Discussion in 'Article Discussion' started by bit-tech, 27 Apr 2018.
Intel is really struggling with 10nm ain't they, not that i blame them or anything, it just shows how damn hard engineering at that scale is and makes me wonder if those problems will be overcome or just end up costing more.
After having the market monopoly for so long, personally I hope it costs them a bloody fortune - it's not like they've not had enough consumer funding over the years and milked it for all it's worth.
When i mentioned costing more i was thinking about us consumers as the difficulties Intel are facing aren't going to be exclusive to them, there problems of physics, and will be faced by everyone.
And before anyone says how GloFlo hasn't had any problems with their 7nm or Samsung, or *insert A N Other company here*, the size fabrication companies attach to the size of their nod's lost most of their meaning years ago.
When you consider that's a transistor less than 1/3rd the size of a Sandy Balls transistor it's hardly surprising.
I can see a future where process stops getting smaller and chips just get bigger, leading to hilarious chips the size of an A5 sheet of paper...
Thats where glue
mcm comes to the rescue.
Next step is carbon, actually. But that will be a while away. And also, they will still have to make the chips bigger to put more into them.
I hope that Intel has a lot of trouble and invests a boat load of money into RnD.
The ideal result is that a breakthrough is needed to overcome the obstsacle which will result in significant performance incease.
I can dream right o/
I find it a bit weird that people are gleefully wishing for Intel to struggle in their R&D when it comes to lithography. It's not like Intel having trouble means that we're magically going to get better products from everyone else, and whilst competition is good if it's a race to the top, it's less good if it's a race to the bottom.
I get that people have a desire to see the underdog succeed or the big boys struggle, but at the end of the day I want to see the pace of improvement in technology be faster for everyone rather than slower for the big boys so that the others can catch up.
As @Corky42 says, it looks like continued node shrinks are becoming exponentially more difficult. If the likes of Intel are finding it hard going then they won't be alone.
I honestly think it's due to how Intel acted when they had no competition after doing dodgy business practices to achieve that.
The more intel spends the more we the consumer pays in the end. If they do develop some super chip expect it to be super expensive.
Cannot really compare process size from different manufactures. So would not go down that route.
It's precisely THAT ^^^
Even more recently a certain Intel CEO dumping as many shares as physically possible prior to the announcement of certain vulnerabilities with Intel products - that certainly hasn't helped any either.
Yeah, I know why people dislike Intel. But hoping that Intel do badly won't magically make everyone else do better, so the end result will be that overall progress will be slowed, and as @rollo says the more that has to be spent on R&D then the more those costs will be passed on the consumer. Having a more level playing field isn't necessarily a good thing if overall quality or innovation suffers.
I'd rather see Intel continue to make progress with 10nm and beyond and for everyone else to also make similar progress. If Intel are toiling with 10nm then it's probable that everyone else will too.
It just means they'll need to milk 10nm like they have 14nm, partly to recoup the money they've had to chuck at it... and they'll probably need to eke 10nm out like they have with 14nm anyway as i can't see 7nm being any easier to realise.
I am quite pleased about it for a couple of reasons.
The first is that it drives a coach and horse through the conspiracy theory that Intel has several years worth of new products just sitting around waiting to be released. For all their other skeevy acts the idea that they had an insurmountable advantage they just weren't bothering to release is a bit silly.
The second is that it with transistor size proving difficult to shrink other ways have to be found to improve the product. Just like the inability to make single cores run ever faster led to multi core CPUs, inability to make things smaller makes power, thermals, multi threaded performance and so on much more important. When the obvious upgrade path reaches it's limit we get upgrades in other directions and that usually leads to a better product overall.
Wasn't the core thing due to AMD not them not being able to make improvements. They kept us at 4 core for ages while they could of easily took us to 6.
AMD was running rings around them, so I don’t believe so...?
It was a clock speed thing, IDK the timelines involved so can't comment on it being an AMD or Intel thing but both of them discovered, at some point, that they weren't seeing the big increases in the switching speed of transistors that they had seen in the past.
What i find odd is that AMD released the dual core Athlon 64 X2 (presumably because they couldn't get much more performance from a single core) and yet still made the same(ish) mistake as Intel by going down the route of a Netburst type design that focused on clock speeds.
If intel struggles they will increaase RnD funding which could lead to new innovations. Intel has the case to spare so nothing to worry about. I think most would prefer 10nm products today, but if now, spend more in RnD and we will get better prducts in the future.
It is also giving AMD a chance to make some bank to fund the RnD needed to remain competitive.
Yes, but going back further than that multi core CPUs came to consumer markets because 4 to 4.5GHz is the practical limit for a retail processor on consumer grade cooling to run reliably. Around 2004-5 retail processors hit that speed limit (when overclocked) and the only way to make a worthwhile bump in speed was to put two cores on one chip. So in came the Athlon X2, whilst Intel did weird things with Pentium 4s then released Core 2 a bit later.
Before that every computer I ever bought was advertised solely on being X-Mhz faster than the previous one.
Separate names with a comma.