Time to lower that expectation some more.
https://www.semiaccurate.com/2020/10/31/intel-might-be-launching-discrete-mobile-gpus/
"We see nothing that this does better than an AMD Ryzen 4000 mobile which is an 8-core device that tends to cost significantly less the Intel solution."
Costs more performs less, winner.
@1nko+17H4QQLy No, you CAN'T even compare Intel's GPU team with the likes of Nvidia's or AMD's.
@5rvw+17H4QQLy When problematic practices and non-practices persists for 12+ years, what makes you think 5 years would suffice?
Raja hasn’t delivered a single competitive product since he was hired. A single product! Since he left, AMD has been doing better. This is Murthy in the making. Hired by Murthy by the way. Another reject riding the gravy train.
On top of it, he has promoted some of the weakest people, his friends and brought AMD rejects. While there is no single competitive discrete GFX product, the useless, waste of space fellow, so called architect (DB) now is senior fellow. And other promotions. These people should have been fired not promoted.
there are senior pe and fellows in ccdo that would not get hired as an entry grade engineer in a real chip development company. the nepotism and clique tribalism there is out of control.
So funny to see Broxton mentioned as a failure by CCDO, which is completely true. It may not have been completely CCDO's fault, but the overpromise and complete underdeliver was definitely a contributor overall to Broxton's failure. You had PE level people doing the most menial work in terms of SoC dev because all they knew from before was their little cog work in Core dev and could not adapt. Promos were given to people who were "in" to the CCDO clique. You have managers who's sole expertise is to lie and backstab.
If anything CCDO is a shell of it's former self since the glory days of Pentium. Broxton was the nail in the coffin. Most of the better design/architect/validation folks have since left CCDO (it is/was called DDG now) to greener pastures, though the clique is still alive and well at different companies in OR.
Understand the shortcomings mentioned behind Broxton failure but that project was over 5 years ago. Key question is has things changed since?
The semiaccurate.com article is pointing out it has not. And this raises the question Is Intel's GPU design competitive today? Tigerlake Xe GPU is pretty good, but this discrete mobile GPU is a head scratcher. Read somewhere that there is a possibility this discrete GPU could work in concert with the onchip CPU Xe GPU to deliver higher performance. If true, that would be a pretty smart move.
Anyone know for sure?
broxton failed because ccdo was and is a massively corrupt organization where people were promoted for backstabbing and politics
Broxton failed because of lack of open mindedness towards third party IP
"I had always considered Intel to be very inefficient, but the many projects I worked on always moved forward and succeeded."
Intel threw money and people at problems, and those projects went into products that were essentially uncontested on the market.
There was something else about Broxton that was different besides the late attempt at SOC. I had always considered Intel to be very inefficient, but the many projects I worked on always moved forward and succeeded. The disorganization and mismanagement (as in No Management) of Broxton was something I had never seen before. It was FUBAR coming out of the gate and the janitor would have been able to see it. When that happens you know upper management is coming completely unglued.
"we should still have a very strong, if not the best, CPU team"
LOL
How good is our GPU team? Is it as good as our CPU team? Put our manufacturing fiasco aside, we should still have a very strong, if not the best, CPU team. How does our GPU team compare with Nvidia and AMD? Do we have enough people? Do we have good enough people working on GPU?
intel was always a corrupt hellhole, broxton was simply the natural result of intel attempting to enter a market where it did not operate with an overwhelming advantage.
After 20 years of successful Intel cpu development I ended up working on Broxton for a year. I mark that project as the start of an accelerated era of flailing and failure. It didn't sound right when I started and was infinitely worse seeing it in person. I left for good.
Is Intel dead completely?
By the way, most of the money on the consumer side is made on mobile.