05-26 12:03 · Score: 0

Except that the program was a hoax that promised to turn a 386DX-40 into a 486DX/2-66. Read the documentation that came with it for some real laughs.
I especially loved this bit of bullshit here

What this program does, is, it adds a mini TSR program into
a protected memory area and this RESIDENT program acts as a CPU,
it analyses the program being run and takes over the work, does
its own calculations, compresses the program in memory, changes
certain commands, all in realtime! All this frees up your
regular CPU. So your regular CPU does its chores and the
EMULATED CPU does its work too. It's like having a math co-processor,
but in this case it's a CPU co-processor.
Re:Groan by walshy007 · 2010-05-26 13:19 · Score: 1

The rest of the time is spent doing transforms, running shaders, doing depth testing, etc,
and what do you think those transforms are being done on? large arrays of numbers stored in gpu memory

There once was a time when main memory to GPU memory was a serious throttle, but those days died around AGP 4x and haven't returned.
Ah, this is where the misunderstanding would hvae been, he was likely talking about gpu to gpu memory bandwidth

but the latencies negate any speed gain, same as for processor memory.
The advantage most gpu's have is their memory access patterns are very linear, they already know what they'll be working on next before they're finished what they're doing, so while they're working on what they are they can chuck the next address on the bus so when they're done they have new data.
Re:Groan by Calinous · 2010-05-26 18:22 · Score: 1

This depends a lot by the graphic load - one has examples of linear performance increase with memory bandwidth, and in some cases the increase is hardly there. Some games benefit from more memory bandwidth, some don't (see http://www.firingsquad.com/hardware/nvidiageforce/page10.asp, going from GeForce 256 to GeForce 256 DDR doubles bandwidth but do nothing for performance in Quake2 or Quake3)
Re:Groan by yoyhed · 2010-05-26 21:00 · Score: 1

So you're saying the current generation, late-2009 to 2010 integrated Intel solutions run "general games" very well? Assuming "general games" means anything more than The Sims or Starcraft 1, I say: FINALLY! However, I don't think the GP was referring to the very newest Intel integrated graphics, but rather everything that's out there from them.

Every single Intel integrated graphics solution I've used up through the middle of 2009 blew ass (one I can specifically remember is the HD 4500, as it's in one of my newer laptops). It can only run Half-Life 1 at playable framerates in 800x600 or less (and only in Direct3D, OGL barely runs at all.) That game is TWELVE YEARS OLD. It ran perfectly on my GeForce 2 MX 400 that I bought for $50 nine years ago.

Granted, there have been HL1 engine updates since then (I'm testing the Steam version) but those updates have also updated it to work on modern hardware.

By contrast, all the low-end integrated Radeon and GeForce chipsets I've tried in the past few years have run HL just fine. Face it: Intel's integrated graphics suck. If they've finally gotten their shit together, good for them - doesn't change the fact that 95% of consumer laptops out there have their shitty GPUs in them.

--
WHO NEEDS SHIFT WHEN YOU HAVE CAPSLOCK/ DAMN1
Re:Groan by makomk · 2010-05-27 07:17 · Score: 1

Sadly, this idea hasn't gone away. Some older Intel and AMD processors - mostly Intel ones - can't run the 64-bit version of Flash because they don't support an instruction it requires. So there's a library you can get that traps the instruction and emulates it.

So Intels next cpu will the same suck video build by Joe+The+Dragon · 2010-05-26 06:23 · Score: 0, Troll

So Intel's next cpu will the same suck video build in?

AMD will kill them.

Intel is a great manufacturer.. not designer. by Anonymous Coward · 2010-05-26 06:28 · Score: 2, Interesting

They've never been able to bring the most innovative designs to market.. they bring 'good enough' wrapped in the x86 instruction set.

If x86 was available to all I think we'd see Intel regress to a foundry business model.

Re:Intel is a great manufacturer.. not designer. by gyrogeerloose · 2010-05-26 06:32 · Score: 1

They've never been able to bring the most innovative designs to market.. they bring 'good enough' wrapped in the x86 instruction set.
And, judging by what I've heard, a lot of people would say that the x86 instruction set itself is nothing more than 'good enough.'

--
This ain't rocket surgery.
Re:Intel is a great manufacturer.. not designer. by hedwards · 2010-05-26 07:40 · Score: 1

It's not so much 'good enough' as it brings along several decades worth of cruft that aren't really necessary in the modern era. While Intel had a bright idea in Itanium and ditching the x86 instruction set, they greatly underestimated the amount of effort that it would take to port the code and ensure that the necessary applications were available. Ultimately it was more or less DOA as a result, it just took some time for it to become formalized.
Re:Intel is a great manufacturer.. not designer. by Jackie_Chan_Fan · 2010-05-26 07:51 · Score: 2, Interesting

I disagree. Intel has been destroying AMD these past 4 years.
AMD's 64bit instruction set, and athlons were a huge improvement where Intel had failed...
But now.. Intel's chips are faster, and AMD has been playing catch up. For a while there AMD didnt have an answer for intel's core line of cpus.
Now they do, and they're slightly cheaper than intel but they do not perform as fast as intel.
Re:Intel is a great manufacturer.. not designer. by Jeng · 2010-05-26 08:38 · Score: 1

Intels chips are faster because Intel has much better production facilities.

--
Don't know something? Look it up. Still don't know? Then ask.
Re:Intel is a great manufacturer.. not designer. by Nadaka · 2010-05-26 09:01 · Score: 1

I think you may be a bit off here. AMD meets or exceeds the performance available from Intel chips at every point of the price curve except the very high end where they do not compete at all.
The i7 920 is the only real competitor to AMDs chips in price/performance, coming in a bit faster than the AMD 955/965 in both performance and cost.. Above that point, incrementally more power from Intel comes at exponentially higher costs. Below that point and AMDs chips beat everything Intel has at each price point.
Re:Intel is a great manufacturer.. not designer. by Jackie_Chan_Fan · 2010-05-26 09:22 · Score: 1

AMD does beat intel on the price curve... but not in performance. If you want the performance.. AMD has no answer for intel's cpus. I recently built a system for someone and looked at all of the cpu options. Ultimately I went with an AMD cpu for him because of his price range....
Like you said, AMD beats intel on the price... but not on the performance. If you want performance, you have to pay intels prices.
Thats why they cost more. The cpu's intel have put out these past 3 years are incredible. For a good time it looked like AMD had intel beat, but Intel really stepped up.. so much that it virtually squashed all of the good will AMD had earned with Athlon
AMD's still in the game because of their prices. They're good cpus, but intel's performance is still better.
Re:Intel is a great manufacturer.. not designer. by Tumbleweed · 2010-05-26 09:38 · Score: 1

I think you may be a bit off here. AMD meets or exceeds the performance available from Intel chips at every point of the price curve except the very high end where they do not compete at all.
Performance/price != Performance
And of course, it also depends mightily on exactly WHICH performance characteristics you're talking about.
Re:Intel is a great manufacturer.. not designer. by adolf · 2010-05-26 09:45 · Score: 1

It doesn't matter what's at the top.
If performance were the only metric one needed when selecting a product, we'd all be driving Bugatti Veyrons when we wanted to go fast, Unimogs when we want to move lots of stuff slow, and AMG Mercedes-Benz SUVs when we want to move stuff along with people.
Over here in reality, though, price is a factor. And so, Toyotas, the Hyundais, and the Chevys are a much better deal for most folks.
So, even if Bugatti made a more inexpensive and practical vehicle that I might be interested in, the fact that they also may produce the Veyron does not influence my buying decisions when comparing their offerings to those of more serene brands like Toyota.
Likewise with CPUs: I don't care what manufacturer makes the fastest chip. If the features that I'm interested in are close enough to the same, then I only I care about how much performance I can get for the amount that I'm willing to spend, and sometimes with an eye toward the power consumption.
To that end, AMD generally wins. That AMD does not have a metaphorical Veyron in their lineup does not mean that they do not produce chips which are cheap and fast for the entire gamut of typical home applications.

--
Kid-proof tablet..
Re:Intel is a great manufacturer.. not designer. by Nadaka · 2010-05-26 09:47 · Score: 1

I agree that for everything from the i7 920 and up, intel is unquestionable faster. Even the new 6 core AMD chips will be able to match/beat the i7 920 at a some tasks despite similar system costs.
The AMD 955 at ~$160 outperforms the intel E8400, Q8200 and Q8400 and i5-650 available in the range of ~150 to ~190. The same goes for just about every lower price point as well.
I think that the larger section of the market lies in the low to mid range chips. I am not just talking about price, but value as well. In that range, AMD certainly has better value by having higher performance per dollar spent for just about any price point in that range.
Re:Intel is a great manufacturer.. not designer. by RatherBeAnonymous · 2010-05-26 10:03 · Score: 1

IANAEE (I am not an electrical engineer), but I've always wondered why Intel doesn't expose their micro-op architecture to programmers. It seems like compilers could do a lot of optimization ahead of time, much like they can with Itanium, if they could bypass the x86 instruction set. It would also let programmers transition slowly, rather than jumping in with both feet.
Re:Intel is a great manufacturer.. not designer. by JamesP · 2010-05-26 10:27 · Score: 1

Well, not really
The P6 was a really good architecture. it's what AMD battled with the K7 arch (really good as well)
Of course, that, until Intel shot itself in the foot with the Netburst architecture (AKA Pentium 4)
A 1GHz P3 could run circles around the 1.4GHz, 1.6GHz even higher clocked Willamette P4
But the P6 arch carried on and Core 2 is based on it (with a lot of improvements on top)

--
how long until /. fixes commenting on Chrome?
Re:Intel is a great manufacturer.. not designer. by Rockoon · 2010-05-26 10:49 · Score: 1

Because it pretty much already is. Yes there are plenty of instructions that generate many uops, but there are also plenty of single uop instructions. I am more familiar with AMD's architecture than Intels, and there I cant think of a single vector path instruction that cant also be described by fast path instructions.

