AMD sent Techreport an Intel system to test their GPU's

[Idontcare]

Orly? Because we both looked at the same sheet that had a $1000 Intel chip on it.

Lol @ you calling price performance ratio arbitrary. Fanboi.

"Orly"? "Lol @ you"? "Fanboi"?

Are you here just to troll and make inflammatory posts or are you here for a serious discussion of the thread's topic?

Your banter makes me think perhaps you are just here to stir the pot and try and irritate as many people as possible while not adding anything germane to the thread topic.

2005 just called, they want their flamewar back. Please consider indulging their request.

 

[Torn Mind]

Not sure what you are trying to say. Because your original statement is still invalid. AMD took all the money they could when they had the lead. And in no way did they give any "charity". It was first in july 2006 we got a huge price change when AMD lost the performance lead we got cheapish CPUs again with the E6600. And that was simply because Intel understood more sold CPUs was better than higher prices in terms of profit. And AMD had to cut prices in the following months with above 75%.

In 2006 a FX62 was 791$ and performned like a 228$ E6400 if you want to play the EE game. 4x4 was AMDs desperate move afterwards.

The problem with AMD fanboys is that they have adopted a part of the AMD culture which has grown into a cancer: no accountability, just blame big bad Intel or some other scapegoat and ignore all problems with their product. Or dredge up their previous successes to mask their current failures.

I knew what I knew now in 2004, I would have sccoped up AMD CPUs like hot cakes because THEY WERE BETTER. Now, I don't want a damn AMD cpu because they suck. I have to pay more for a PSU because an FX is a guzzler or I have to buy more expensive RAM if I'm getting an APU, and all of their CPUs have flaky properties.

Hah, the original Bulldozers was priced above their comparable Intel counterparts. In fact, the FX-8150 were going for 269.99 for quite a while on Amazon(link), far more than the i5-2500k ever did throughout its entire lifespan(link).

 

[mrmt]

Lol @ you calling price performance ratio arbitrary.

No, I'm calling *YOUR* price/performance ratio arbitrary, because so far you didn't outlay anything about this ratio, and I'm saying that cheaper than something means costing less than something, not better price/performance ratio than something as you are willing to imply.

Fanboi.

Fanboi. You can't even look at the mirror, do you?

A more reasonable person could just say something like "Yes, when they had better chips compared to Intel offers they could charge more, when they have the better performing desktop chips, they would obviously have the more expensive chip on the market, but that was a brief time, for the most time of AMD existence they offered cheaper chips compared to Intel's, and as they do a lot of die-salvaging they offer some nicer deal, or at least chips with less features fused off for the folks in the bottom market".

Instead what we got is a torrent of disconnected posts trying to prove that despite AMD chips costing nominally more than Intel's, in the end they tend to be cheaper once you factor an arbitrary Terry's price/performance ratio.

It is *exactly* that kind of insistence in deny a minute and obvious point like "AMD's processors sometimes were more expensive than Intel's" that makes you a fanboy. It may be too much for you to admit that sometimes AMD isn't the best deal around and they do charge more when they can.

 

[TerryMathews]

No, I'm calling *YOUR* price/performance ratio arbitrary, because so far you didn't outlay anything about this ratio, and I'm saying that cheaper than something means costing less than something, not better price/performance ratio than something as you are willing to imply.



Fanboi. You can't even look at the mirror, do you?

A more reasonable person could just say something like "Yes, when they had better chips compared to Intel offers they could charge more, when they have the better performing desktop chips, they would obviously have the more expensive chip on the market, but that was a brief time, for the most time of AMD existence they offered cheaper chips compared to Intel's, and as they do a lot of die-salvaging they offer some nicer deal, or at least chips with less features fused off for the folks in the bottom market".

Instead what we got is a torrent of disconnected posts trying to prove that despite AMD chips costing nominally more than Intel's, in the end they tend to be cheaper once you factor an arbitrary Terry's price/performance ratio.

