I'd love to be a fly on the wall in the board meeting room.
[0] https://www.statista.com/statistics/267873/amds-expenditure-...
[1] https://www.investopedia.com/news/amazons-23b-rd-budget-sets...
I'd love to be a fly on the wall in the board meeting room.
[0] https://www.statista.com/statistics/267873/amds-expenditure-...
[1] https://www.investopedia.com/news/amazons-23b-rd-budget-sets...
I know there's a non-trivial mobile performance penalty, but a fourth gen mid-range i5 I got on eBay for $30 has nearly the performance of the 8750H in last year's Dell XPS (my work laptop). Even when you get away from mobile, compare two i7 chips that were released five years apart for basically the same price, and userbenchmark.com shows just a 21% overall performance difference in them:
https://cpu.userbenchmark.com/Compare/Intel-Core-i7-4790-vs-...
This windows shrinks even further when you look at what overclockers get out of the older chips with fancy cooling setups. But yeah, when else in the history of personal computing has five years of progress yielded such small gains?
I know eventually I'll upgrade my system for real, probably to something AM4-based, but when I saw these numbers, it became a no brainer to just bump up the CPU on an old LGA1150 system I had lying around rather than spend a bunch of money on an all-new motherboard, new RAM, etc.
Intel really has got to be in an internal panic. The last time AMD pushed Intel this hard, Intel went to extraordinary tactics to push AMD out of the market. I wouldn't be surprised if pushing their workers well beyond safety limits was in the current playbook.
There's a funny series of LTT videos from this past fall where they bring up a 2008 Skulltrail mobo and try running some modern games on it, with and without overclocking:
The last big microarch jump for intel was Skylake (6th gen) and then 7, 8, 9 and some of 10th gen (but not all, but they needed "10th gen" for marketing reasons) are just incremental improvemnts over it. So Skylake was about a 15-20% increase in performance (real) but took a few shortcuts (Meltdown and Spectre). From then on, you had a 5-10% performance improvement per generation (let's round it at 7%). That's not bad per se, but it was done out of comfort. The "cursed" 10nm node which never came is also part of the reason why Intel stalled, they insisted on "tick-tock" and were unwilling (or unable?) to deliver a new microarch over 14nm+++, the super refined node (and the extra refinement allowed them to pull the clock higher on each new gen, now alowing those sweet close-to and with this new weird turbo [that seems more marketing than substance] surpass the 5Ghz barrier).
And that's how you come to Intel's situation on 2020: no 10nm node, no real new microarch in the last 5 years.
Edit:
As for why we never saw something like this... I think we never had a situation before where one of the competitors had such an advantage (the microarchs before Zen on AMDs part were bad, and that hit them hard for a long time) that they could pull this duckish moves for so long. For a long time, Intel chose not to improve their offering, because they were so far ahead of competition and they could keep milking the Whatever-Lake microarch family. They'd blame on the difficulty of extracting more IPC, Moores Law coming to an end (or not coming to an end, depending on what marketing was trying to sell).
In the end it came to AMD+TSMC offering superb processors for Intel to slash the prices in half, and then still being crushed performance wise, in desktop, server and now mobile too. And all Intel has to show is AVX512, which after having it gatekeeped for years in the server market, they've released on consumer processors - but hey, nobody's using it because there was no support!
Chips fabbed on 10 nm have been shipping in volume since last summer. The Sunny Cove CPU cores found in Ice Lake chips, fabbed on the 10+ node, average roughly around 16 % higher performance vs. Skylake, clock-for-clock, before accounting for improvements in memory bandwidth. Tiger Lake (fabbed on 10++) will be shipping this year, alongside with Ice Lake Xeons.
Extract from Anandtech[0] comparing Comet Lake vs Ice Lake:
Some of these tests rely heavily on turbo, such as the PCMark tests, and so the Comet Lake i7-10710U can hit 4.7 GHz on the latest variant of Skylake, while the Ice Lake i7-1065G7, despite its higher IPC difference, can only do 3.9 GHz. This means in a lot of bursty workloads (which a lot of business workloads are), the Comet Lake wins and we see that play out.
Again, that means that some chips aren't produced in 10nm, because if they were, they'd end up being straight up slower, despite having an 18% IPC improvement.As I said before: I really hope Intel sorts this mess up and comes along with something that can compete against AMDs offering. Otherwise we might see a turn of the tables, with AMD eventually stalling progress, just because they aren't pushed hard enough by the competition.
