The Cortex A8 core in the A4 with the cache is about as large as the Tegra 2's 2 cores + cache. A9 on TSMC's 40nm is twice as dense as A9 on Samsung's 45nm.
iPad 2
- Thread starter AlphaWolf
- Start date
The Cortex A8 core in the A4 with the cache is about as large as the Tegra 2's 2 cores + cache. A9 on TSMC's 40nm is twice as dense as A9 on Samsung's 45nm.
But perhaps all that proves in that intrinsity's tech that optimised the cortex A8 inside Apple's A4 for max speed/power profile did so at the expense of chip area ?, as opposed to necessarily being because of samsung's manufacturing process.
Powerpoint in Office 2010 was a significant upgrade which uses Direct3D for tons of effects actually. It's not that archaic any longer.
The A9 is smaller than the A8 as is. The A8 also included the NEON engine as well as all 32 DP-FP registers.
Plus domino circuits balloon area. I think there are way more variables to conclude on the density of the process.
An interesting comparison would be with Tegra 3 as that will include both NEON and VFPv3-D32 on TSMC's 40nm. You can then get an idea of how much of a density difference exists between the two processes.
ltcommander.data
Regular
http://ca.gizmodo.com/5789093/the-near+future-of-mobile-gaming-is-going-to-be-pretty-epic
It was talked about a bit in the NGP thread, but this article is relevant here since it's primarily Tim Sweeney on the iPad 2. Basically he believes the 9x GPU performance improvement claims, prefers the iOS platform to Android, is satisfied wit h 512MB of RAM, but would like to see more predictability in free RAM available instead of being at the whim of background apps.
Interestingly, he points out a major limitation right now is the OpenGL ES drivers which have too much overhead and limit the amount of objects that can be point on screen. Carmack has also pointed to OpenGL ES driver overhead as limitations as well. It's been mentioned before that new PowerVR drivers for the SGX540 significantly improve it's performance relative to Tegra 2. Can new PowerVR drivers directly address this or is it a fundamental limitation of Apple's OpenGL ES pipeline or even OpenGL ES itself?
On a side note, Tim Sweeney advocates for a future of OpenGL everywhere, but does UE3 even natively support OpenGL (non-ES) yet or be able to target OS X or Linux? Despite Sweeney's talk, Epic seems to be moving backwards in this area since UE2.x natively had an OpenGL codepath and OS X and Linux support.
It was talked about a bit in the NGP thread, but this article is relevant here since it's primarily Tim Sweeney on the iPad 2. Basically he believes the 9x GPU performance improvement claims, prefers the iOS platform to Android, is satisfied wit h 512MB of RAM, but would like to see more predictability in free RAM available instead of being at the whim of background apps.
Interestingly, he points out a major limitation right now is the OpenGL ES drivers which have too much overhead and limit the amount of objects that can be point on screen. Carmack has also pointed to OpenGL ES driver overhead as limitations as well. It's been mentioned before that new PowerVR drivers for the SGX540 significantly improve it's performance relative to Tegra 2. Can new PowerVR drivers directly address this or is it a fundamental limitation of Apple's OpenGL ES pipeline or even OpenGL ES itself?
On a side note, Tim Sweeney advocates for a future of OpenGL everywhere, but does UE3 even natively support OpenGL (non-ES) yet or be able to target OS X or Linux? Despite Sweeney's talk, Epic seems to be moving backwards in this area since UE2.x natively had an OpenGL codepath and OS X and Linux support.
My first (probably dumb) question would be whether the overhead is due to the API itself or if it's more of a generic trend for all OGL_ES drivers out there. If it should be the first case I don't think there's all too much a driver itself could do; necessary changes will most likely come in future API versions.
As for the MP2 in the iPad2 vs. the SGX535 in iPad GLBenchmark2.0 shows an overall difference of roughly 5.6x times. Unfortunately the frequencies are still a question mark for either/or.
http://www.glbenchmark.com/compare.jsp?benchmark=glpro20&showhide=true&certified_only=1&D1=Apple%20iPad%202&D2=Apple%20iPad
The texture fetch test gives 920822 kTexels/s for the iPad2 (4 TMUs) vs. 179099 kTexels/s for the iPad (2 TMUs), which might not be due to a possible frequency increase (if there even is any) alone.
As for the MP2 in the iPad2 vs. the SGX535 in iPad GLBenchmark2.0 shows an overall difference of roughly 5.6x times. Unfortunately the frequencies are still a question mark for either/or.
http://www.glbenchmark.com/compare.jsp?benchmark=glpro20&showhide=true&certified_only=1&D1=Apple%20iPad%202&D2=Apple%20iPad
The texture fetch test gives 920822 kTexels/s for the iPad2 (4 TMUs) vs. 179099 kTexels/s for the iPad (2 TMUs), which might not be due to a possible frequency increase (if there even is any) alone.
$99 - that's pretty ridiculous.
For something that is designed to seemlessly be carried with your iPad and with an aluminum shell, it's not that ridiculous. I'd buy one if I had an iPad. Also considering that the wireless BT keyboard that Apple sells for the Mac cost around $80, it's not all that expensive.
not that expensive compared to incredibly overpriced apple peripherals isn't exactly much of an endorsement
Yes a lot of Apple branded accessories are overpriced but their aluminum wireless keyboard is not one of them. I paid $75 for Logitech's illuminated ultrathin USB keyboard when they were first available and it's made of plastic...
Do you think aluminum is expensive or something? Not that $75 for that logitech wasn't overpriced as well.
Bluetooth is your friend.
Best keyboards I've ever used over the last 25 years. All our desktops in the house have one.
Yes, but my ipad2 is for business trips to Asia so it must be small. I think I'm close to being able to go laptop free for a week in China.
The raw material may not, but casting and milling costs might be <shrug>
With that shape I'd say it must be cast and/or milled, further casting/milling would give a much more rigid design.