Announcement
Collapse
No announcement yet.
NVIDIA To Begin Publishing Open GPU Documentation
Collapse
X
-
Originally posted by gamerk2 View PostDisagree. Due to power draw, APU's are never going to advance past entry-level performance. Desktops and high end systems are still going to need dedicated GPU's as a result.
I remember exactly the same arguments about every bit of hardware that's been integrated. How many people still buy discrete sound cards? Discrete ethernet cards? Yes they still exist, but it is a very small minority. The integrated part doesn't need to be better, it just needs to be cheaper and "good enough".
Edit: Seems the new 21” Haswell iMac has an integrated GPU - "In terms of processing power the base $1,299 USD 21.5-inch model trades last generation's NVIDIA GeForce GT 640M for the on-die Iris Pro GPU inside the Intel Haswell 2.7 GHz quad-core Core i5-4570R processor." http://arstechnica.com/apple/2013/09...-and-802-11ac/ Surely an iMac is powerful enough for the typical desktop user.Last edited by chrisb; 24 September 2013, 07:38 PM.
Comment
-
Originally posted by chrisb View PostMy guess: NVIDIA Loses Huge GPU Order Due To Linux Blob combined with NY Times: NSA Should Be Barred From Requiring Companies To Introduce Surveillance Backdoors. Given the revelations about the NSA being able to force US companies to include backdoors in their software, foreign governments would have to be crazy to keep on using closed source binaries from US vendors. There is no way China can trust US software now. Losing an order for over 10 million GPUs is pretty bad, losing every future foreign government order would be a disaster.
The other thing threatening Nvidia is APUs. Intel has an onboard GPU. AMD has an onboard GPU. Integration can result in lower power and potentially higher performance through unified memory. Apple already moved the Macbook 11" and 13" to integrated GPU, both Sony and MS next-gen consoles are integrated GPU... The writing is on the wall for discrete cards in consumer systems. This only leaves Nvidia with the possibility of licensing their core for ARM mobile, or selling high performance cards to select customers. They need to adapt fast because people aren't going to pay the discrete card premium once Sony+MS show the world that cheaper integrated graphics is good enough for the top videogames consoles.
Onboard APU and GPU is going to force nvidia to make their own cpu's. Nvidia have always been about big performance, so maybe this is the start of more open source to come? I honestly think they are doing this in prep for Valves new steambox.
Comment
-
There was some speculation a while ago about AMD drivers that maybe they want to drop making catalyst for GNU. Maybe its the same thing here?
We're in an important moment of GNU/Linux history - X is going out the door and there are 2 contenders waiting to take it's place. For GPU makers it means more work to support the new stuff, and probably more work than it was for X. So maybe both AMD and nVidia look into stopping making their propertiary drivers for GNU, and instead throw the docs + keeping 1-2 guys for supporting the FLOSS developers.
It would be both cheaper for the companies (less people employed and the community does all the work) and the FLOSS community would be happy cause the drivers were free. Everybody wins.
(these are just my thoughts)
Comment
-
-
Originally posted by gamerk2 View PostDisagree. Due to power draw, APU's are never going to advance past entry-level performance. Desktops and high end systems are still going to need dedicated GPU's as a result.
You couldn't be more wrong, seriously...
With each die shrink, the gpu portion of the apu die will become bigger and bigger. I anticipate a percentage like 60-70% a few years ahead.
Plus, there is no real reason why TDPs for consumer APU's have to be in the range of 100W... Why not 150 or 200 watts? If that means no discrete GPU, why not?
Discrete gpus will go the way of discrete fpus. And unless NVIDIA's project Denver works, it is dead...
Comment
Comment