It is *exactly* that kind of insistence in deny a minute and obvious point like "AMD's processors sometimes were more expensive than Intel's" that makes you a fanboy. It may be too much for you to admit that sometimes AMD isn't the best deal around and they do charge more when they can.

I fail to see how it's arbitrary.

IPC, especially today, is an accepted performance metric. Number of instructions per clock cycle. A measure of efficiency, of sorts.

Frequency is a measure of oscillations per second.

Those two numbers multiplied together give you operations per second.

Divided by price gives you a reasonable approximation of its price/performance ratio. Bang for the buck if you will.

The excel spreadsheet I quickly did jived pretty well with how I remember the landscape being in 2005.

At any rate, I'm done with this line of discussion. The whole thing is moot, because AMD hasn't effectively invested in R&D since the K7. You think AMD somehow "charged more" even though you've been shown two Intel products priced above anything AMD ever charged at the time. Used math to show that AMD also offered better bang for the buck across their product line. Called simple math arbitrary.

Someone else pointed out that over clocking was easier too. That was ignored.

So yes. Fanboi.

 

[TerryMathews]

"Orly"? "Lol @ you"? "Fanboi"?

Are you here just to troll and make inflammatory posts or are you here for a serious discussion of the thread's topic?

Your banter makes me think perhaps you are just here to stir the pot and try and irritate as many people as possible while not adding anything germane to the thread topic.

2005 just called, they want their flamewar back. Please consider indulging their request.

I have made concrete posts on this topic.
On topic, I don't see anything unusual about a discrete GPU being tested on Intel. In general, that is the reality of the market. Doubly so in the field of discrete mobile graphics. For better or worse, Llano and Trinity is AMD's mobile future.

I made one comment related to another comment in this thread: that AMD has been cheaper, even when they held the performance crown. I stand behind that comment. I justified it. Mathematically it's obvious. Logically, it has to be true.

AMD doesn't have the name recognition Intel does or did. They also didn't advertise as much. And certainly they didn't have as many OEM contract units - thanks to Dell. How else can you explain market share growth?

Anyway, I'm done with this line. You're right, it is off topic. And one side is never going to convince the other of anything.

For the record, I remember Intel absent any effective competition. It wasn't pretty. I hope we never get back there.

 

[Idontcare]

I made one comment related to another comment in this thread: that AMD has been cheaper, even when they held the performance crown. I stand behind that comment. I justified it. Mathematically it's obvious. Logically, it has to be true.

I'm not taking exception to this portion of your posts, rather it is the inflammatory quips that are undermining your very message.

I paid $1500 for a QX6700 in Nov 2006, it was outrageous to be sure but it was a product of supply/demand. Intel had a limited supply and the demand was so high that even at $1500 a pop, Newegg sold out shortly after I snagged mine.

My point is just that the market sets the going price more than the seller, unless the seller wants to be sitting on a bunch of inventory that doesn't move.

AMD never had issues moving X2's at $1k until Core 2 Duo showed up. I don't fault AMD for selling what they sold, nor do I fault Intel for selling at what they sell at.

AMD did have the choice though of expanding capacity and producing more X2's versus spending $5B on ATI. They chose to buy ATI and continue to remain capacity constrained in the X2's. I don't recall Intel ever making a direct business choice which resulted in prop'ing up the supply/demand curve as AMD did.

 

[TerryMathews]

I'm not taking exception to this portion of your posts, rather it is the inflammatory quips that are undermining your very message.

I paid $1500 for a QX6700 in Nov 2006, it was outrageous to be sure but it was a product of supply/demand. Intel had a limited supply and the demand was so high that even at $1500 a pop, Newegg sold out shortly after I snagged mine.

My point is just that the market sets the going price more than the seller, unless the seller wants to be sitting on a bunch of inventory that doesn't move.

AMD never had issues moving X2's at $1k until Core 2 Duo showed up. I don't fault AMD for selling what they sold, nor do I fault Intel for selling at what they sell at.