[0] https://www.anandtech.com/show/15385/intels-confusing-messag...
They seem to have learned from their 10nm mistake, so they're apparently planning to have + and ++ versions of the following nodes, that could handle production of a new microarch if the next node isn't ready, but only time will tell how that goes along.
[0] https://en.wikipedia.org/wiki/7_nm_process#7_nm_process_node...
[1] https://www.anandtech.com/show/15217/intels-manufacturing-ro...
Last month, I started getting some kernel_trap BSODs so I reduced voltage a bit and also the clock speed down to 4.7ghz.
I don't see a reason for me to upgrade my CPU for another few years unless it burns out or something. The only component I've upgraded is my GPU, from a GTX 670 to a 980 TI.
The newer chips may be great for specific needs, but for gaming and light development work I don't see any reason why I need to upgrade. Sure, a quad core from 2012 isn't the hottest thing anymore, but along with my 980TI it runs all my games at 2k without any issues.
Intel is a very odd company, and for all that there are truly brilliant people there, it feels increadingly like IBM of 2003.
Probably explained by: https://www.youtube.com/watch?v=P4VBqTViEx4
On another point. I'm not sure how much this is true, but I've read that most of AMD's chips are made from the same basic building blocks, and their different CPUs are mostly just a matter of how many these blocks are stitched together along with whatever tuning/other hardware. Such a design is probably very scalable, which simplifies manufacturing and reduces costs. Whereas I heard intel chips have more separate customizations requiring more separate tooling for each of their chipsets, which complicates manufacturing...
The bigger TSMC failure was on 32nm, which had to be killed outright.
Their problem right now is that they have an inferior process and their high margin datacenter parts[1] are in critical danger of facing competition from devices fabbed by TSMC and (to a lesser extent) Samsung.
Their offerings in the market now, though, are still winning and Intel is making a ton of money.
So the net effect of a downturn is that the future market, the one where AMD is competetive and Intel needs to drop prices, is going to shrink in total dollars. That's a win for Intel, comparatively. It means the window into which AMD could have jumped to steal revenue is smaller, and Intel has more time to catch up.
[1] Remember that most of the money in the CPU industry is made in the datacenter. AMD having better desktop parts doesn't affect Intel's bottom line much as long as they remain competitive in servers. (AMD still lags with laptop devices too, which are also a bigger market than desktops).
Uh... because their revenues are down 20%, they have a hiring freeze, and are delaying all capital expenditures? That was my point: there may be technical reasons for people to move away from Intel's products. But the economic reality means that fewer customers will.
Intel is Xilinx + AMD ( CPU + GPU ) + Nvidia ( Competing with CUDA ) TSMC ( And more since Intel does everything but TSMC relying on many of its partner forming an ecosystem ) + Qualcomm Modem / WIFi + AI Chip + Autonomous Vehicle + Broadcom's Networking ( Consumer to Enterprise Networking Controller ) and many many more smaller things that are not listed.
And since the R&D budget happened during the era where current CEO was the CFO of the company. I wouldn't say Intel doesn't know much about it.
And if we are looking at fiscal results, [1][2], Intel is earning exactly 10x the revenue from AMD. So in terms of dollar generated by R&D they are pretty much in line with industry standards. Now of course in reality Intel is earning likely 80% of their profits from ˜50% of the R&D budget purely on Fabs and x86 CPU. The rest of the R&D budget still provides little returns at the moment, but for example it has been a long time that anyone has the potential to challenge Nvidia 's CUDA or GPU/GPGPU dominance. And Intel's current leaked gigantic GPU offering seems to be just that.
Having said all that. The other side of argument is that there is currently no reason why AMD cant doubl, triple or even 5 times the revenue with its existing R&D budget. Their EPYC line is still doing poorly by (my) standards, only at less than 5% marketing share. Their Desktop are doing well in consumer market but not OEM market ( The majority of Desktop Market ). And laptop has barely stated. And that is all due to AMD's ( comparatively speaking ) incompetence in Sales and Marketing. ( You could properly argue that it was Intel's Sales and Marketing doing its job far too well ).
[1] https://www.anandtech.com/show/15747/intel-reports-q1-2020-e...
[2] https://www.anandtech.com/show/15754/amd-reports-q1-2020-ear...
