Is AMD Still Relevant?

[rudy]

Its sad to see. AMD just doesn't want to do the hard work to beat Intel in IPC. They've done it before, it's not impossible.

They only beat intel, because intel screwed up. As soon as intel was able to react and fix their mistake the game was over. AMD could have been in alot better position if they had not screwed up the phenom launch so bad.

The fact is AMD in the CPU world does not have a good enough talent pool to design, and produce CPUs. They are always at least 1 node behind, their yield is typically worse as such their only option is to make an architecture that is better. But unless it is alot better it wont overcome their other 2 deficiencies. And most of the time they do not pull it off.

 

[demingo]

Sometimes I wonder if AMD just doesn't have enough money to compete. Between the technical costs, and salary cost of top designers I wonder if Intel can just throw more money at CPUs and come out ahead.

I remember reading an article a couple years ago about how Intel, even through the economic downturn kept R&D spending high, while AMD made significant cuts. It might be that we are seeing the results of those decisions now.

 

[ebduncan]

lol what is with all the doom and gloom?

Amd will be around for many years to come.

 

[Jagger100]

Even if you think Intel is awesome and all being in front of AMD. Intel pushes the envelope only because AMD exists and is on their heels.

Hard to say how far back we would be without the competition. And Windows needs Intel's advances to offset its bloat.

 

[Azhar]

lol what is with all the doom and gloom?

Amd will be around for many years to come.

Haven't you heard? It's cool to say that AMD will go bankrupt next year every year.

 

[Zarathustra[H]]

They only beat intel, because intel screwed up. As soon as intel was able to react and fix their mistake the game was over. AMD could have been in alot better position if they had not screwed up the phenom launch so bad.

Exactly.

Intel screwed the pooch big time with P4 and netburst, and this - together with help from recent hires from Digital and licencing the Alpha EV6 bus resulted in the highly successful K7 line.

AMD had a rare opportunity to cement themselves in the top tier CPU business but instead of investing big in CPU R&D, they blew $6B on purchasing ATI, and had to reduce their CPU R&D spending as a result leading to lackluster Phenom through Bulldozer product lines.

Essentially they:
a.) Didn't have enough experience to realize what it actually took to develop a successful CPU architecture from scratch, after purchasing all or parts of their last few designs, and hiring work forces of other companies
b.) Got greedy and bought ATi instead of spending the money on R&D because of th elack of understanding in a.)

Got greedy; and

 

[Dan_D]

Exactly.

Intel screwed the pooch big time with P4 and netburst, and this - together with help from recent hires from Digital and licencing the Alpha EV6 bus resulted in the highly successful K7 line.

AMD had a rare opportunity to cement themselves in the top tier CPU business but instead of investing big in CPU R&D, they blew $6B on purchasing ATI, and had to reduce their CPU R&D spending as a result leading to lackluster Phenom through Bulldozer product lines.

Essentially they:
a.) Didn't have enough experience to realize what it actually took to develop a successful CPU architecture from scratch, after purchasing all or parts of their last few designs, and hiring work forces of other companies
b.) Got greedy and bought ATi instead of spending the money on R&D because of th elack of understanding in a.)

Got greedy; and

Pretty much what I said earlier.

 

[rudy]

Even if you think Intel is awesome and all being in front of AMD. Intel pushes the envelope only because AMD exists and is on their heels.

Hard to say how far back we would be without the competition. And Windows needs Intel's advances to offset its bloat.

Can you make money, IE survive if you serve no purpose other than to push another company? No I do not think you can, certainly not in the CPU world where being the best not only means you have a faster product it typically means you have a cheaper product and the combination of those 2 is very cut throat. It is not like cars or other products where making a better product typically costs alot more and there is room for the value player to make a cheaper product. Amd is forced to cut its prices, it does not choose to do so.

 

[DJ Big T]

Got greedy; and

and????