The many-uop model isnt a bad one, either. It makes sense that if there are common pairs of uops, that shorthand be used to encode them. This improve code density, among other things.

While there are many uncommon pairs and triplets of uops that have shorthands, their existence only has an very minor effect on efficiency (just don't use them.) This is why RISC doesnt beat CISC.

Is the x86 instruction set optimal? Nope. In spite of it not being optimal, Intel mass produces the fastest CPU's on the planet.

--
"His name was James Damore."
Re:Intel is a great manufacturer.. not designer. by Rockoon · 2010-05-26 11:12 · Score: 2, Interesting

AMD does beat intel on the price curve... but not in performance.

AMD does seem to have an edge in the multiprocessor arena, although I am not sure why.

According to PassMark, the fastest machines clocked using their software is a 4 x Opteron 6168 (4 x 12 cores = 48 cores) system and a 8 x Opteron 8435 (8 x 6 cores = 48 cores)

The actual numbers are:

4 x Opteron 6168 : 23,784 Passmarks.
8 x Opteron 8435 : 22,745 Passmarks.
4 x Xeon X7460 : 18,304 Passmarks.
2 x Xeon X5680 : 17,910 Passmarks.

That $200 AMD chip that everyone is raving about, the Phenom II 1055T, scores 5,661 Passmarks. If AMD keeps that up, Intel might be in some trouble soon even in the high end market unless Intel can cut prices dramatically. Intel doesnt offer anything comparable for the money.

--
"His name was James Damore."
Re:Intel is a great manufacturer.. not designer. by Anonymous Coward · 2010-05-26 11:45 · Score: 0

This issue is in that Intel has smaller transistor technology not in their architecture and design. If you were to build an AMD chip on 32 nm or even 45 nm it would be more comparable.
Re:Intel is a great manufacturer.. not designer. by kramulous · 2010-05-26 14:56 · Score: 1

As somebody whose sole job is to squeeze maximum floating point performance out of Intel chips, I can tell you those benchmarks are absolute crap.
How was the code for the benchmark written? Did it use the compiler that intel puts out? Does it use ippxMalloc to create the datastructures for the number crunching? If the answer to any of the last two questions is no, then you are not getting even slightly close to the full throughput of the chips.

--
.
Re:Intel is a great manufacturer.. not designer. by Rockoon · 2010-05-26 15:47 · Score: 1

How was the code for the benchmark written?
Its fucking Passmark. Are you new to the benchmarking scene?

Did it use the compiler that intel puts out?
The one that intentionally outputs code that cripples performance on non-Intels? Are you new to the benchmarking scene? Yes, their compiler is great.. as long as you trick it into not crippling performance on non-Intels.

Does it use ippxMalloc to create the datastructures for the number crunching?
Its fucking Passmark. Are you a dipshit or something?

If the answer to any of the last two questions is no, then you are not getting even slightly close to the full throughput of the chips.
First you say the results are crap, and then later you say its only crap if the answer to blah blah blah is no? Make up your mind, dipshit.

--
"His name was James Damore."
Re:Intel is a great manufacturer.. not designer. by kramulous · 2010-05-26 15:58 · Score: 1

Hello troll.

Yes, their compiler is great.. as long as you trick it into not crippling performance on non-Intels.
Well you could use the compiler that AMD puts out. I'm sure they have one that is specific to their architectures.

Its fucking Passmark.
Doesn't mean anything unless you can pop the hood and examine the code. To get the most out of recent chips, the coding style/methodologies have changed substantially. There are whole things you just shouldn't do now. You code to the compiler.
I generally write micro-benchmarks to test even simple things (std::max, if greater than do this otherwise do that, etc) and the difference in the number and what instructions are used varies a lot. Then I put them together. So, no, I'm not new to the benchmarking 'scene'.
I have a feeling I'll regret responding to a troll.

--
.
Re:Intel is a great manufacturer.. not designer. by Calinous · 2010-05-26 18:58 · Score: 1

The same situation happened when AMD had the performance crown - they had the higher prices, and Intel competed on price. However, AMD didn't had enough production capacity to really profit from those times, and now they're again selling cheap.
Re:Intel is a great manufacturer.. not designer. by Calinous · 2010-05-26 19:01 · Score: 1

AMD has the performance lead in server market, but only barely and only in some very specific benchmarks. As for the rest, since they integrated the memory controller, Intel is king.
Re:Intel is a great manufacturer.. not designer. by Calinous · 2010-05-26 19:03 · Score: 1

"Its ****** Passmark"
What the ******** is Passmark?
Re:Intel is a great manufacturer.. not designer. by Rockoon · 2010-05-27 01:22 · Score: 1

I generally write micro-benchmarks to test even simple things (std::max, if greater than do this otherwise do that, etc)

This is called benchmarking the compiler, dipshit.

--
"His name was James Damore."
Re:Intel is a great manufacturer.. not designer. by kramulous · 2010-05-27 08:50 · Score: 1

You're a fucking idiot and have no idea about high performance computing. Stick to your corner computer shop job you monkey.
Other compilers have no idea about the new instructions fucktard. What is there to benchmark when only one can do it. Get an education and stop peddling shit.

--
.
Re:Intel is a great manufacturer.. not designer. by Rockoon · 2010-05-27 09:27 · Score: 1

You're a fucking idiot and have no idea about high performance computing
...says the dipshit that benchmarks by changing multiple variables at the same time, and also isolation-benchmarks single functions.

Here is a fucking clue.

When you benchmarks CPU's, you dont change the software. You use the same software on all of them, or else the measurement is bullshit.

And when you benchmark a compiler, you use real world code and don't change the processor. This std::max measurement bullshit is just that, bullshit. You don't know dick about its performance until its used in practice, surrounded on one side by the actual code that feeds it, and on the other by the actual code that needs the result. You are like a VB programmer that learned that you have some impact on overall performance of the program.. but unlike a real programmer, you still havent figured out that synthetic tests are useless.

--
"His name was James Damore."
Re:Intel is a great manufacturer.. not designer. by kramulous · 2010-05-27 10:47 · Score: 1

You're still missing the point. You said [Insert Random Benchmark Test Here] found AMD was better than Intel Xeons.
When I have a single scientific program that runs absolute(a) 1x10^15 times (as well as thousands of other operations), I need to make sure that operation is done in the best possible way for the cpu's I have available.
eg
fabs(a) - the gnu math.h call is always resolved to a library call by the intel compiler - it assumes you want that particular implementation.
if (a 0)
a *= -1.0f
The compiler detected this as an absolute operation that does not use the math.h call and hence uses the hardware instruction instead. Because the X5500 has 256bit wide registers and provided my arrays are such that each next item in the array is *exactly* one unit stride from the next in L2 cache, I can pack the register and perform 8 absolute operations per cycle, per core. This is at least 64 times more throughput than the math.h call. I then move onto the next operation that can be 'fixed'.
max, min, sin,cos, log, other exponentiation, and hundreds of others are the same. Hence the micro benchmarks. Generating the assembly, flipping through the cpu engineering specification and making sure that the minimum number and appropriate instructions are being used. You then start putting them together to solve the big nonlinear problems.
That is how you write extremely efficient code.
If your beloved benchmark doesn't provide or advertise a vectorization report then they suck at benchmarking. They are not taking advantage of the architectures of the cpus. The same old shitty code only takes advantage of the architectures the programmer knew how to write for at the time. You absolutely must change the code to do correct benchmarking.

--
.
Re:Intel is a great manufacturer.. not designer. by Rockoon · 2010-05-27 11:53 · Score: 1

You're still missing the point. You said [Insert Random Benchmark Test Here] found AMD was better than Intel Xeons.
No, I didn't say that at all.

a) I cited a specific benchmark. You havent cited any.
b) The benchmark was not "AMD" vs "Xeon" .. it was specifically "4 x 12 CORE Opteron Systems" and "8 x 6 CORE Opteron System" vs "4 x 6 CORE Xeon Systems" and "2 x 12 THREAD Xeon Systems"

See that? Its very specific, and shows what the fuck you know. Its 48 core systems vs 24 core/thread system. You are such a dumbfuck.

You even called the benchmark "crap."

You are a fucking moron. If Intel is such a performance leader, where does Intels 48 core/thread solution land on any metric? Thats right, they dont even fucking have one other than their "single-chip cloud computer" which they only demo'd. You cant actually own one.

I can pack the register and perform 8 absolute operations per cycle, per core.
How quaint. You think that other people dont know what SIMD is. Heres an idea.. stop using Intel's auto-vectorization.. then clue up that it is not rare knowledge that masking off the high bit of a float removes the sign. If you were worth your shit, you would have masked the bit off to begin with. What a fucking tard you are writing inefficient code and then praising yourself for optimizing it.

