• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

AMD prepares 12 core 24 Thread 5.1Ghz Mainstream CPU

Status
Not open for further replies.
humbug said:
With Zen1 being a brand new architecture and on a 3Ghz process node i do think there is significant room for improvement, where AMD lacks most is not IPC, as i said its a bit behind in low threaded tasks, its not a lot; its actually clock speeds, Intel are a full Ghz ahead in clock speed, that combined with the 5% or so low threading IPC difference is what puts Intel that 20% ahead in most games at 1080P or lower.

Low hanging fruit should get them a 5% IPC gain and on a proper processing node clocks should also be normalised with industry standard, 4.3 to 4.5Ghz on all cores out of the box and 4.7Ghz to 5Ghz overclocks would put the 1600X replacement just about on par with the 8700K, probably beat it in MT Cinebench.

And yes i think AMD did remarkably well to come from nowhere and get this close to Intel's best with the first iteration of the new architecture, Intel have been working on the Core architecture for a decade.

I'm looking forward to Zen2, hopefully early next year so i can swap my i5 out for one sooner rather than later.
Click to expand...
....and then I woke up to find Bobby Ewing in the shower and it was all a dream.
 
Can't remember who but an AMD guy did say in an interview that there was a tangible amount of performance left on the table with Zen. It was their first shot with IF, a new core and new process, they needed a product shipping asap
 
I think we will see this back and fourth for a while with AMD and Intel. Itll take a few releases from either side until we reach the technology point and things will return to normal. Great time for old tech builds to upgrade indeed. Looking forward to seeing what it all brings over the next year or so.
 
The end part is quite interesting.

2.5 Timing
The GF 7nm process is expected in the second half of 2018. The Intel 10nm is already late and I am hearing late 2018 and possibly even 2019 before it enters production. This presents a fascinating change in the semiconductor industry. Intel introduced 45nm, 32nm, 22nm and 14nm in 2007, 2009, 2011 and 2014 respectively. For many generations Intel was on a 2-year process introduction cadence, now they have gone to 3 years and 4+ years and while their scaling at 14nm was such that even at 3-years their yearly scaling pace was unchanged, they are now drifting off of that at 4+ year. It also begs the question of when Intel will introduce 7nm, are we now looking at 2022 or 2023?

In the mean time TSMC introduced 10nm in 2016/2017 and 7nm in 2017/2018 with 5nm due in 2019 and 3nm development underway. Samsung also introduced 10nm in 2017 with 8nm due 2017/2018, 7nm due 2018/2019, 6nm and 5nm due in 2019 and 4nm in 2020. GF is introducing 7nm in 2018 with a shrunk version due around 2019. GF hasn't discussed 5nm yet but I would expect it this decade. With the foundry 7nm processes similar in density to Intel's 10nm process and several foundry generations likely to come out by the time Intel introduces 7nm, I would expect a significant density advantage for the foundries over the next several years.

2.6 Performance
I wish I could offer a performance comparison between the processes but based on disclosures to-date I can't. Intel is focused on microprocessor performance and GF and the other foundries are more focused on the mobile space and power for performance, beyond that I don't have anything definitive to say on performance.

3.0 Conclusion
Comparing Intel's 10nm process to GF's 7nm process they are more similar than they are different. Since both companies are solving the same difficult physics problems this is in some ways not surprising.

The surprising part in my opinion is that GF at 14nm stumbled so badly they had to license it from Samsung. Now they have an internally developed 7nm process that matches up well with Intel's latest 10nm process.

It is also surprising to me to see how far Intel has fallen from the process lead they had. First with HKMG by several years, first with FinFet by several year, I suppose they are still first to do cobalt interconnect but in terms of process density the foundries have caught them and appear poised to take a substantial lead over the next several years.

With Intel offering foundry processes and GF, Samung and TSMC all offering leading edge processes the industry now has four viable leading edge process options.
Click to expand...
 
Particularly the second from last line.

"It is also surprising to me to see how far Intel has fallen from the process lead they had. First with HKMG by several years, first with FinFet by several year, I suppose they are still first to do cobalt interconnect but in terms of process density the foundries have caught them and appear poised to take a substantial lead over the next several years."

I do think that's down to Intel just re-releasing the same architecture rather than actually bringing something new and original to the table. We all know they had to be dragged screaming and kicking to release anything more than quad cores. In the determination to milk the market for every penny they can, they completely lost sight of the fact that they are a Tech company and Tech company's only ever stay at the top by innovation. With nothing really new to offer, they may well just end up the same as Kodak did in the 1990's as i said in another thread ages ago now.
Peeps ridiculed me for saying that by pointing out Intel is so large and cash rich that will never happen. Well, obviously they are looking for savings otherwise "the bribe" that is the Insider program would not be being slashed. Kodak did the same thing when Fuji started eating into their market share and profits as well............by the way, at the time Kodak were a much bigger company than Intel is now.
 
Status
Not open for further replies.
Back
Top Bottom