 

[Cmdrmonkey]

They haven't been relevant for a long time. That the Core 2 Conroe and Geforce 8800, two of the most revolutionary products of all time, came out when they were preoccupied with merging with ATI in 2006 spelled doom for the company. They've never really recovered from it.

 

[Zarathustra[H]]

and????

Typo and no edit.

I accidentally typed portions of my b.) section twice.

 

[Acer_Sheep]

That article is retarded, who can tell that some company, something, someone is relevant?? There is NO point in writing this garbage as well as reading it.

 

[Zarathustra[H]]

They haven't been relevant for a long time. That the Core 2 Conroe and Geforce 8800, two of the most revolutionary products of all time, came out when they were preoccupied with merging with ATI in 2006 spelled doom for the company. They've never really recovered from it.

On the GPU side they aren't doing badly.

Remember how much sooner than Nvidia they came out with their DX11 5000 series GPUs when Fermi was delayed badly.

Currently Nvidia may have a slight edge with their 680, but AMD isn't far behind.

(they really do still need to work on those GPU drivers though, especially for CFX)

 

[DJ Big T]

Yes they are still relevant, Cnet... Lol.

Also im going to go on record and say i don't see all this bad drives/software everyone talks about, And have never had any problems.

About Brooke Crothers (author)

Brooke Crothers writes about small devices and the hardware inside. He has served as editor at large at CNET News and a contributing reporter to The New York Times' Bits and Technology sections. His interest in things small began when living in Tokyo in a very small apartment for a very long time. He is a member of the CNET Blog Network and is not an employee of CNET.

Someone on [H] posted a blog on Forbes from another one of these "contributors" and now this? These guys get paid from writing these "controversial" topics.

How about, is this article relevant? Does it use any real facts or sources?

 

[ktk_ace]

110USD for a bulldozer FX6100 , 120USD MB

how is that NOT relevant?

Hell, even the 6770/6850 sells like hotcakes in my area.

writer must be an intel c*cksucker.

 

[Dan_D]

On the GPU side they aren't doing badly.

Remember how much sooner than Nvidia they came out with their DX11 5000 series GPUs when Fermi was delayed badly.

Currently Nvidia may have a slight edge with their 680, but AMD isn't far behind.

(they really do still need to work on those GPU drivers though, especially for CFX)

This. I'd buy another Radeon HD 7970 right now if they fixed the CrossfireX drivers. That's a cheaper solution than replacing my 7970 and buying two GTX 680's etc.

 

[PRIME1]

On the GPU side they aren't doing badly.

Their top chip is getting beat by NVIDIA's mid-range chip. I would say that's pretty bad.

 

[Zarathustra[H]]

110USD for a bulldozer FX6100 , 120USD MB

how is that NOT relevant?

When a $119 Core i3-2100 blows it away is the things most people do?

 

[BababooeyHTJ]

Their top chip is getting beat by NVIDIA's mid-range chip. I would say that's pretty bad.

Huh? Then where is "Nvidia's top chip"? Six months away with super limited availability at best?

 

[Zarathustra[H]]

Their top chip is getting beat by NVIDIA's mid-range chip. I would say that's pretty bad.

The 680 is Nvidia's top end single GPU board right now, and the direct successor to the GTX580.

It is irrelevant to compare AMD products to Nvidia products that don't exist, regardless of what plans Nvidia may have had.

 

[KingMing]

When a $119 Core i3-2100 blows it away is the things most people do?

Dont' slay me if I am wrong but seems usually even if an "older" cheaper intel can beat a similiar priced Amd chip the price on unused motherboards still goes AMDs way.... so overall love or hate either AMD or INTEL it is still bang for buck cheaper to build AMD...

 

[Dan_D]

The 680 is Nvidia's top end single GPU board right now, and the direct successor to the GTX580.

It is irrelevant to compare AMD products to Nvidia products that don't exist, regardless of what plans Nvidia may have had.

I couldn't agree more. I don't know why the hell people keep refering to the GTX 680 as a "midrange GPU."