--
"His name was James Damore."
Re:Intel is a great manufacturer.. not designer. by Anonymous Coward · 2010-05-28 16:27 · Score: 0

Are you 12?

missed milestones by LinuxIsGarbage · 2010-05-26 06:30 · Score: 4, Informative

' He added that Intel had 'missed some key product milestones' in the development of the discrete Larrabee product,

Like proof that they were even capable of making an integrated graphics product that wasn't a pile of garbage?

GMA910: Couldn't run WDDM, thus couldn't run Aero, central to the "Vista capable" Lawsuits

GMA500: decent hardware, crappy drivers under Windows, virtually non-existant Linux drivers, worse performance than GMA950 in Netbooks.

Pressure to lockout competing video chipsets. We're lucky ION saw the light of day. http://www.pcgameshardware.com/aid,680035/Nvidia-versus-Intel-Nvidia-files-lawsuit-against-Intel/News/

Re:missed milestones by KillShill · 2010-05-26 07:41 · Score: 1

So the answer is to buy more Intel cpu's. We wouldn't want them to die and that 3 time convicted monopolist AMD be left alone to dominate...
oh wait...
"I run Linux on my shiny new i7 cpu... stick it to Microsoft..."
wait again...
Support those who support compeition, not monopolists who destroy it.

--
Science : Proprietary , Knowledge : Open Source
Re:missed milestones by soppsa · 2010-05-28 23:57 · Score: 1

Support those who support compeition, not monopolists who destroy it.
Don't be ignorant. If AMD were in Intel's shoes, they'd act like a monopolist too. The purpose of any company responsible to shareholders is to maximize profits. Give dear Google some time... Don't blindly put faith in the underdog as 'Good Guys', sure they won't be Apple Evil but they'll come around..

--
Ogre Wedding Planners llc.

Re:So Intels next cpu will the same suck video bui by Anonymous Coward · 2010-05-26 06:31 · Score: 0

ignoring you're complete inability to form a sentence, this article is about their discrete graphics line (like it says in the fucking title), not their integrated graphics. So while their next chip may or may not have "the same suck video" it's completely irrelevant to this conversation.

Mod parent "Likely." by Spazntwich · 2010-05-26 06:33 · Score: 4, Informative

Short of buying out Nvidia I don't see Intel having a consumer's chance in America of competing with AMD in the value sector for the next few generations of chips.

CPUs have been "fast enough" for years, but GPUs have not. AMD is going to laugh all the way to the bank being able to offer a $50 package that can run The Sims.

Re:Mod parent "Likely." by TheRaven64 · 2010-05-26 06:40 · Score: 4, Insightful

CPUs have been "fast enough" for years, but GPUs have not.
Really? I think you might want to take a look at what most people use their GPUs for. Unless you are a gamer, or want to watch 1080p H.264 on a slightly older CPU, a 4-5 generation old GPU is more than adequate. My current laptop is 3.5 years old, and I can't remember ever doing anything on it that the GPU couldn't handle. As long as you've got decent compositing speed and pixel shaders for a few GUI effects, pretty much any GPU from the last few years is fast enough for a typical user.

--
I am TheRaven on Soylent News
Re:Mod parent "Likely." by Anonymous Coward · 2010-05-26 06:42 · Score: 0

They wouldn't need to buy them in order to create an integrated offering. Heck, they're already doing it with the Atom/ION product line. If AMD makes headway with all-in-one solutions, it will be damaging to both Intel and Nvidia and I'd put money on them teaming up to compete.
Re:Mod parent "Likely." by Korin43 · 2010-05-26 06:55 · Score: 4, Informative

As long as you've got decent compositing speed and pixel shaders for a few GUI effects, pretty much any ATI or nVidia GPU from the last few years is fast enough for a typical user.
Fixed that for you. Intel cards are fine for "normal" computer usage, but they still suck pretty bad at most games.
Re:Mod parent "Likely." by LinuxIsGarbage · 2010-05-26 06:56 · Score: 1

Indeed. GMA950 was crappy during it's prime compared to ATI and nVidia shared memory bargain bin offerings 4 years ago, however it's still being shipped in N270/N280 netbooks, and is capable of running Aero fine. Probably H.264 decoding might be (or will be) beneficial to casual users than gaming performance, and I believe many of Intel's mainstream desktop and notebook GPUs provide it.
Re:Mod parent "Likely." by tepples · 2010-05-26 06:56 · Score: 1

Unless you are a gamer, or want to watch 1080p H.264 on a slightly older CPU, a 4-5 generation old GPU is more than adequate.
But if you're a game developer, you have an interest in members of the public having PCs with more powerful GPUs, not the Voodoo3-equivalent without even hardware T&L that is a GMA 950.
Re:Mod parent "Likely." by Anonymous Coward · 2010-05-26 07:07 · Score: 0

The Voodoo 3 was bitching for its time. The GMA 950, however was not. Not to mention that the Voodoo 3 can play Doom 3 (well it's a Voodoo 2, but can a GMA950 play Doom 3?). And it doesn't require any "flashlight mod".
Re:Mod parent "Likely." by LinuxIsGarbage · 2010-05-26 07:27 · Score: 1

Intel resisted letting ION be bundled with Atom chips. They preferred to continue releasing anemic GMA950. The CPU wasn't capable of rendering HD, it couldn't be offloaded to the GPU, and they wanted to keep it that way.
Re:Mod parent "Likely." by LinuxIsGarbage · 2010-05-26 07:30 · Score: 1

Also, this sounds like Intel wants to team up with nVidia. nVidia manages to put out something better than Intel, so Intel pulls their licence. http://arstechnica.com/hardware/news/2009/03/nvidia-countersues-intelright-on-schedule.ars
Re:Mod parent "Likely." by Spazntwich · 2010-05-26 07:31 · Score: 1

All I really meant by that is incremental improvements in GPU performance hold significantly greater "value" than commensurate increases in CPU speed. A bottom barrel consumer computer is going to be able to handle anything a common consumer is going to throw at it except gaming. Even the lightest gaming is about impossible on the most common consumer-class (Intel) GPUs, and unfortunately Intel is still the graphics chip in the vast majority of consumer computers.
Now bottom-barrel will still mean "game capable."
Intel's and AMD's offerings at the same price point will likely offer a 10%-20% discrepancy in overall CPU performance that will favor Intel but nobody will ever notice, while at the same price point AMD will be able to offer graphical performance eclipsing Intel's by several fold.
AMD is going to eat Intel's lunch, and this will be good news for all of us as the continuously floating "average" performance developers can target increases, but I digress.
Re:Mod parent "Likely." by Jackie_Chan_Fan · 2010-05-26 07:49 · Score: 1

Cpu's have never been fast enough :) My Quad Core could be a dual 8 core.. and I'm still at the mercy of the CPU while rendering.
Anyone doing music production at home, or professional, or 3d graphics at home or professional... or photoshop work at home or.. well you get it.
CPU is and will always be a factor. That son of a bitch is never fast enough for me :)
Re:Mod parent "Likely." by ElectricTurtle · 2010-05-26 08:00 · Score: 1

Intel and nVidia are pretty openly hostile to each other. (Though as represented by Jen-Hsun Huang, nVidia is hostile to just about everybody. nVidia was actually AMD's first choice for their merger to create the Fusion platform, but Jen-Hsun Huang's ego was too huge to compromise and AMD was a little short on the capital necessary at the time to be in a strong bargaining position.) However with Larrabee in the trash heap of history, Intel needs nVidia now more than ever, but unless Intel really comes groveling back to Jen-Hsun with a killer sweetheart deal of some kind, I can't see any kind of productive partnership between the two companies in time to compete with AMD's Fusion.

As painful as the AMD/ATI merger was, I do think that Fusion is poised to be a game changing architecture, just as important if not significantly moreso than the integration of floating point processing and memory controllers into CPU dies.

