So everybody is expecting broad well. Its certainly coming early 2015 from apple. Probably late January early february. Great way to start off the year, right guys? Anyways, the reason I know that the graphics performance would increase by 50% for the retina macbook pro is because the graphics tdp will drop by half.
So if you don't know. tdp is thermal design power. its how they measure how big to make the heatsink and such. If you don't know apple as well as i do, then you'll know in order to get the maximum tdp of your chip, you will have to burn your cpu to 100C. Unlike other companies who keep it around 70C, which is nominal, apple likes to push their laptops to its limits in order to get thinner and more battery.
The tdp of the graphics card for ivy bridge was 45W. So in order to get the full performance out of the graphics on ivybridge you would have to have heat sink that can take care of 45W of power. obviously at the time, macbook pro had a 35w limit and the air had 17w limit. Haswell dropped that power to 30W. which is almost 33% faster in graphics performance based on tdp alone +15~20% in the architectural change. Keep in mind though the tdp limit for the 13.3" models is 28w so the macbook has to share cpu power and gpu power to achieve the maximum performance of the gt3 architecture. rMBP cpu can take up to 20-22W of the 28W allowed. so basically in order to achieve full combined performance the tdp limit should be 22+30+4(this is for the pch controller now integrated in the chip itself)=56W TDP.
The New Broadwell CPU is now 15W TDP for the graphics. so lets say they keep the same 28W tdp limit. 28W-4W = 24W available to share between the CPU and GPU. 15(GPU)+17W(CPU (based on my estimates making the cpu more efficient)) + 4W PCH= 36W of TDP is need for maximum combined performance. 28W/36W is only 80% of away from peak performance. Where as 28/56W is 50% away from peak performance. 30% increase in peak performance. 10% more peak performance could could from the l4 cache integrated for the graphics. its like a memory buffer for the graphics card. this helps a lot because the memory they use in graphics cards are way more sophisticated and powerful than the shared video memory we use for day to day operations. it won't be as fast as the dedicated graphics in memory performance, but will certainly close the gap.
I am talking about the 13.3" rMBP and the 6100 iris that will coming for it.
Highly opinionated Information. Please be easy on me.
So if you don't know. tdp is thermal design power. its how they measure how big to make the heatsink and such. If you don't know apple as well as i do, then you'll know in order to get the maximum tdp of your chip, you will have to burn your cpu to 100C. Unlike other companies who keep it around 70C, which is nominal, apple likes to push their laptops to its limits in order to get thinner and more battery.
The tdp of the graphics card for ivy bridge was 45W. So in order to get the full performance out of the graphics on ivybridge you would have to have heat sink that can take care of 45W of power. obviously at the time, macbook pro had a 35w limit and the air had 17w limit. Haswell dropped that power to 30W. which is almost 33% faster in graphics performance based on tdp alone +15~20% in the architectural change. Keep in mind though the tdp limit for the 13.3" models is 28w so the macbook has to share cpu power and gpu power to achieve the maximum performance of the gt3 architecture. rMBP cpu can take up to 20-22W of the 28W allowed. so basically in order to achieve full combined performance the tdp limit should be 22+30+4(this is for the pch controller now integrated in the chip itself)=56W TDP.
The New Broadwell CPU is now 15W TDP for the graphics. so lets say they keep the same 28W tdp limit. 28W-4W = 24W available to share between the CPU and GPU. 15(GPU)+17W(CPU (based on my estimates making the cpu more efficient)) + 4W PCH= 36W of TDP is need for maximum combined performance. 28W/36W is only 80% of away from peak performance. Where as 28/56W is 50% away from peak performance. 30% increase in peak performance. 10% more peak performance could could from the l4 cache integrated for the graphics. its like a memory buffer for the graphics card. this helps a lot because the memory they use in graphics cards are way more sophisticated and powerful than the shared video memory we use for day to day operations. it won't be as fast as the dedicated graphics in memory performance, but will certainly close the gap.
I am talking about the 13.3" rMBP and the 6100 iris that will coming for it.
Highly opinionated Information. Please be easy on me.