 

[Zarathustra[H]]

Dont' slay me if I am wrong but seems usually even if an "older" cheaper intel can beat a similiar priced Amd chip the price on unused motherboards still goes AMDs way.... so overall love or hate either AMD or INTEL it is still bang for buck cheaper to build AMD...

I'm not much for slaying.

This used to be the case, but Intel Socket 1155 motherboards have actually become pretty cheap now. You can even find some at the $45 pricepoint after rebate on the egg...

AMD has lost any bang-for-the-buck advantage it ever had, unfortunately. for the holidays in 2010 I built both my stepsons rigs based on overclocked Phenom II X4's, and they performed admirably, and were great bang-for-the-buck systems.

If I were doing the same thing today, I wouldn't go AMD. A low end Sabdy bridge will be both cheaper and faster in almost everything.

 

[dr.kevin]

based on their stock price, most people must think not.

that said, i have 3 amd machines, all bought on the cheap.

 

[Trimlock]

I couldn't agree more. I don't know why the hell people keep refering to the GTX 680 as a "midrange GPU."

Prime is a known nvidia troll, you would go blue in the face argueing with him. I find it much easier to have him on ignore and talk objectively with the rest.

 

[PRIME1]

I couldn't agree more. I don't know why the hell people keep refering to the GTX 680 as a "midrange GPU."

Based on the size of the chip for one. The name "GK104" which would make it the successor to the GF104 which was the 460. Also there is the GK110 which is clearly a monster chip like the 480/580 before it. It's pure logic. Seriously look how small the PCB is on the 670..

The 680 was supposed to be a 660 but once they saw how poorly the 7970 performed they are now able to sell a $300 part for $500. The GK110 they can now sell purely to the professional market for upwards of $2000.

 

[DejaWiz]

Exactly.

Intel screwed the pooch big time with P4 and netburst, and this - together with help from recent hires from Digital and licencing the Alpha EV6 bus resulted in the highly successful K7 line.

AMD had a rare opportunity to cement themselves in the top tier CPU business but instead of investing big in CPU R&D, they blew $6B on purchasing ATI, and had to reduce their CPU R&D spending as a result leading to lackluster Phenom through Bulldozer product lines.

Essentially they:
a.) Didn't have enough experience to realize what it actually took to develop a successful CPU architecture from scratch, after purchasing all or parts of their last few designs, and hiring work forces of other companies
b.) Got greedy and bought ATi instead of spending the money on R&D because of th elack of understanding in a.)

Got greedy; and

I'm going to disagree with ATi being a bad purchase. AMD has since developed one hell of a good performing APU thanks to integrating ATi technology into the CPU die. Amazing performance for the speed that the APU and IGP are clocked at, plus it's only going to get better as AMD focuses R&D on improving things past Trinity. That's where the bread and butter is for both intel and AMD: consumer and business SFF and mobile/ultra-portable computing.

The only worry I have is AMD's ability to keep improving it's APU performance by a significant margin with each new generation. intel's HD4000 is certainly no slouch in many tasks that are more typical home and business use oriented, and a great leap over HD3000, so it's a safe assumption that Haswell will be bringing it's A-game in this department.

 

[Acer_Sheep]

Their top chip is getting beat by NVIDIA's mid-range chip. I would say that's pretty bad.

That fanboism is bad, GTX680 is max offered high end.

 

[Zarathustra[H]]

I couldn't agree more. I don't know why the hell people keep refering to the GTX 680 as a "midrange GPU."

Probably because GK104 was originally intended to be a midrange GPU, but then NVIDIA had trouble manufacturing the larger GK100 chip, and had to bin and clock up GK104 higher than it was originally intended in order to make it their new GTX680.

I feel like Nvidia was shaking in their boots thinking they were going to miss out on this generation due to their GK100 problems, until the 7970 launched and it was a little bit more sedate than they expected, and they breathed a sigh of relief.