--
I support the Slashcott and will not be reading or commenting from 2/10/14 to 2/17/14. Beta is steaming pile of dog shit
Re:Mod parent "Likely." by tepples · 2010-05-26 08:08 · Score: 1

The Voodoo 3 was bitching for its time.
And its time was a decade ago.
Re:Mod parent "Likely." by FreonTrip · 2010-05-26 08:31 · Score: 1

*sigh* All right, I've got karma to burn.
The Voodoo3's limitations weren't trivial even for its time: 256x256 texture dimensions, forced 16-bit color, no real stencil buffer support, framebuffer size maxed out at 16 MB... A former 3dfx employee literally told me that it was a die-shrunk, bug-fixed Voodoo Banshee with an extra TMU popped onto its single lonely rendering pipeline. I owned one, and liked it tremendously, but it's based on technology first debuted 13 years ago.
As for Doom 3: yes, a MesaGL --> Glide wrapper exists that provides basic rendering functionality for Doom 3 and tricks the engine into sending its lighting straight to /dev/null/*. There are other wrappers that perform similar tricks for any vendor's GL drivers, and they will let you run on a GMA950 substantially better than the Voodoo3 could manage. And for whatever meager stakes we're playing for, a GMA950 can actually run Doom 3 with lighting when backed up by a dual-core CPU.++
* Metaphorically speaking.
++ Flexibility aside, software vertex shaders suck.
Re:Mod parent "Likely." by Anonymous Coward · 2010-05-26 09:56 · Score: 0

When I bought my Voodoo 3 3000 I was upgrading from a pair of Voodoo 2 cards in SLI. The performance increase was close to nothing, it just consolidated two cards into one. Since the Voodoo 3 also didn't support SLI, I felt a bit ripped off. At the same time, the PC I had at work had a Riva TNT2 Ultra that ran circles around the Voodoo 3 and supported 32-bit colour.
So no, the Voodoo 3 was NOT bitching for its time. It was when the cracks in 3dfx's armour started becoming visible.
Re:Mod parent "Likely." by AHuxley · 2010-05-26 12:07 · Score: 1

http://arstechnica.com/apple/news/2010/05/new-intel-ulv-processors-still-a-bad-fit-for-macbook-air.ars
"Because Intel is still battling NVIDIA in court over whether it has the necessary license to make chipsets for Intel's latest processors,
Apple can't pair these new Core i5 processors with the new NVIDIA 320M used in the new 13" MacBook Pro and white MacBook."
AMD is back by default :)

--
Domestic spying is now "Benign Information Gathering"
Re:Mod parent "Likely." by Coppit · 2010-05-26 16:58 · Score: 1

Your perspective is a little short-sighted... I remember when CPUs were barely able to play MP3s. Have you heard that CS5 will be GPU accelerated?
(Disclaimer: I work for NVIDIA)
Re:Mod parent "Likely." by TheRaven64 · 2010-05-26 22:20 · Score: 1

Doom 3 and tricks the engine into sending its lighting straight to /dev/null/
I was under the impression that this was the default behaviour in Doom 3.

--
I am TheRaven on Soylent News
Re:Mod parent "Likely." by TheRaven64 · 2010-05-26 22:29 · Score: 1

And what percentage of users run something like CS5? If you work for nVidia, then you should remember why nVidia was founded - because SGI refused to accept that their high-performance niche was about to become a commodity market and that the consumer-grade hardware was enough for a large chunk of their market.
I'm not saying that an old card will always be fast enough for everyone (by the way, CS5 is only using pixel shaders, and these run more okay on even Intel's last generation - if all you're doing is applying filters to still images, they're much faster than the CPU, and you don't care if you only get one frame a second). I'm saying that integrated graphics are fast enough for most people, and the speed of integrated graphics is increasing faster than the requirements of most people.
Remember when x86 chips didn't come with an FPU? Remember how you had several options for an external FPU? Now, how many x87 vendors are left? None. If I were in nVidia management, I would be focussing all of my attention on Tegra, and licensing the GPU design to companies like TI for use in their SoCs.

--
I am TheRaven on Soylent News
Re:Mod parent "Likely." by drsmithy · 2010-05-27 05:57 · Score: 1

CPUs have been "fast enough" for years, but GPUs have not.
GPUs are already quite "fast enough" for the majority of users, and have been for years.
What is it you think the average computer using is doing that they coudl be doing faster with a better GPU ?
Re:Mod parent "Likely." by soppsa · 2010-05-28 23:43 · Score: 1

Except that your 'normal' computer user just wants to play Farmville and the like, for which the Intel GMA is fine, these are all CPU bound games.

--
Ogre Wedding Planners llc.
Re:Mod parent "Likely." by Coppit · 2010-05-29 13:38 · Score: 1

The big deal for CS5 is apparently Premiere, which runs like a slideshow without GPU acceleration.
You seem to be suggesting that GPUs will become commodities and integrated into the CPU. For a lot of people that is true... I'm sure that there are plenty of laptop users for whom Aero is the toughest graphics they'll ever require. And I agree that trying to hold on to a market while there is a race to the commodity bottom is stupid.
But the discrete GPU is far from being subsumed by the CPU. See the Larrabee failure. Besides it being really hard to do video and 3D efficiently using a general-purpose CPU, there are tons of patents covering the tricky bits of making it go fast.
Re:Mod parent "Likely." by Spazntwich · 2010-05-30 14:24 · Score: 1

Your sig is all I need to read to know you're a complete faggot lol.
Slow Down Cowboy!
Slashdot requires you to wait longer between hitting 'reply' and submitting a comment.
It's been 15 seconds since you hit 'reply'.
Chances are, you're behind a firewall or proxy, or clicked the Back button to accidentally reuse a form. Please try again. If the problem persists, and all other options have been tried, contact the site administrator.

Wait, what? This is news? by jtownatpunk.net · 2010-05-26 06:35 · Score: 1, Insightful

A company that hasn't produced a discrete graphics card in over a decade (I'm pretty sure I remember seeing an Intel graphics card once. Back in the 90s.) is going to continue to not produce discrete graphics cards. Wow. Stop the presses. Has Ric Romero been alerted?

I wonder by halestock · 2010-05-26 06:35 · Score: 1

Does this mean that they'll be focusing on continuous graphics instead?

Re:I wonder by Anonymous Coward · 2010-05-26 06:48 · Score: 1, Insightful

Does this mean that they'll be focusing on continuous graphics instead?
More directly, what the hell is "discrete graphics"? I've been working with computers since the TRS-80 days and this is the first time I've seen the term. The writeup makes it sound like "discrete graphics" is some revolutionary kind of new display method that will make our old idea of viewing "pixels" on a "screen" obsolete, but the Google tells me that "discrete graphics" just means a video card as opposed to an onboard chip. Call it a video card! (Or more precisely, the chip that would run on a video card as opposed to being integrated into the motherboard.)
Re:I wonder by Anonymous Coward · 2010-05-26 07:02 · Score: 0

Hillarious. I applaud you.
Re:I wonder by TheRaven64 · 2010-05-26 07:12 · Score: 3, Informative

It's not a new term, and it's not unique to GPUs. The distinction between integrated and discrete coprocessors has been around for at least 25 years. If you read something like Byte from the early '90s, you will find discussions about the relative merits of integrated and discrete FPUs. You'll find a similar discussion on integrated and discrete MMUs and various other components if you look a few years earlier.

--
I am TheRaven on Soylent News
Re:I wonder by jandrese · 2010-05-26 08:32 · Score: 1

Discrete Graphics are the opposite of Integrated Graphics. Generally it refers to using a PCIe card to do the graphics instead of something built into your Northbridge (or whatever Intel calls their Northbridge now) or the CPU. Just as Integrated Graphics are synonymous with "crap", Discrete Graphics generally imply that you've got something at least somewhat capable under the hood.

--

I read the internet for the articles.
Re:I wonder by bill_mcgonigle · 2010-05-26 08:37 · Score: 2, Funny

If you read something like Byte from the early '90s, you will find discussions about the relative merits of integrated and discrete FPUs.
yikes, my memory of installing an 80387 has been completely un-accessed for at least a decade. Thanks for the scrub. :)

--
My God, it's Full of Source!
OUTSIDE_IP=$(dig +short my.ip @outsideip.net)
Re:I wonder by networkBoy · 2010-05-26 09:34 · Score: 1

and 80287 and 8087
meh
I miss those days some times.
then I use my computer to do something and I cease missing them.

--
whois gawk date unzip strip find touch finger mount join nice man top fsck grep eject more yes exit umount sleep dump
Re:I wonder by petermgreen · 2010-05-26 10:54 · Score: 1

Afaict graphics for PCs come into two main categories*. Discrete graphics chips have their own memory and communicate with the rest of the system over standard busses (PCI AGP or PCIe depending on age). Integrated graphics are integrated with some part of the chipset and don't usually have any dedicated ram.
I think the term originated in the laptop market. Many laptops have discrete graphics soloutions soldered to the motherboard.
*There are graphics setups that don't fall nicely into either category. Servers often have graphics integrated with the out of band management chipset, so they aren't really discrete (they are integrated with something) but they aren't integrated with the main chipset either.

--
note: i'm known as plugwash most places but i screwd up registering that here somehow in the past and now can't register

Re:So Intels next cpu will the same suck video bui by Mr.+Underbridge · 2010-05-26 06:36 · Score: 2, Funny

ignoring you're complete inability to form a sentence

Hey everybody, 'tard fight! Come watch!

Both good and bad by TheRealQuestor · 2010-05-26 06:37 · Score: 3, Interesting

This is bad news for one reason. Competition. There are only 2 major players in discreet graphics right now and that is horrible for the consumer. Now the good. Intel SUCKS at making gpus. I mean seriously. So either way Intel has no hope of making a 120 core GPU based off of x86 being cheap or fast enough to compete. Go big or stay at home. Intel stay at home.

Re:Both good and bad by keeboo · 2010-05-26 06:47 · Score: 1

This is bad news for one reason. Competition. There are only 2 major players in discreet graphics right now and that is horrible for the consumer.
What about VIA and Matrox?
Re:Both good and bad by h4rr4r · 2010-05-26 07:04 · Score: 1

They produce joke cards.
Re:Both good and bad by ElectricTurtle · 2010-05-26 07:06 · Score: 1

He said "major players". VIA's IGPs are generations behind and only end up in 'budget' computers or embedded appliances. Matrox serves a niche market primarily focused on professional workstation rendering. Neither competes head to head with nVidia or AMD/ATI.