I would argue that fab capacity at TSMC is very much a limiting factor [0], putting AMD in the same spot as Intel by selling practically every working chip produced. This puts a cap on the amount of units sold, and with that a soft cap on the revenue. Doubling revenue at this point would mean doubling prices, which would be very unhealthy for their turnover.
The fab production capacity problem is even worse if you consider that basically all current and coming products by AMD (CPUs and GPUs for any of PS5, laptop, PC, server, the new Xbox) are produced on 7nm or 7P, and therefore compete for the limited fab-capacity available.
[0] https://wccftech.com/amd-7nm-wafer-production-set-to-double-...
Intel performance will always be one step ahead of AMD, intel have faster tech in their pocket waiting to be released.
No, they delayed time and time again because the 10nm process utterly failed to deliver the necessary yields and performance to be viable. It's not some ingenious display of sandbagging - only now in the past year or so has the 10nm process yields risen high enough to make low-power chips viable, and it is still not viable for high-power desktop chips, and probably will never be viable.
>Intel performance will always be one step ahead of AMD
A completely ridiculous assertion, and one that is already largely false at the present moment.
>intel have faster tech in their pocket waiting to be released.
Likely true, but only because of their failure to execute on 10nm. CPU designs are closely tied to the process they are designed to be manufactured on -- especially for Intel, who have never been concerned about being too tied to a single manufacturer (since it is themselves). This tight coupling can be good when the process is good, but the failure of 10nm meant that for years and years those architectural improvements you speak of are unshippable and useless.
Intel is not completely screwed by any measure, but the idea that they have the situation entirely under control and are intentionally not competing with now-superior AMD chips is entirely ridiculous.
I really had never imagined this could be so. Could you expand somewhat on why it is? Thanks.
This just isn't true. When Ryzen first came out Intel was still ahead in many use cases, especially if $$ was no object to you. But after several years AMD is now taking the crown from intel in all areas they compete in. If they really had some magical faster tech Intel would have released it years ago, much less this year, but they haven't.
Instead they announce the same old chips then hook them up to industrial strength cooling units to make it look like they can compete with AMD. This isn't the behavior of a company that is just biding its time.
They've even had processors that basically only shipped on paper.
https://twitter.com/chiakokhua/status/1258393158766886913 https://twitter.com/chiakokhua/status/1258393160784330754
Citation very much needed.
Show me an Intel CPU with PCIe 4.0 or with the performance per watt of the latest AMD Ryzen mobile chips. You can't, because to date they don't exist.
I think if they had anything in reserve it would have been released by now, because AMD is trouncing them in reviews.
After that the story is well known: their microarch were tied to their node, and Intel's 10nm was a failure, and still is not up to what the competition is doing. They started to backport the next microarch to 14 too late IMO. They may release new ones on 14nm; or they may suddenly manage to strongly improve their 10nm (not likely, it is usually very gradual at this stage)
And it is again obvious that even in the last 3 years, they had nothing (on the microarch side that they could produce on their mainstream node) they were able to rush as a fallback to counteract Zen, otherwise why would they not have done it?
Now they are clearly behind on the process, and on the microarch Zen 2 is good enough and does not need to compete much against Sunny Cove, plus if Intel releases a backport of Sunny Cove for their 14nm it may be against Zen 3...
Hmm. I think if Intel had a significant architecture enhancement waiting in the wings we would have seen it by now. They need it badly.
I just updated a 2013 quad core laptop with a late 2019 8 core laptop. Tasks that really use the cores are of course faster. Other than that, no substantial performance increase. 7 years and it is just incremental. Their process issues of late are horrific but there is not much happening architecturally either. Sad to see. They just seem a bit lost.
Meanwhile in a couple of months AMD will be on 7nm+
(However, it should be said that Intel 10nm is roughly equivalent to TSMC 7nm in terms of density. AMD is still ahead but not by as much as the numbers would imply).
https://en.wikichip.org/wiki/technology_node
"Recent technology nodes such as 22 nm, 16 nm, 14 nm, and 10 nm refer purely to a specific generation of chips made in a particular technology. It does not correspond to any gate length or half pitch."
"At the 45 nm process, Intel reached a gate length of 25 nm on a traditional planar transistor. At that node the gate length scaling effectively stalled; any further scaling to the gate length would produce less desirable results. Following the 32 nm process node, while other aspects of the transistor shrunk, the gate length was actually increased."
Nodes mean 3 things: density, elecrical performance (ion, max f, leakage, etc), and processing flexibility (how many supported feature primitives or designs).