Will we ever see a dGPU in a 13in Macbook?

[MagicBoy]

So you've now set the cutoff point to be that the GPU can't have ANY memory of it's own, no matter if it's on it's on it's own piece of silicon or even it's own PCB?

I've not changed "the cutoff point". OK, so I didn't explicitly mention the lack of dedicated memory, I shouldn't have to. It's obvious to anyone that's been around computers with integrated graphics that they never had dedicated memory. Ever.

The lack of performance of integrated graphics compared to dedicated graphics cards of whatever era is primary due to :
1) limited memory bandwidth because of the use of slower memory. VRAM/GDDRx on a discrete card generally has twice the bandwidth of system memory.
2) the GPU is contending for that memory bandwidth with the CPU which is doing it's normal operations.

The 320M might borrow the core config of the 9600M (I don't know, I haven't checked .. but why re-design something when you don't have to... ), but it will have slower memory access and bandwidth then the dedicated GPU.

 

[SarcasticJoe]

I've not changed "the cutoff point". OK, so I didn't explicitly mention the lack of dedicated memory, I shouldn't have to. It's obvious to anyone that's been around computers with integrated graphics that they never had dedicated memory. Ever.

I've worked with machines that have 9400M or 320M graphics and I can say from personal experience that they're not horribly slow in any way, specially for their time. Compared to the Intel X3100 and specially GMA950 machines that came before them, the 9400M ones were miles ahead in terms of graphical performance and the 320M machines were even further ahead. You're trying to discredit these chips by some by complaining that they didn't have dedicated memory when in reality they performed well even without dedicated memory.

The lack of performance of integrated graphics compared to dedicated graphics cards of whatever era is primary due to :
1) limited memory bandwidth because of the use of slower memory. VRAM/GDDRx on a discrete card generally has twice the bandwidth of system memory.
2) the GPU is contending for that memory bandwidth with the CPU which is doing it's normal operations.

So your argument is just memory bandwidth, repeated twice over, as if that was the only factor in play...

The 320M might borrow the core config of the 9600M (I don't know, I haven't checked .. but why re-design something when you don't have to... ), but it will have slower memory access and bandwidth then the dedicated GPU.

Sure, it's going to be slower than exactly the same chip with memory dedicated to itself, but it's not an automatic "This chip is really slow"-stamp. Performance wise the 9400M and 320M were able to compete with other chips that had dedicated memory, not mid or upper range chips, but still chips with dedicated memory.

 

[MagicBoy]

I've worked with machines that have 9400M or 320M graphics and I can say from personal experience that they're not horribly slow in any way, specially for their time. Compared to the Intel X3100 and specially GMA950 machines that came before them, the 9400M ones were miles ahead in terms of graphical performance and the 320M machines were even further ahead. You're trying to discredit these chips by some by complaining that they didn't have dedicated memory when in reality they performed well even without dedicated memory.

Thanks for answering a question I didn't ask! I'm not trying to discredit anything other than your flawed logic.

I never took a pop at the 9400M/320M - they were far better than anything the other chipset makers were producing at the time. Check the Mac mini in my sig. It was bought on launch day and has never missed a beat in 5 years. I'd been waiting what seemed an eternity for Apple to make one without the awful Intel GMA950. Now retired from desktop duties it very happily functions as an XBMC HTPC playing 1080p content without breaking sweat. It runs a 9400M, as did my old uMB.

 

[joe-h2o]

I've worked with machines that have 9400M or 320M graphics and I can say from personal experience that they're not horribly slow in any way, specially for their time. Compared to the Intel X3100 and specially GMA950 machines that came before them, the 9400M ones were miles ahead in terms of graphical performance and the 320M machines were even further ahead. You're trying to discredit these chips by some by complaining that they didn't have dedicated memory when in reality they performed well even without dedicated memory.


So your argument is just memory bandwidth, repeated twice over, as if that was the only factor in play...