--
I support the Slashcott and will not be reading or commenting from 2/10/14 to 2/17/14. Beta is steaming pile of dog shit
Re:Both good and bad by Anonymous Coward · 2010-05-26 07:15 · Score: 1, Interesting

What about them?
VIA's chips suck, like everything else they put out (e.g. their ridiculous re-badged Cyrix CPUs and stability-challenged motherboard chipsets).
Matrox is a niche player (multi-monitor etc.). Their performance is on level with Intel's GMA, and their prices are a lot higher than "free" which is what a GMA core basically costs when buying a modern Intel CPU or chipset.
Larrabee was indeed the only serious contender in discrete GFX we've seen for the better part of a decade or so, but it seems the Larrabee project management over-promised and under-delivered.
Intel corp. basically chose to scrap the project rather than suffer the embarrassment of a card that would probably have been one or two generations slower than NVIDIA and AMD's enthusiast parts, while using more power.
It would have been kick-ass for specialty graphics (truly fast vector graphics, voxel rendering, anything that doesn't conform to the standard vertex+effects shader pipeline of todays 3D graphics) and GPGPU though.
Shaders are a horrible kludge compared to just having a bunch of real CPUs with real memory and just running your algorithm like on a normal cluster.
Re:Both good and bad by FreonTrip · 2010-05-26 07:36 · Score: 3, Informative

VIA stopped designing motherboards for AMD and Intel CPUs about two years ago. Consequently, you can't find its GPUs in many places aside from embedded systems or ultra low-budget netbooks and the like. Weirdly they still sell a miniscule number of discrete cards, primarily overseas, but without divine intervention they'll never become a serious player again.
Matrox serves niche markets, mostly in the way of professional workstations, medical imaging equipment, and the odd sale of their TripleHead system to the ever-eroding hardcore PC gamer market.
In case anyone wonders what happened to the others: Oak Technologies' graphics division was acquired by ATI many moons ago; Rendition was eaten by Micron in 1998 and their name is now used to sell budget RAM; SiS bought Trident's graphics division, spun off their graphical company as XGI Technologies, had a series of disastrous product releases, and had their foundries bought by ATI, who let them continue to sell their unremarkable products to eager low-bidders; and 3dfx was mismanaged into oblivion a decade ago.
Re:Both good and bad by Jackie_Chan_Fan · 2010-05-26 07:56 · Score: 1

Nvidia has been incredible for the consumer for a long time now.
I would like to see their quadro products come down in price though. They are ridiculously overpriced.
Re:Both good and bad by Jeng · 2010-05-26 08:23 · Score: 1

and 3dfx was mismanaged into oblivion a decade ago
And Nvidia picked up the pieces from 3dfx.

--
Don't know something? Look it up. Still don't know? Then ask.
Re:Both good and bad by FreonTrip · 2010-05-26 09:03 · Score: 1

Right you are; sorry I forgot to put that in.
Re:Both good and bad by FreonTrip · 2010-05-26 09:06 · Score: 1

Right you are; I forgot to mention that. Thanks!
Re:Both good and bad by Anonymous Coward · 2010-05-26 10:07 · Score: 0

What are you talking about? The discreet graphics card market is in a very healthy state of competition. Just look at the pace at which AMD and NVidia are releasing new (plus, GOOD) products. The last thing I want is Intel joining the graphics card segment. They have way too much money and have already shown that they are willing to abuse it in order to slow down competitors.
Re:Both good and bad by kimvette · 2010-05-26 11:30 · Score: 1

Matrox has been irrelevant ever since 3D became important.

--
The Christian Right is Neither (Christian nor right). See: Matthew 23, Matthew 25, Ezekiel 16:48-50
Re:Both good and bad by BikeHelmet · 2010-05-26 13:29 · Score: 2, Interesting

Are you kidding me? This is great for consumers.
If Intel got their claws in the discreet graphics market (which is already showing signs of stagnation rather than growth), then they'd take a huge chunk of nVidia and ATI's R&D budgets away. Unable to put as much money towards advancement, GPU generations (and their pricedrops) would come slower. Meanwhile Intel would utilize their advanced (and cheap) fabbing to make a killing on that market, just as they do IGPs.
End result? Slower progress, nVidia and ATI suffer, Intel rakes in cash.
Our current duopoly is GOOD. Videocards drop in price and increase in performance quicker than CPUs do. Adding a behemoth competitor will hurt the industry.

Intel planed to put this tech into the next cpu an by Joe+The+Dragon · 2010-05-26 06:43 · Score: 1

Intel planed to put this tech into the next cpu and this seems to be dead so what will intel do?

Hrmmm. by SLot · 2010-05-26 06:44 · Score: 1

Doesn't bode well for the future of Project Offset.

Re:Wait, what? This is news? by Anonymous Coward · 2010-05-26 06:44 · Score: 3, Insightful

A large, publicly announced project with a great deal of media hype that had the potential to shake up the industry was cancelled. So, yeah, stop the presses.

Re:So Intels next cpu will the same suck video bui by odin1899 · 2010-05-26 06:46 · Score: 0, Offtopic

I'll just ignore your complete inability to distinguish between a possessive and a verb contraction then.

Re: Discreet by TaoPhoenix · 2010-05-26 06:47 · Score: 1

I almost let this slide until you put the other half of the pun in capitals!

There is lots of tasty competition producing NSFW "Discreet Graphics" that Sucks!

--
My first Journal Entry ever, in 8 years! http://slashdot.org/journal/365947/aphelion-scifi-fantasy-horror-poetry-webzine

Re:Intel planed to put this tech into the next cpu by Anonymous Coward · 2010-05-26 06:47 · Score: 0

Intel planed to put this tech into the next cpu and this seems to be dead so what will intel do?

According to who? Intel has always said Larrabee would be a discrete GPGPU based of x86 instruction set. I've never seen anything from Intel saying this would be used on an integrated chip.

Re:So Intels next cpu will the same suck video bui by oatworm · 2010-05-26 06:49 · Score: 0, Offtopic

First person to count to "potato" wins!

Re:So Intels next cpu will the same suck video bui by Anonymous Coward · 2010-05-26 06:49 · Score: 0

I buy AMD chips, *but* I would not pay for AMD graphics. Reason is it doesn't run nearly as well under Linux as nVidia chips.

And no, I don't really care about opensource drivers, I just want hardware I bought to work under OS I use. So for now, that will be AMD + nVidia for best bang for the buck. If AMD puts integrated graphics on their chips, then that's a negative for me as I'm not willing to pay for something that doesn't work. Intel integrated graphics work better than AMD under Linux at the moment.

Not really by Sycraft-fu · 2010-05-26 06:50 · Score: 4, Insightful

Everyone gets up on Intel integrated GPUs because they are slow, but they are looking at it from a gamer perspective. Yes, they suck ass for games, however that is NOT what they are for. Their intended purpose is to be cheap solutions for basic video, including things like Aero. This they do quite well. A modern Intel GMA does a fine job of this. They are also extremely low power, especially new newest ones that you find right on the Core i5 line in laptops.

Now what AMD may do well in is a budget gaming market. Perhaps they will roll out solutions that cost less than a discreet graphics card, but perform better than a GMA for games. That may be a market they could do well in. However they aren't going to "kill" Intel by any stretch of the imagination. For low power, non-gaming stuff using minimal power is the key and the GMA chips are great at that. For the majority of gaming, a discreet solution isn't a problem ($100 gets you a very nice gaming card these days) and can be upgraded.

Re:Not really by quanticle · 2010-05-26 07:34 · Score: 2, Informative

Their intended purpose is to be cheap solutions for basic video, including things like Aero.
Well, it depends on your definition of basic video, of course. I mean, I've seen Intel GMA chipsets struggle to display a 1080p Blu-Ray movie. Given that consumers increasingly are going to be hooking up their laptops to TVs and other larger displays, saying, "Oh, that's not basic video," isn't going to cut it.

