Recent content by thomase

I wouldn't be surprised if the SOCs themselves were more expensive, but the improved thermals and integration allowed them to cut costs elsewhere. Sent from my Nexus 6P using Tapatalk

Sorry, I missed the smiley... Sent from my Nexus 6P using Tapatalk

Don't read into it too much. It doesn't necessarily mean top performing compared to their other products that they offer. Language like this typically means top performing relative to their competitors, which we know can be subjective depending on what is being measured. Sent from my Nexus...

My concern is that 16nm is still, and will most likely remain, more expensive than earlier process nodes. Moore's law really is nearing its end. Nintendo is very cost sensitive. Sent from my Nexus 6P using Tapatalk

The more important question is whether it's 20nm planar or 16nm FinFet. Sent from my Nexus 6P using Tapatalk

Not so sure that they will be using NAND flash. http://wccftech.com/nintendo-switch-cartridge-16gb-launch/ Sent from my Nexus 6P using Tapatalk

Don't forget that Nintendo is risk-averse. The risk of bugs or other issues with an older silicon-proven microarchitecture may be lower than with something newer. Sent from my Nexus 6P using Tapatalk

If they want to scale down power consumption and performance in undocked mode, it may be more efficient to cut the power entirely to a fraction of the CUDA cores. For example, due to leakage effects, turning off half of the CUDA cores while keeping the clock speed unchanged may result in the...

If the dock has active cooling, it may be that some of the GPU cores are power-gated off in portable mode, but turned on in docked mode. That would allow power to scale much more than simply by changing the clock speed. GPU cores may be power-gated off in portable mode for both battery and...

I just checked out a Wii demo kiosk at my local Gamestop. One thing that kind of baffled me was that although the kiosk is set up with a nice little Samsung HDTV, it is clearly in 480i instead of 480p mode. I wasn't able to change to 480p in the settings, which suggests that the kiosk only has...

Its not the CAS latency that I'm talking about. Acessing a DRAM (GDDR3 included) is complicated. The memory is orgainzed into banks, rows, columns. You access a memory location by first "opening" the associated bank with the associated row. Then you can do consecutive burst accesses to columns...

If its not a discrete component on the motherboard, the 24MB of 1T must be embedded in Hollywood. That would explain the increase in GPU die size despite the smaller feature size (90nm vs. 180nm) and the (supposed) lack of new GPU features. 1T is essentially DRAM with some additional banking...

So is there still an NDA in effect or something? Now that its launched, couldn't a developer simply step forward and describe the differences between Hollywood and Flipper item by item?

I'm assuming this starts in Japan first? If so, should we be getting news on the East Coast by tomorrow night?

If the 24MB 1-T SRAM is really integrated on the Hollywood die, that would explain everything for me. Flipper was ~51M transistors on a 0.18 process. About half was logic and half was 1-T. For the sake of argument, lets say that a 0.09 process allows 4x the number of transistors with the same...