Sure, it's going to be slower than exactly the same chip with memory dedicated to itself, but it's not an automatic "This chip is really slow"-stamp. Performance wise the 9400M and 320M were able to compete with other chips that had dedicated memory, not mid or upper range chips, but still chips with dedicated memory.

Dude, you lost the argument. Man up and accept it kid.

Changing the argument and putting words in your opponent's mouth isn't going to cover it up.

 

[Count Blah]

People want it, but not enough. So apple makes sure you pay TOP dollar to get a dGPU by forcing you to go 15" and NOW even forcing you to go top of the line 15"

If you want performance in a 13", you have to buy elsewhere. Only low-mid to mid performance allowed for such a price at apple.

 

[SarcasticJoe]

It's obvious to anyone that's been around computers with integrated graphics that they never had dedicated memory. Ever.

Thanks for answering a question I didn't ask! I'm not trying to discredit anything other than your flawed logic.

In your post you talked about how it was clear that anyone who has worked with dedicated chips knows that they never had dedicated memory. I countered this by pointing out that the chips you wanted to label as "integrated" (9400M and 320M) were dedicated chips that just had their dedicated memory ripped out and put on the same die as a chipset. To continue on the Land/Range Rover theme of comparisons, it's still like saying the more luxurious ones aren't terrain vehicles because they took out the mud guards and made the carpets too nice.

I never took a pop at the 9400M/320M - they were far better than anything the other chipset makers were producing at the time. Check the Mac mini in my sig. It was bought on launch day and has never missed a beat in 5 years. I'd been waiting what seemed an eternity for Apple to make one without the awful Intel GMA950. Now retired from desktop duties it very happily functions as an XBMC HTPC playing 1080p content without breaking sweat. It runs a 9400M, as did my old uMB.

Trying the old charm offensive are we? Well it's not going to work as the discussion between the two of us was about the 9400 and 320M from the get-go. Just go back to the sentence I quoted at the beginning of my post and your previous claims that the 9400 and 320M are integrated chips.

 

[MagicBoy]

LOL. No. The fundamental point is ... you don't know what integrated graphics are.

I'm done with this. You're just making yourself look silly now.

/unsubscribes from thread.

 

[leman]

Trying the old charm offensive are we? Well it's not going to work as the discussion between the two of us was about the 9400 and 320M from the get-go. Just go back to the sentence I quoted at the beginning of my post and your previous claims that the 9400 and 320M are integrated chips.

Frankly, you are acting beyond stupid now. Nobody in this thread has ever denied that 9400M and 320M were great GPUs of in their class. The only argument was about your delusional denial of the fact that they are integrated GPUs.

 

[ha1o2surfer]

The Iris and the Iris + 750m models are being heavily throttled so it doesn't matter anyways.. You aren't getting full performance out of them..

I don't even have Iris (which apparently is just as powerful as my dGPU lol)

 

[orestes1984]

A dedicated GPU has its own RAM
An Integrated GPU does not except in some weird cases such as the Nforce and Nforce2 chipsets which were respectively pretty much dedicated GPUs onboard anyway.

Really is as simple as that ladies and gentlemen, the Gefore 9400m and Geforce 320 were IGPS While I'd love to see a dedicated GPU on 13" MacBook Pros, I'm not holding my breath.

There are a lot of factors involved particularly with Intel and them not licensing out the rights to make a North/South bridge chipset layout like Nvidia did with the Geforce320m and 9400m on the LGA775 platform. Then there is the fact that Apple exclusively uses Intel chips with no AMD options, then there is fitting it and the heat/power consumption issues, and then there is the fact that Apple isn't known to go all out and support high end dedicated GPUs anyway particularly not on its mid range computers.

So, we're stuck here, but mostly unless anything changes and Intel decides to relax their licensing and allow Nvidia to design its own boards, we probably will never see another Nvidia IGP on an Apple motherboard, or any other one for that matter.