--
We all know what to do, but we don't know how to get re-elected once we have done it
Re:Not really by Anonymous Coward · 2010-05-26 07:40 · Score: 1, Insightful

Now what AMD may do well in is a budget gaming market. Perhaps they will roll out solutions that cost less than a discreet graphics card, but perform better than a GMA for games. That may be a market they could do well in. However they aren't going to "kill" Intel by any stretch of the imagination.
I think you're selling AMD short. Intel has always relied on its fabs to keep things competitive even when their designs weren't. Now that die shrinks are beginning to reach a point of diminishing returns, Intel cannot rely on their fabs as heavily. Also, Fusion certainly has the potential to break Intel's lock on the low power IGP market.
Ever since they ran Ruiz out, AMD has been executing brilliantly. If AMD is able to come close to equaling Intel's CPU tech and considering that ATI's GPU tech spanks Intel, how much better do you all think Bulldozer will ultimately be than Larabee? I'm guessing metric tons considering applications (i.e. Photoshop) are now utilizing GPU acceleration for stream processing (not just HD accel. and Aero).
If AMD can keep this pace up, Intel is in for some deep hurting... You can double the hurting if (and it is a big if) ARM starts moving up the hardware stack (e.g. iPad) in the next couple of years.
Re:Not really by blahplusplus · 2010-05-26 08:41 · Score: 1

"Everyone gets up on Intel integrated GPUs because they are slow, but they are looking at it from a gamer perspective. "
Everyone SHOULD get up on intel. Intel doesn't approach computers as a platform like it should, and that's a huge problem for the #1 player in the industry when it comes to CPU's and motherboard chipsets. Nvidia and AMD are the few companies approaching the PC as a _platform in itself_. The idea that the "gamers perspective" doesn't matter is short-sighted and NAIVE, imagine you told Matrox, ATI, 3Dfx and nvidia that they should not "focus on gamers", Nvidia and 3Dfx practically pioneered and industry into existence based on providing hardware to make games run faster and prettier since even today NO CPU can compete with dedicated graphics hardware in software rendering.
Neither will integration be the death knelll for discrete GPU's because people forget 1) Heat 2) Power and 3) Bandwidth, I really don't care what the naysayer say - there is no fucking way you are going to be able to push computation in an integrated CPU/GPU hybrid like you can with dedicated hardware - the bandwidth problem is ALWAYS overlooked by the integrations. Mark Reign of epic games and many others predicted "the death of discrete graphics" in 2-5 years and they've bee nsaying that since 1997-98 or so and it's now 2010 and there is NO END IN SIGHT for discrete graphics since now they are targetting high performance computing space.
Who thought a bunch of hardware guys providing chips for graphic acceleration for gamers would grow into such multi-purpose behemoths?
Re:Not really by Klintus+Fang · 2010-05-26 09:09 · Score: 1

You are correct that discrete graphics isn't going anywhere, but your argument for why gaming performance is so important in integrated graphics contradicts that. I think the same argument applies there: the thing that people riding on Intel for the performance of their integrated gpu's and who are predicting that AMD will spank them in the integrated gpu space are forgetting is: 1) heat, 2) power, and 3) bw. When you integrate the GPU with the CPU you have to share the heat, power, and bw budget with the CPU.
It will be interesting to see what AMD comes up with when they have an integrated GPU, but frankly, the fact that it has taken them so long to get to that step since acquiring ATI tells me that they are likely finding it rather difficult to do well. I will be very happy if they somehow come up with something that can "spank" Intel in the integrated gpu space, but I wouldn't be surprised to find that they had to cut out so many features to get it into the power/bw budget that it ends up being...about the same.
Other point worth keeping in mind is that a huge portion of the market where intel's gpus do reign is the business desktop space where the only thing the machine needs to do is run web browsers, email apps, and office applications. Blue ray decoding? In that space, it doesn't matter.

--
In a minute there is time For decisions and revisions which a minute will reverse. -T.S. Eliot
Re:Not really by forkazoo · 2010-05-26 09:09 · Score: 5, Insightful

Everyone gets up on Intel integrated GPUs because they are slow, but they are looking at it from a gamer perspective. Yes, they suck ass for games, however that is NOT what they are for. Their intended purpose is to be cheap solutions for basic video, including things like Aero. This they do quite well. A modern Intel GMA does a fine job of this. They are also extremely low power, especially new newest ones that you find right on the Core i5 line in laptops.
Funny, at this point, I thought the purpose of Intel graphics was to try and make sure that OpenCL never becomes a viable solution. Seriously, Intel does everything in their power to make their terrible graphics chips universal. They've done some pretty shady dealing over the years to try and make it happen. At this point, they have even put their GPU's right on the CPU's of their current generation laptop chips. Apple and nVidia had to come up with dual-GPU solutions that can't be as power efficient as an Intel-only solution because they have to leave the Intel GPU also running and burning power. Intel is trying to sue nVidia out of the integrated chipset market. Examples go on and on.
Why? It isn't like Intel makes all that much money on their GPU's. It's nothing to sneeze at. Intel makes more money in a year on GPU's than I'll probably make in a lifetime, but that's peanuts on the scale of Intel. It's also not enough cash to justify the effort. But, if you look at it as a strategic move to make sure that the average consumer will never have a system that can run GPGPU code out of the box, it starts to make a little more sense. Intel is trying to compete on sheer terribleness of their GPU's, because if the average consumer has an nVidia integrated GPU in their chipset, then developers will bother to learn how to take advantage of GPU computing, which will marginalize Intel's importance.
I know it sounds kind of like a crazy conspiracy theory, but after the last several years of Intel-watching, it really does seem like quietly strangling GPGPU is a serious strategic goal for Intel.
Re:Not really by Anonymous Coward · 2010-05-26 09:17 · Score: 0

My thoughts exactly. I believe many people here vastly overestimate the importance of 3D performance in the mass market. As you say, when that's not thought of as the primary feature, GMAs start looking pretty good.
Re:Not really by Anonymous Coward · 2010-05-26 11:58 · Score: 0

Except that graphics cards are now starting to be used for processing capabilities. The better your card the more you can do especially if you are running some of the high end software that can take advantage of this.
Re:Not really by dave420 · 2010-05-26 13:13 · Score: 1

AMD is nowhere near Intel when it comes to CPUs. Not even close.
Re:Not really by Calinous · 2010-05-26 18:47 · Score: 1

AMD is performance-competitive at any price point below $150 for the processor (even more so as the mainboards are a bit cheaper). Intel is better at performance per watt (at any performance point), and AMD can't reach the performance of Intel's top processors/platforms.
Re:Not really by Calinous · 2010-05-26 18:50 · Score: 1

I think AMD's Fusion in the first step will be the same as integrated graphics now, just cheaper. I hope to be wrong, though.
Re:Not really by Calinous · 2010-05-26 18:51 · Score: 1

"Funny, at this point, I thought the purpose of Intel graphics was to try and make sure that OpenCL never becomes a viable solution."
Why write your applications in OpenCL now, to run on cards available now, when you can write them the next year for a platform incompatible with anything else existing?
Re:Not really by drinkypoo · 2010-05-27 01:15 · Score: 1

I know it sounds kind of like a crazy conspiracy theory, but after the last several years of Intel-watching, it really does seem like quietly strangling GPGPU is a serious strategic goal for Intel.
It's only a crazy conspiracy theory if you forget history. Intel is planning a future without a GPU and AMD's ATI division and nVidia are both inconvenient impediments on the road that intel envisions. The unfortunate thing is that such a future would probably be quite a bit better for users and developers alike.

--
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
Re:Not really by haruchai · 2010-05-27 06:56 · Score: 1

Old slogan: Intel Inside
New slogan: Intel Graphics - S3 Virge inside
Perhaps they'll consider buying S3 - even that would be an improvement on the GMA.
Or, perhaps, Matrox - that would actually be cool if they include multi-head as standard.

--
Pain is merely failure leaving the body

Re:Wait, what? This is news? by kdekorte · 2010-05-26 06:52 · Score: 2, Informative

The i740 card.... great expections, poor real world experience.

Re:So Intels next cpu will the same suck video bui by Anonymous Coward · 2010-05-26 06:53 · Score: 2, Funny

its potatoe you dumb fuck.

Re:Wait, what? This is news? by 0123456 · 2010-05-26 06:55 · Score: 1

The i740 card.... great expections, poor real world experience.

Everyone I knew in the graphics business thought that Intel had gone completely insane with the i740; other companies were trying to cram more and more faster and faster RAM onto their cards while Intel were going to use slow system RAM over a snail-like AGP bus.

So I'd say the expectations were pretty low, at least among those who knew what they were talking about.

Re:So Intels next cpu will the same suck video bui by Anonymous Coward · 2010-05-26 06:56 · Score: 0

ignoring you're complete inability to form a sentence

Never fails. Guy pulls a grammar faux pas when being a grammar Nazi to another poster. FFAS.

A discrete component by tepples · 2010-05-26 07:01 · Score: 4, Informative

More directly, what the hell is "discrete graphics"?

It refers to a graphics processor as a separate (discrete) component of a computer system. A chip that does nothing but graphics can be more powerful than integrated graphics because the GPU circuitry doesn't have to share a die with the rest of the northbridge.

I think they mean by Anonymous Coward · 2010-05-26 07:02 · Score: 0

a discrete (meaning separate) card, as opposed to a graphics system that's integrated with the CPU's processing power.

So basically, like ATI and nVidia have been doing for years.

Please someone tell me if I'm wrong... TFA doesn't really define the term "discrete graphics card" very well so I'm just using my understanding of the word.

Re:I think they mean by hedwards · 2010-05-26 07:47 · Score: 1

It's discrete as in not continuous. At least I think that's how the term came to be.
Re:I think they mean by Anonymous Coward · 2010-05-26 08:02 · Score: 0

Yeah, those old-fashioned analogue graphics cards were great for picture fidelity, especially with gold-plated graphics cables.

Re:Wait, what? This is news? by TheRaven64 · 2010-05-26 07:10 · Score: 2, Informative

To be fair to Intel, most graphics cards then were on the PCI bus, not AGP, so they didn't have the opportunity to use the host RAM except via a very slow mechanism. At the time, the amount of RAM was far more of a limitation than the speed, and a card using 8MB of host RAM via AGP was likely to have an advantage over a card with 4MB of local RAM on the PCI bus. While it was much slower than competing solutions, it was also much cheaper. The RAM on something like the VooDoo 2 was a significant proportion of the cost. A 740 cost about 20% of a VooDoo 2 and using system RAM had the advantage that you didn't have a load of RAM doing nothing while you were not doing 3D stuff. At the time the 740 was introduced, I had an 8MB VooDoo 2 and only 32MB of main memory. Having 8MB of RAM sitting doing nothing during the 90% of the time that I wasn't playing 3D games was a massive waste.