This doesn't mean that the 680 is a mid range product. The binned and higher clocked GK104's in the 680 are likely performing much higher than the mid range Kepler was originally intended to perform.

Nvidia will likely eventually come out with a full sized GK100, but by the time they do, it may be a respun GK110 to fix the issues, and it may be named GTX780 as part of the next gen, at which point AMD's 8000 GPU's are likely to be on the horizon...

So, GTX680 is NOT a midrange product, even though the GPU around which it is based was originally intended for midrange boards.

 

[Deleted whining member 223597]

If only their CPU's could finally get good like in the A64 vs. P4 days.

I want to use AMD again

 

[BababooeyHTJ]

Based on the size of the chip for one. The name "GK104" which would make it the successor to the GF104 which was the 460. Also there is the GK110 which is clearly a monster chip like the 480/580 before it. It's pure logic. Seriously look how small the PCB is on the 670..

The 680 was supposed to be a 660 but once they saw how poorly the 7970 performed they are now able to sell a $300 part for $500. The GK110 they can now sell purely to the professional market for upwards of $2000.

Talk to me when I can buy a GK110. Availability is just starting to pick up on GK104. Do you really think that Nvidia could supply a gpu twice that size anytime soon?

If code names mean so much what was G92?

 

[Zarathustra[H]]

I'm going to disagree with ATi being a bad purchase. AMD has since developed one hell of a good performing APU thanks to integrating ATi technology into the CPU die. Amazing performance for the speed that the APU and IGP are clocked at, plus it's only going to get better as AMD focuses R&D on improving things past Trinity. That's where the bread and butter is for both intel and AMD: consumer and business SFF and mobile/ultra-portable computing.

The only worry I have is AMD's ability to keep improving it's APU performance by a significant margin with each new generation. intel's HD4000 is certainly no slouch in many tasks that are more typical home and business use oriented, and a great leap over HD3000, so it's a safe assumption that Haswell will be bringing it's A-game in this department.

I'm not saying that their APU's are bad. They aren't.

But was it worth sacrificing being able to compete with their main rival in their core business of CPU's?

 

[DejaWiz]

Based on the size of the chip for one. Seriously look how small the PCB is on the 670..

PCB size has nothing to do with the class and category of a GPU. Computer parts get smaller with each die shrink and technology advancement, because that's the way it goes.

Hell, compare the physical size and computing power of ENIAC to Raspberry PI.

 

[Jagger100]

Can you make money, IE survive if you serve no purpose other than to push another company? No I do not think you can, certainly not in the CPU world where being the best not only means you have a faster product it typically means you have a cheaper product and the combination of those 2 is very cut throat. It is not like cars or other products where making a better product typically costs alot more and there is room for the value player to make a cheaper product. Amd is forced to cut its prices, it does not choose to do so.

Didn't say it was viable. Just saying they provide a service, even to diehard intel fans.

And a monopolies make me cry.

 

[Dan_D]

That fanboism is bad, GTX680 is max offered high end.

Probably because GK104 was originally intended to be a midrange GPU, but then NVIDIA had trouble manufacturing the larger GK100 chip, and had to bin and clock up GK104 higher than it was originally intended in order to make it their new GTX680.

I feel like Nvidia was shaking in their boots thinking they were going to miss out on this generation due to their GK100 problems, until the 7970 launched and it was a little bit more sedate than they expected, and they breathed a sigh of relief.

This doesn't mean that the 680 is a mid range product. The binned and higher clocked GK104's in the 680 are likely performing much higher than the mid range Kepler was originally intended to perform.

Nvidia will likely eventually come out with a full sized GK100, but by the time they do, it may be a respun GK110 to fix the issues, and it may be named GTX780 as part of the next gen, at which point AMD's 8000 GPU's are likely to be on the horizon...

So, GTX680 is NOT a midrange product, even though the GPU around which it is based was originally intended for midrange boards.

