| Errorist76 said: http://cpu.userbenchmark.com/Compare/AMD-Ryzen-7-1700-vs-AMD-Athlon-5350-APU-R3/3917vsm10020 Ok, it’s more like 400%....my bad. |
Still. You were close enough.
Although... Those are synthetics, not real-world so your mileage will still vary.
| Errorist76 said: So you’ve chosen to completely disregard my arguments for price gap and leaving behind base console users...?! I wonder why. |
Well. Mostly because I don't care about pricing. Obviously.
I am a little biased in this aspect because I own a dozen consoles in conjunction with high-end PC's and thousands of games across them all, $100 is stuff all money in 2017 at the end of the day.
But... We also need to remember the Xbox One X is also marketed as a Premium product, it is meant to have a higher price.
The Xbox One S is the more budget orientated product and will push the lower price... That is generally why companies have multiple products in a product lineup, so they can hit various price points and the needs/wants/desires of as many consumers as possible to make a buck.
| Errorist76 said: Considering that vid..I’ve read the according DF article about that intelligent delivery a while ago already. It’s a concept in development atm , which they will mostly use for their excessively huge first party titles in order to shrink their sizes. |
It's more than just a concept, it's in active development.
| Errorist76 said: why would anyone understand he has to download 4K assets on his base X1?! It’s crazy X1 users even had to download those updates. |
You don't have to download 4k assets on an Xbox One. That is a false assertion on your behalf.
You have the option to download 4k assets on the Xbox One so you can migrate them over to an Xbox One X at a later date.
| taus90 said: The only issue I could think of is LPDDR rams bandwidth but again with a closed platform and a fast SDXC card or an onboard nvme solution those can be easily offset. |
SDXC/NVMe isn't going to offset memory bandwidth. Not in a mobile device.
Besides... NVMe solutions typically top out at 3GB/s... Which isn't going to make much of a difference at the end of the day when Ram is easily 10x that in the super low-end these days.
Also NVMe @ 3GB/s requires power, not something that is ideal in a mobile device.
Not to mention that NVMe drives obtain such transfer rates thanks to the use of a ton of memory chips, memory operations can be highly parallel.
SDXC is even more depressing. Most SDXC cards top out at around 100MB/s.
Some premium cards can do 280MB/s... But all pale in comparison to the 20,000-60,000MB/s that ram can potentially offer in a mobile device.
| taus90 said: And this is not even on 10nm, even a mass produced snapdragon 835 is outputting 500+ GF at 2k resolution, bring that too 1080p it goes even higher. |
Not entirely sure I understand your reasoning.
Are you suggesting a device gains more "GF" the lower the resolution goes?
| taus90 said:
Again apple with its A11 chip has showcased that very same thing which you are skeptical about, infact A11 is derived from the PSvita hardware successors which i mentioned in my previous post.. an in house custom design 3 core GPU, based on PowerVR cores and for CPU 4 high power cores and 2 low power cores and that think is outperforming a 2017 mac pro with i5. I am an iOS and Android Developer and the things A11 can do really makes me wish sony should consider a vita2. And if u talk about the price of iphone 8, remember we are talking about apple here.
|
Apple seems to be ahead of the industry with it's mobile chips... And they really are not directly comparable to the competition...
Mostly because Apples ARM cores tend to be extremely wide along the lines of an Intel Core chip.
Would it be out-performing a Core i5 that can turbo to 3.5-3.6Ghz though? Hell no.
But it's still a potent chip all things considered.
| taus90 said:
Also if Sony and MS can redesign a far worse Jaguar APU and produce the result which we have on PS4... u think they cant redesign a far better and more enegy efficient ryzen apu to fit onto a tablet size device?? Bottom line is a system designed around close platform and ability to code directly to hardware will always be better than general purpose device based on the same hardware.
|
No one codes directly to the hardware anymore, no one writes in Assembly/Machine in 2017, there is no need for it.
But the good thing about that is... It's far easier to retain backwards compatibility and develop software and games.
With that... A Pascal based Tegra can provide 50% more performance for the same power over the Maxwell Tegra (Aka. The chip in the Switch).
And provide 128% more bandwidth. (The biggest game changer for performance at 1080P verses the Switch.)
A Volta based Tegra has already been demonstrated for a few markets like cars... And is set to double the Pascal based Tegra again and is using 16nm at TSMC.
There would be power/performance gains if nVidia shifted Volta Tegra to "12nm" as well of a around 10%...
In short, I doubt we will get a chip that is a Playstation 4 in identical capabilities for many years yet, but we can get "close enough" with hardware available today/soon.

www.youtube.com/@Pemalite














