They are still using GF 14nm on their IOD due to Wafer Supply Agreement. We will have to wait and see how they fit their iOD with PCI-Express 5 and DDR5 with Zen 4.
And I cant wait to see Netflix plays around with these in their FreeBSD box.
It is interesting the biggest upgrade from Zen 4 won't actually the Core part but the iOD. As I am expecting Zen 4 to be some tweaks and a Die Shrink to 5nm. TSMC also somehow unexpectedly announce doubling their 5nm capacity with new Fab capacity being built. My guess would be an aggregate demand from Apple and AMD exceed certain threshold to be worth doing.
5nm will be Zen 4 which should bring 10-20% IPC uplift if AMD's current trend continues.
TSMC's N5 5nm transistors are 85% smaller than their N7 transistors which should lower power consumption significantly though SRAM only shrinks a modest 35% (this especially affects desktop Ryzen with tons of cache compared to their laptop versions).
AMD currently makes the Zen 2/3 IO die on Global Foundries 12nm for contractual reasons. When they finally shrink that to 7 or 5nm, the power savings should be significant.
Zen 4 is expected to bring DDR5 support which will both drastically increase bandwidth and lower RAM power consumption. Likewise, it is expected to support PCIe 5 which doubles the bandwidth per lane to a little shy of 4GB/s.
All of these things together could mean a decent improvement in IPC and total performance and a very big improvement in performance per watt.
Meanwhile, I suspect we'll start seeing large "Infinity Cache" additions to their APUs that is shared between the CPU and GPU as the bus width of DDR just doesn't offer the bandwidth to keep larger GPUs from fighting the CPU for bandwidth. This should not only improve APU total performance, but fewer trips to RAM has a significant effect on power consumption (it costs more to move 2 bytes than to add them together).
Sure, but it's not Intel 3nm. AMD is moving to TSMC 5nm and will go to 3nm in the future as well. It really bothers me when either side compares tomorrows products to competitors todays products.
I've heard for a long time that Intel 10nm =~ TSMC 7nm. This is something they probably just have to do to avoid the perception of being further behind than they actually are.
With how much they've been struggling with 7nm, doesn't look positive. Zen 3 is already 7nm, a significant leap forward and TSMC and Samsung are already messing around with 3nm.
I mean, 2022 for their first 7nm chips to hit market? That's crazy. Zen 3 hit the market two weeks ago.
>Intel is off by 4-5 years according to their initial estimates (in 2011) of hitting 10nm
In 2011 they were expecting 10nm in 2016. So it was not really off by 4 - 5 years. More like 3 - 4 years.
>TSMC's chief technologist said Tuesday (March 12).
Somewhere along the line TSMC's node naming no longer follows Intel's node naming. So that 9nm they mention is more like today's TSMC 5nm. So if you look at 2024 for TSMC's 5nm it isn't too far off. TSMC is planning to have 5nm in 2020. The 4 years speed up has been from heavy investment of Mobile SoC and industry scales changes that were not foreseen at the time.
That’s just marketing fluff in terms of measurements. Most agree that in terms of gate size, intel’s 7nm is similar to TSMCs 5nm. So you could say that Intel moving to 5 would give them a 30% boost but the TSMC goes to 3nm.
So there is a weird factor here that we only know what intel tells us about 10nm. The actual parts that have shipped on 10nm have not been competitive with AMD parts at 7nm, or Apple at 5nm.
When Apple/AMD were shipping Intel beating parts folks were quick to dismiss it as a byproduct of TSMC's process. Is intel's chip design so much worse that AMD can trounce them on a substantially worse process? Or is Intel 10nm just not very good.
We know real Intel's 10nm / TSMC 7nm has finished and is matter of yielding.
We know TSMC Intel 7nm / TSMC 5nm is pretty close to complete. This is coming out to market in roughly 2020.
We know 3nm is coming in 2022 / 2023.
But what comes after 3nm?
Will we need some material science breakthrough? Process and Material that can run at 10Ghz with the same power usage.
More transistor hasn't given us more performance. IPC, Core Count, Clockspeed, Special Instruction Set (Its funny how we swing from RISC to CSIC again ), and larger cache. It seems we have reached a plateau where we cant have more performance from CPU Hardware. GPU is different since it scaled very nicely with transistor count, and is more limited by bandwidth.
And fast, simple, high performance, easy to programmed for Programming languages + framework hasn't really come along.
But cost for Fabs, Wafer and Designing is rising.
Or have we reached a stage, performance no longer matter for majority of people?
reply