--
I am TheRaven on Soylent News

Intel's NotToBee GPU by julie-h · 2010-05-26 07:18 · Score: 2, Funny

Actually Intel have changed the name to NotToBee.

Larrabee was a hedge anyway by Funk_dat69 · 2010-05-26 07:20 · Score: 5, Interesting

I kind of think Larrabee was a hedge.

If you think about it, around the time it was announced (very early on in development, which is not normal), you had a bunch of potentially scary things going on in the market.
Cell came out with a potentially disruptive design, Nvidia was gaining ground in the HPC market, OpenCL was being brought forth by Apple to request a standard in hybrid computing.

All of sudden it looked like maybe Intel was a little too far behind.

Solution: Announce a new design of their own to crush the competition! In Intel-land, sometimes the announcement is as big as the GA. Heck, the announcement of Itanium was enough to kill off a few architectures. They would announce Larrabee as a discrete graphics chip to get gamers to subsidize development and....profit!

Lucky for them, Cell never found a big enough market and Nvidia had a few missteps of their own. Also, Nehalem turned out to be successful. Add all that up, and it becomes kind of clear that Larrebee was no longer needed, negating the fact that it was a huge failure, performance-wise.

Intel is the only company that can afford such huge hedge bets. Looks like maybe another one is coming to attack the ARM threat. We'll see.

--
FUNK!

Re:Larrabee was a hedge anyway by soppsa · 2010-05-28 23:59 · Score: 1

The new Atom's already do well against the ARM threat. If only they could price/package them sanely...

--
Ogre Wedding Planners llc.

Re:Wait, what? This is news? by 91degrees · 2010-05-26 07:22 · Score: 1

Thing is, Intel were never in the same market as AMD and nVidia. Sure, nVidia had a few budget parts but that's not their main product. They're really making money from mid-range chips for PC gamers.

Anyone who would be satisfied with Intel would consider AMD and nVidia to be hopelessly expensive. Anyone who would consider paying for a decent graphics card would consider Intel chips to be worthless.

Limited to 950 by manekineko2 · 2010-05-26 07:23 · Score: 5, Informative

Note for anyone else whose curiosity was piqued, this only works with 32bit systems with 950 chipset based systems, and does not work with GMA X3100, GMA X4500, GMA 500, or GMA 900.

Re:Limited to 950 by Svartalf · 2010-05-26 10:00 · Score: 1

Not only that, but people shouldn't confuse the GMA500, which is a rebadged PowerVR in the mix.

--
I am not merely a "consumer" or a "taxpayer". I am a Citizen of the State of Texas
Re:Limited to 950 by BikeHelmet · 2010-05-26 13:21 · Score: 1

It's interesting how although it boosts the clockspeed from 133/166mhz to 400mhz, the performance boost is approximately 25%.
It says right on the website, this is because it's RAM bandwidth starved - and as it gobbles more bandwidth, your CPU gets less, but it does result in a net gain.
That means in theory the performance gains could be higher on desktop systems with higher speed RAM, as opposed to laptops. However, being able to feed it more data also means it works harder, so the recommendation for decent cooling is good advice!

Fucking journalism by Anonymous Coward · 2010-05-26 07:24 · Score: 0

There's a difference between abandoning and postponing. But that would not be catchy anymore so you wouldn't make so much money on your website ads.

Re:Fucking journalism by itsybitsy · 2010-05-26 14:02 · Score: 1

"There's a difference between abandoning and postponing. But that would not be catchy anymore"
Right on target.
Now maybe Intel will cancel or abandon the Larrabee after all is said and done... or not... their failure so far was in the ring bus that limited their first chips to about 1 teraflops or so... unfortunately NVidia had already beat that target... so if Intel is to continue they need to be able to leap frog NVidia and ATI on performance.
From a chip perspective the Larrabee rocks. Maybe like the Itanium we won't see it happen...
Let's cross our fingers for it's a cool tech for us programmers to get our hands on and for product makers to produce cool stuff with.

Re:So Intels next cpu will the same suck video bui by LinuxIsGarbage · 2010-05-26 07:25 · Score: 1

That's good criteria, however I don't like a vendor that releases defective junk that delaminates from the mounting package, rendering the entire system useless under any OS: http://hplies.com/

Re:Wait, what? This is news? by 0123456 · 2010-05-26 07:25 · Score: 1

To be fair to Intel, most graphics cards then were on the PCI bus, not AGP, so they didn't have the opportunity to use the host RAM except via a very slow mechanism.

If by 'most' you mean 'Voodoo-2', yes. From what I remember all the cards I was using at the time Intel was trying to sell the i740 (Permedia-2, TNT, etc) were on the AGP bus.

I believe 3dfx were pretty much the last holdouts on PCI, because game developers had deliberately restricted their games to run well on Voodoo cards, thereby ensuring that they didn't need much bus bandwidth (any game which actually took advantage of AGP features so it ran well on a TNT but badly on a Voodoo was slated in reviews).

Re:Wait, what? This is news? by gman003 · 2010-05-26 07:33 · Score: 2, Interesting

The Larrabee chips actually looked pretty good. There was a lot of hype, especially from Intel. They demoed things like Quake Wars running a custom real-time ray-tracing renderer at a pretty decent resolution. Being able to use even a partial x86 ISA for shaders would have been a massive improvement as well, both in capabilities and performance.

From what I've been able to piece together, the problem wasn't even the hardware, it was the drivers. Apparently, writing what amounts to a software renderer for OpenGL/DirectX that got good performance was beyond them.

Another part was an odd insistence on doing all the rendering in software, even stuff like texel lookup and blitting, but that's another story.

Re:Wait, what? This is news? by hedwards · 2010-05-26 07:45 · Score: 1

You mean ATI and nVidia. AMD only recently took over ATI and the AMD chips were budget friendly compared with the typically much more expensive chips that Intel was selling. I realize that you're technically correct, it just a tad misleading to suggest that Intel was competing with AMD over that time period when it was a completely different company.

Re:So Intels next cpu will the same suck video bui by ElectricTurtle · 2010-05-26 07:47 · Score: 1

This is important, and a reason that I'm likely to buy an AMD/ATI card for my next upgrade after being an nVidia loyalist for the better part of a decade.

--
I support the Slashcott and will not be reading or commenting from 2/10/14 to 2/17/14. Beta is steaming pile of dog shit

Re:Wait, what? This is news? by TheRaven64 · 2010-05-26 07:52 · Score: 2, Insightful

From what I remember all the cards I was using at the time Intel was trying to sell the i740 (Permedia-2, TNT, etc) were on the AGP bus.

Check the dates. The i740 was one of the very first cards to use AGP. Not sure about the Permedia-2, but the TNT was introduced six months after the i740 and cost significantly more (about four times as much, as I recall). It performed a lot better, but that wasn't really surprising.

--
I am TheRaven on Soylent News

Re:So Intels next cpu will the same suck video bui by Anonymous Coward · 2010-05-26 07:57 · Score: 0

Actually the RadeonHD project has drivers available for the AMD chipsets sold on motherboards today. Just last month I bought an Asus mainboard with integrated ATi graphics in its AMD chipset and it works flawlessly under Ubuntu 10.04. Just Google the specs of the chipset and then the Linux/X.org/RadeonHD support for it. These days I'd rather have an integrated ATi that works conveniently and efficiently for less then having an nVidia add-on board with its own cooling and a proprietary blob of code that no one except nVidia can fix. I'd actually be suprised if you can buy an AMD chipset with integrated graphics that's not supported by the latest distros. I haven't tried ATi add-on boards but RadeonHD has support for those as well.

As for Intel graphics, well, you get what you pay for.

Re:Wait, what? This is news? by ElectricTurtle · 2010-05-26 08:10 · Score: 1

Considering that the context here is Larrabee which didn't publicly exist until after the AMD/ATI merger, the usage is appropriate. Further, that merger was half a decade ago. In the tech world that really doesn't count as 'recent' anymore. You really need to get over it and live in the present.

--
I support the Slashcott and will not be reading or commenting from 2/10/14 to 2/17/14. Beta is steaming pile of dog shit

Re:Intel planed to put this tech into the next cpu by Anonymous Coward · 2010-05-26 08:30 · Score: 0

Larrabee was never intended to go into any next-gen Intel CPU, so it's not a problem for them at least in that respect.

Re:So Intels next cpu will the same suck video bui by Nadaka · 2010-05-26 08:49 · Score: 1

Same here. I am running a Asus motherboard with the Radeon HD 3300 graphics chipset, dual booting 9.10, now 10.04 and windows 7. I have drivers available under both OS's and the hardware is plenty good enough for everything that I have thrown at it (admittedly, I am not exactly playing crysis, but it decodes 720p without a hiccup and plays all the games that I both have and like).