AMD did have the choice though of expanding capacity and producing more X2's versus spending $5B on ATI. They chose to buy ATI and continue to remain capacity constrained in the X2's. I don't recall Intel ever making a direct business choice which resulted in prop'ing up the supply/demand curve as AMD did.

Sure and I do see your point. It's hard to not get frustrated when you show someone mathematically that they sky is blue and they still argue that it's orange.

I wish AMD had continue to develop and innovate. They really only had that one big breakthrough in the K7.

I wonder where we would be if K8 was as revolutionary as K7 was. And I give all the credit in the world to Intel for abandoning Net burst and starting back over with the PPro architecture. Having the guts to start back over was very shrewd. And profitable.

It always seemed to me that Intel maintained their margins through technology more than supply restrictions: new, incompatible sockets. Very restrictive chipset licensing. Multiplier locks. Feature differentiation (cache-less Celerons, VTx missing on low-end processors at one point). The only real AMD example of that I can remember is the Opteron 2xx vs 8xx.

Have a Merry Christmas IDC!

 

[Idontcare]

Have a Merry Christmas IDC!

Merry Christmas to you too :thumbsup:

 

[mrmt]

I fail to see how it's arbitrary.

You don't because you are an AMD fanboy, because you cannot admit to yourself that AMD will charge more when they can, regardless of how much *you* bend the definition of cheaper.

AMD will charge the price they need to maximize their profit. If you think they don't do that you either think that AMD sales execs are idiots or you have no clue on how sales work. If AMD can charge more, they will charge, if they don't they won't. It is simple as that. When you say that AMD execs should have charged more and didn't charge is the same of saying that AMD execs were idiots at the time.

 

[Homeles]

You don't because you are an AMD fanboy, because you cannot admit to yourself that AMD will charge more when they can, regardless of how much *you* bend the definition of cheaper.

AMD will charge the price they need to maximize their profit. If you think they don't do that you either think that AMD sales execs are idiots or you have no clue on how sales work. If AMD can charge more, they will charge, if they don't they won't. It is simple as that. When you say that AMD execs should have charged more and didn't charge is the same of saying that AMD execs were idiots at the time.

I have no idea what the context is of this discussion, but a good case to your point would be the HD 7970 upon release.

 

[Sleepingforest]

Terrymatthews: Here's the problem with purely instructions per dollar: you don't factor in TDP and the electricity costs, you don't factor in how many instructions per second may actually be needed (as in, you may need greater power, but going up the graph in power may decrease the per dollar efficency), and you aren't factoring in the rest of the system that the CPU is going into.

 

[TerryMathews]

You don't because you are an AMD fanboy, because you cannot admit to yourself that AMD will charge more when they can, regardless of how much *you* bend the definition of cheaper.

AMD will charge the price they need to maximize their profit. If you think they don't do that you either think that AMD sales execs are idiots or you have no clue on how sales work. If AMD can charge more, they will charge, if they don't they won't. It is simple as that. When you say that AMD execs should have charged more and didn't charge is the same of saying that AMD execs were idiots at the time.

Yes. I actually do think AMD's execs were and currently are idiots.

 

[ViRGE]

I've already knocked a few heads together for this, but as a reminder to everyone we're cracking down on these inane "fanboy" insults. If you can't post in CPUs without attacking the other party (and calling someone a fanboy is virtually always an attack), then you need to reevaluate your posting habits. This forum is better than petty insults.

I would also like to remind everyone that this thread has a topic. If you want to rehash K8, I would be more than happy to spin off your posts.

-ViRGE

 

[Ferzerp]

:wub: Virge

 

[Chiropteran]

This has what to do with anything? Let alone Nvidia designs ARM, not x86?

So nvidia cpu + nvidia gpu = 0 performance. Nvidia has the brains to realize that nobody would buy a GPU that can't function at all, as they would in their own tegra based systems, so they use test systems with intel CPU.

