What actually makes a 7970m more expansive than a 7870?

I am sure this has been asked before but i have never come across it being answered. Why is the 7970m, which is just an under clocked 7780. 3 times more money? The 7780 is 200 bucks and the 7970m is 600 IIRC at time of release. Is it just price gouging? If anything it would cost a little less due to less material needed to build and less labor or is the marginal smaller board magically cost 3x the amount?

HF

 

Edit that post first please. Notebook cards are usually more expensive than their desktop counterparts because

A) They have to run at a strict thermal limit
B) Their demand is MUCH lower than their desktop counterparts

 

A how does that change the actual hardware? The laptop 3720qm 3920xm are essentially the exact same thing as a 3770k just packaged differently. Maybe binned differently but that is debatable.
B...again how does that actually make them more expensive besides "supply and demand" and "price gauge" because they can.

You provided nothing useful -_-

 

Because they can charge more for it. A desktop gpu can be swapped out at any time for any pci-e card made in the last ten years and probably will be made in the next ten years. There are lots of suppliers competing with each other also. Also you are comparing a top tier mobile card to a mid range desktop card, so they are sold at different price points.When it comes to laptop gpus, they basically have your balls in a vise.


Sent from my HUAWEI G520-0000 using Tapatalk

 

reason why i am asking is because it is the same GPU for the most part right? Their profit margins must be through the roof for those mobile cards. 400 bucks more per card from what it looks like I just find it hard to believe that duopoly has that crazy prices. The lack of competition in that market is gross to think i am paying 600 bucks for a card that cost 100-150 to make -_- It makes apples profit margins a joke in that regard.

 

Not always. As stated MXM cards sell nowhere near the volume as your counterpart PCI-E cards. They are also essentially putting a desktop class GPU and restricting it to 100W power draw max. It takes engineers to do that, and hence why it costs $$$ and adds onto the cost of the card.

Why can they charge 600 for the same card and get away with it? Because they can. Unless you want to have a contract with TSMC and fab your own MXM PCB, sell them for far cheaper than anyone else, let me know.

 

How does under clocking and undervoltimg require engineers?

 

It requires engineers to:

1) custom fabricate a MXM module + PCB to fit that GPU fabbed by TSMC, and ensure it is what OEMs want
2) set upper/low boundaries on power consumption for the GPU
3) get some kind of vBIOS for testing for Clevo/Alienware/MSI machines
4) testing the MXM modules.

Again if you think you know better, contact TSMC, say you want to buy a million GPUs, custom fab your own MXM cards, write custom vBIOS for all the myriad of different specifications/requirements, and sell them for 150, let me know and I'll buy some. In fact I think the whole Alienware/Clevo/MSI community would thank if you did. There's more it than just saying they undervolt/clock cards, if it was that easy then anyone could do it. So do it.

 

never said i did...just want to learn and understand and want to sift through the BS that people spew. People tend to restate the same jibberish over and over and never think if they are just spreading bad info. Have you ever heard of any sort of volume figures for the 7970m/680m vs 7780? I dount the process to bring together a 7970m for production vs a 7870m is horrible more expansive. Granted if 7970m sells 1000 while a 7780 sells 100000. I could easily see why the price difference but without actual figures it is hard to see if there is a notable cost difference or just monopolizing on the market. I am just curious because i have always wondered why there was such a huge difference...if it was a really cost difference or profit difference.

 

Of course the actual production costs realistically aren't that much more. MXM 3.0b hasn't really changed much, they can essentially reuse the same technology, but they already have the specifications. Nvidia already spends XX money on engineering/research, they have the equipment, they get the fabbed GPUs from TSMC, so everything in the assembly line is already done for Nvidia/AMD. But as I stated, unless you wish to go in the buisness of fabbing, we are stuck purchasing MXM cards that require a 2nd mortgage. The actual reason it costs more is because only a few people make MXM cards, it's not like every OEM goes out and produces them, and it's essentially mini-monopoly. If you want XX product and only 2-3 people make them, prepare to pay $$$.

 

I think the cost to produce an mxm card is the same as a pci-e card. Companies come out with custom pcb versions of desktop cards all the time. Changing the layout of a pcb is not a big deal. That is basically the difference between pcie and mxm cards. Designing the chip and how to get everything working with it is the hard part. Putting the gpu into a different form factor is not much work relative to that. Also you have discrete gpus on motherboards. It is all the same thing. They have a design and they just need to shoehorn it in to whatever form factor they need.

Sent from my HUAWEI G520-0000 using Tapatalk

 

Fixed vs variable costs over smaller quantities and supply vs demand, this is simply an economics question. There is a little bit of chip binning involved but it's mostly economics.

 

If the 8970M is worth 600$ then the rest of the MSI GX60 is worth 500$.

 

8970M is not worth 600, it's a 7970M.