Re:So Intels next cpu will the same suck video bui by petermgreen · 2010-05-26 09:17 · Score: 1

So Intel's next cpu will the same suck video build in?
Intel are already building in their sucky (though not quite as sucky as it used to be) video into their dual core i3 and i5 chips (technically it's a multi-die module ATM but from the system integrators POV that doesn't really make any difference). With the next gen I believe they are planning to put it on-die on all their low and mid range chips (maybe the high end too, information on the next gen high end stuff seems very sketchy at the moment)

IMO Integration of graphics onto the CPU was a pretty inevitable result of Intel's decision to put the memory and fast PCIe controllers on the CPU. Putting the graphics elsewhere in such a system would either require dedicated graphics memory or having a very fast link between the CPU and the device containing the graphics with carefully designed prioritisation.

AMD will kill them.
BS, the only market segment intels integration of video into the CPU will really impact are those who currently explicitly buy laptops with nvidia chipsets to get a bit better graphics performance without the size and battery life sacrifices of a fully independent graphics soloution. Afaict those are a fairly small proportion of the laptop market.

From a desktop perspective this is no big deal. Graphics cards with performance comparable to the best integrated graphics aren't exactly expensive. Gamers will go from leaving the graphics integrated in the northbridge disabled to leaving the graphics integrated in the CPU disabled.

Plus even if AMD had a technically better solution than Intel afaict they simply don't have the production capacity to kill Intel any time soon. Nor it seems do they have the marketing.

--
note: i'm known as plugwash most places but i screwd up registering that here somehow in the past and now can't register

depends on GP-GPU by Creepy · 2010-05-26 09:27 · Score: 1

I would say that is currently true, but the average user may care if General Purpose GPU (GP-GPU) takes off and they use applications that use it. For a speed example, I had what is essentially a math problem that kept a dual core CPU busy (and yes, it was threaded) for 2 weeks, 3 days, 14 hours. The same problem tackled by 216 GPU shaders and one CPU took around 25 minutes. While neither

I realize most people aren't doing surface detail analysis involving trillions of points of data like I was (actually, that was a brute force method, too - probably could have it down to about a day, optimized), but I know a lot of people that use Photoshop, and imagine the same sort of gains for certain filters. Filters that were too slow to incorporate 5 years ago may be possible today.

Re:depends on GP-GPU by Calinous · 2010-05-26 18:41 · Score: 1

You probably did floating point work on double precision (in fact 80-bits precision) in the CPU and in 32-bits precision on the graphic card. If this "increased error" doesn't bother you, GPUs offer better computing performance and much higher memory bandwidth compared to any common CPU (triple channel DDR3 1600MHz has less than 40 GB/s, while a 5870 has over 150 GB/s and a Fermi has 177 GB/s)

Why Intel chips are faster by Tumbleweed · 2010-05-26 09:51 · Score: 1

Intels chips are faster because Intel has much better production facilities.

No way, man, it's their Speed Hole(tm) Technology!

Question by Locke2005 · 2010-05-26 09:51 · Score: 1

Does anybody know where the Larrabee development was actually done? I'd be embarrassed if it was done at Intel here in Oregon.

--
I've abandoned my search for truth; now I'm just looking for some useful delusions.

Re:Question by Anonymous Coward · 2010-05-26 11:20 · Score: 0

Most Larrabee R&D was done in Oregon. Sorry.

Isn't this old news? by Locke2005 · 2010-05-26 09:54 · Score: 1

Larrabee was canceled "as a standalone discrete graphics product" on December 4, 2009.

--
I've abandoned my search for truth; now I'm just looking for some useful delusions.

Re:So Intels next cpu will the same suck video bui by Svartalf · 2010-05-26 10:09 · Score: 1

Heh... I'm none too happy with either of them. And "doesn't run nearly as well" is a relative concept- I've had decent results (though not as good as with NVidia) with my AMD parts I've got- though there ARE glitches with the proprietary drivers (which is where all the problems with AMD's stuff arises from- even on Windows.).

So, while you've got decent overall performance with reasonably stable drivers, you've got to deal with a company that did what NVidia did with their packaging a while back.

On the other hand, while you've got better theoretical speed and overall performance (including bang for buck and bang for watt spent...), you've got to deal with a company that can't manage to get their drivers as well off as they ought to and haven't for years running now... I'm hoping that the driver devs with xorg, etc. can get the FOSS AMD drivers really going good so everyone will be ahead of the game on that front, including AMD.

--
I am not merely a "consumer" or a "taxpayer". I am a Citizen of the State of Texas

Good... by BulletMagnet · 2010-05-26 10:14 · Score: 1

At least everyone knows that Intel's integrated stuff sucks. Some of you might have forgotten the i740. Bet Otellini didn't

Re:Good... by Khyber · 2010-05-26 16:32 · Score: 1

The i740 wasn't integrated. It was a discrete AGP-bus card.
God I remember playing POD on that thing.

--
Still waiting on Serviscope_minor to wake up to fucking reality and realize that Jessica Price isn't going to fuck him.

Re:Wait, what? This is news? by hedwards · 2010-05-26 11:42 · Score: 1

Not really, it's been nearly a decade since Intel was competing even nominally in that market. The i740 was the last time I remember them even shipping a discrete card. And that was a PCI card in an AGP slot.

Some more fuel by RelliK · 2010-05-26 12:08 · Score: 2, Interesting

There was a company called Rapid Mind, which built library & tools for writing code to target various GPUs, multi-core CPUs, etc. Something similar OpenCL, I suppose, but easier to program (theoretically -- I never actually tried it). Intel bought it and killed it.

Another company, Havok, developed a successful physics & AI library. They were going to port it to to GPUs. Then Intel bought it and canceled the GPU port.

--
___
If you think big enough, you'll never have to do it.

Re:So Intels next cpu will the same suck video bui by LinuxIsGarbage · 2010-05-26 13:20 · Score: 1

I was recently in the market for an oversized (12"), slightly more powerful netbook. I considered buying an ION powered HP 311, but decided not to. After having the Go 7200 fail in my HP laptop (repaired under extended warranty, now waiting the final death failure), and seeing Go 6150 fail at least once (if not twice) in every HP laptop I've seen so equipped, I wasn't going to jump out and buy another nVidia/HP. Not to say I won't ever buy one, just not now. I ended up with an AMD/ATI powered MSI computer.

Re:So Intels next cpu will the same suck video bui by BobNET · 2010-05-26 16:18 · Score: 1

its potatoe you dumb fuck.

This is what happens when you cross Dan Quayle with Joe Biden.

Re:So Intels next cpu will the same suck video bui by Calinous · 2010-05-26 18:54 · Score: 2, Insightful

"Graphics cards with performance comparable to the best integrated graphics aren't exactly expensive"
You can't find expansion graphic cards with performance comparable to the current integrated graphics - the integrated graphics are slower than anything else (less available memory bandwidth, fewer compute clusters, ...).

Re:So Intels next cpu will the same suck video bui by petermgreen · 2010-05-26 23:42 · Score: 1

I thought some of nvidia's integrated graphics for core 2 processors (which afaict are the best integrated graphics out there for intel processors at the moment) were compararable to nvidia's bottom end cards but i've failed to find any benchmarks confirming or refuting this.

--
note: i'm known as plugwash most places but i screwd up registering that here somehow in the past and now can't register

Re:So Intels next cpu will the same suck video bui by Calinous · 2010-05-27 00:17 · Score: 1

The 9400M has the same number of shaders as the 9400GT (the desktop graphic chip), so here there is somewhat parity (they have a bit faster core and shader clock, but use main memory).
Similar situation for AMD, the cheapest card (3450) has 40 cores, just as the HD4290 integrated graphic core from the 890GX chipset (8 times 5-way).

OpenCL vs. RapidMind by DrYak · 2010-05-27 04:18 · Score: 1

There was a company called Rapid Mind, which built library & tools for writing code to target various GPUs, multi-core CPUs, etc. Something similar OpenCL, I suppose, but easier to program (theoretically -- I never actually tried it).

I've tested it.
- RapidMind is high level stuff, easy to code, relying mostly on C++ Template meta programming and macros.
(Just declare a special type of vector and it mostly works)
- OpenCL is low-lever stuff, looking pretty much similar to OpenGL
(calls to initialise hardware, calls to setup the calculations, calls to allocate memory, etc...)

In fact, last time I've checked (~1 year ago) RapidMind was planning to use OpenCL as a possible backend.

RapidMind can be compared to Brook as used by AMD (both are very high level, except that Brook is a separate language, a dialect of C with stream computing)
OpenCL can be compared to CUDA when using the lower-level "driver" API (whereas CUDA with the higher-level API + Cuda Tools macros is a middle ground between OpenCL and the highlevel)

Now the whole thing is a shame, because Intel was very vocal in the OpenCL development to make sure that this API was also compatible with their weird Larrabee architecture (somewhere mid-way between a multicore CPU with cache coherency, full flow control and other typically CPU features, and simplistic but massively multi-core SIMD / SIMT unit like most modern GPUs - it would have had lots of Atom/Pentium grade x86 core each with a massive SIMD unit attached to it)

--
"Sufficiently advanced satire is indistinguishable from reality." - [Tips: 1DrYakQDKCQ6y52z6QbnkxHXAocMZJE61o ]

Re:Wait, what? This is news? by 91degrees · 2010-05-27 06:42 · Score: 1

Doh! Yes. Initialism fail on my part.

Slashdot Mirror

Intel Abandons Discrete Graphics

165 comments