[edit: actually, tegra does include support for pci-e. nvidia could use it with their gpu, they would need to write drivers and such and performance would be abysmal of course but they could do it]

AMD CPU + AMD GPU = >100% potential performance. AMD also has the brains to realize that few customers would buy a GPU that functions worse than it should, as they would in a system based on their own admittedly slower CPUs, so they use test systems with intel CPU.

Intel GPU + Intel CPU = >100% potential performance. Intel has the brains to realize that few customers would buy their CPU based on castrated performance of mating them with onboard video only, so they use test systems with nvidia or AMD GPU.


Can someone explain why this is thread-worthy when the intel & nvidia equivalent situations, which occur all the time, are not?

It would simply be bad business sense to castrate a product by matching it with anything less than the best possible components. Nvidia doesn't do it, Intel doesn't do it, why should AMD do it? This thread is full of hypocrites.

 

[CHADBOGA]

So nvidia cpu + nvidia gpu = 0 performance. Nvidia has the brains to realize that nobody would buy a GPU that can't function at all, as they would in their own tegra based systems, so they use test systems with intel CPU.

[edit: actually, tegra does include support for pci-e. nvidia could use it with their gpu, they would need to write drivers and such and performance would be abysmal of course but they could do it]

AMD CPU + AMD GPU = >100% potential performance. AMD also has the brains to realize that few customers would buy a GPU that functions worse than it should, as they would in a system based on their own admittedly slower CPUs, so they use test systems with intel CPU.

Intel GPU + Intel CPU = >100% potential performance. Intel has the brains to realize that few customers would buy their CPU based on castrated performance of mating them with onboard video only, so they use test systems with nvidia or AMD GPU.


Can someone explain why this is thread-worthy when the intel & nvidia equivalent situations, which occur all the time, are not?

It would simply be bad business sense to castrate a product by matching it with anything less than the best possible components. Nvidia doesn't do it, Intel doesn't do it, why should AMD do it? This thread is full of hypocrites.

Intel doesn't sell discrete cards for desktop, so of course they are going to use someone else's discrete card.

The testing is meant to show performance on x86 software, so of course Nvidia aren't going to use their ARM cpu.

It takes a real effort to not understand the above and to be crying about hypocrites.

 

[SPBHM]

Intel should send this with their latest CPUs

 

[Idontcare]

:wub: Virge

+1

 

[Chiropteran]

Intel doesn't sell discrete cards for desktop, so of course they are going to use someone else's discrete card.

AMD doesn't sell ultra high end desktop CPU, so of course they are going to use a competitors CPU. No difference whatsoever.

It takes a real effort to try to argue that AMD is somehow doing this "wrong" when the other players are all doing the exact same thing.

 

[Ferzerp]

AMD doesn't sell ultra high end desktop CPU, so of course they are going to use a competitors CPU. No difference whatsoever.

It takes a real effort to try to argue that AMD is somehow doing this "wrong" when the other players are all doing the exact same thing.

This is the same company that tried to claim that the 8150 was as good as Intel's $1k part in their marketing materials (in gpu limited scenarios of course, but they sure weren't pointing that out!). That is why people are amused by it.

 

[Puppies04]

Intel doesn't sell discrete cards for desktop, so of course they are going to use someone else's discrete card.

The testing is meant to show performance on x86 software, so of course Nvidia aren't going to use their ARM cpu.

It takes a real effort to not understand the above and to be crying about hypocrites.

Unfortunately I have seen this almost exact arguement in another thread a while back in this forum. Trust me it takes him no effort at all.

 

[Chiropteran]

Doing something because it's good business sense is somehow seen as worse than doing something because you have no other choice. Maybe hypocrite is the wrong word, I should have said "illogical nonsense".

AMD is clearly in a better position than Nvidia, having the option to use their own CPU at a slight performance penalty or to use Intel CPU and enjoy the best possible performance. Nvidia, in comparison is completely reliant on Intel.