At the end of the day, what the GPU's code name was, and what it was originally intended for is irrelevant. It more than competes with AMD's highest end offering and it's name as the GTX 680 denotes what it is now.

 

[Trimlock]

Zara, nvidia stated that the next high end geforce is slated for q1 2013, gk110 based. Hopefully they can get that one out.

I would also say that I thought the hutu port of ati was good for them, but they might of competed better on the cpu side if they didn't have the lull from the purchase. Also don't forget amd only had the chance to compete because intel was trying to brute force their competition out, the A64 was a great processor but intel was still competing pretty well then took complete dominance with the core technology.

 

[DejaWiz]

I'm not saying that their APU's are bad. They aren't.

But was it worth sacrificing being able to compete with their main rival in their core business of CPU's?

APU's are not a sacrifice to compete in their core business of CPU's, because general consumer and business grade CPU's with IGP's ( aka APU's) has been their core business, is their core businesstoday, and will be their core business from here on out. That's where the volume sales are at. Consumer enthusiast CPU's and GPU's are nothing but a sliver of their overall sales revenue.

First started out with budget slot-based cards to feed the OEM suply chain en masse, then went to budget MoBo integrated graphics, now it's CPU integrated graphics. That's how the business has evolved in order for them to manufacture APU's cheaper thus allowing for more potential profit margin per unit.

 

[Trimlock]

Hutu port = buy out, freaking weird auto correct for droid to make.

 

[XamediX]

I dunno, the 9700/9800 Pro kicked some serious ass.

I agree. One of my favorite video cards was the 9800PRO that I flashed to an XT. But yea I'm not saying they didn't have their high points but I think to continue competing at such a high level, they needed the things a merger with AMD would bring.

I know but he said something about AMD being the "only reason ATI survived". Just pointing out that ATI was capable of great things.

I apologize for the wording, I meant to say AMD being the only reason they continued to be so competitive. And like Dan said, a lot of ATI's successes were gleamed from 3rd party tech/ip.

I was an AMD fan boy all the way from the K7 up to my first Intel rig using a 2600k. Jumping from AMD to Intel for my latest processor was AMAZING. I could not believe the performance increase. This and a very good price point made the switch so easy.

I did the same thing but with a 2500k. Amazing. My last Intel chip was a PIII733 which didn't OC to 933mhz like all the rest seemed to be able to.Haha.

Their top chip is getting beat by NVIDIA's mid-range chip. I would say that's pretty bad.

If you mean the 670 beating the 7970 in certain tests, then I can see where you are coming from but the 670 also almost beats NV's top chip as well. So ... *shrug*

This used to be the case, but Intel Socket 1155 motherboards have actually become pretty cheap now. You can even find some at the $45 pricepoint after rebate on the egg...

AMD has lost any bang-for-the-buck advantage it ever had, unfortunately. for the holidays in 2010 I built both my stepsons rigs based on overclocked Phenom II X4's, and they performed admirably, and were great bang-for-the-buck systems.

If I were doing the same thing today, I wouldn't go AMD. A low end Sabdy bridge will be both cheaper and faster in almost everything.

I noticed the same thing in my transition. My mobo was like $80.

 

[demingo]

On the GPU side they aren't doing badly.

Remember how much sooner than Nvidia they came out with their DX11 5000 series GPUs when Fermi was delayed badly.

Currently Nvidia may have a slight edge with their 680, but AMD isn't far behind.

(they really do still need to work on those GPU drivers though, especially for CFX)

My biggest concern with AMD is developer relations. While the 7970 is a very impressive piece of hardware, the software is lacking. I see NVidia constantly working with devs to improve the performance of Geforce hardware. Now you may or may not agree with this business tactic, but in the end, if Nvidia is investing money to ensure a better experience for their customer, they are going to be successful. This has become painfully evident in the last year where many games on release worked better on Nvidia. If AMD doesn't step up developer relations the divide in performance will just grow.