in worst case the console can deliver 9TF of compute power
"It's easy", then proceeds to present as truth unsubstantiated bullshit. lol
It's no harder then understand, it's between 9 and 10TF. That's a fact.
Actually is very simple, 9TF is a drop of frequency of 13%, which is nowhere near what Cerny has said (oh wait Cerny is a liar right).
Not to mention that the GPU can be kept at maximum frequency at expense of the CPU.
I'd say your statements can be catalogued from "bullshit" to "misleading".
Well, between 9.2TF and 10.2TF then. Just like in case of XSX is called a 12TF console while it's 12.115 or whatever.
How much of the time it spends and when is a guess, but it's a 9.2TF console at worst, and 10.2TF at best.
Cerny aint a liar, but he isn't going to promote the worst case scenerarios for his companies products either.
And the other way around. It's easier to gain power from the GPU though.
Well, a 0.2TF difference as bs lol. There you go. Maybe keep things civil while you're at it, if you want escalations you will get them.
Do you think, that if someone promotes his product he will tell everything that might be "negative" (or not so good) about his product?So he is liar.
No. He stated that it was difficult to guarantee those clocks without restricting power limits specifically. He made it quite clear that the clock limit of the GPU was logic/timing driven. Not contingent on the limitations of power, thermals, or SmartShift.Do you think, that if someone promotes his product he will tell everything that might be "negative" (or not so good) about his product?
No, he will always tell you about the positive sides and maybe some not so bright things but that are still positive or that have a "simple" workaround. That is just not lying, this is just simple PR.
Just, always take such events with a grain of salt.
Even developer events are PR. Well at least all tech-events that I was present as a developer were also meant to promote something you should use/buy/sale/rent/pay, ...
Btw, Cerny stated that it wasn't possible to even get a 2GHz stable frequency without those things they have done. So you can assume, that the frequency will also drop below 2GHz at times.
We don't know under which circumstances or if this will happen often. That's why we a in a speculation forum
Actually is very simple, 9TF is a drop of frequency of 13%, which is nowhere near what Cerny has said (oh wait Cerny is a liar right).
Not to mention that the GPU can be kept at maximum frequency at the expense of the CPU. So yeah, I'd say your statements can be catalogued from "bullshit" to "misleading".
Well what is "brute force"? "Brute force" has such a negative meaning.No need to be worried, even if it drops sometimes or often to the 9TF range, we've seen a glimpse of what it can do, and it's already looking really good.
They focused on other bottlenecks instead of brute force, and i'm sure it'll pay off enventually.
Are we still counting Tflops?
I think that TFlops as a measurement is as useful as using MIPs for processors. It was only ever informative when comparing apples to apples.
Pretty sure they've already been counted.Are we still counting Tflops?