Hacker News Reader

Comment by rwmj | Hacker News Reader

There's a surprising amount of lock-in for corporate customers. They'll be using proprietary software which is only certified for Intel (for reasons which can sometimes be real). They could be using the Intel compiler which as far as I know still pessimizes on AMD. Also it's not necessarily cheaper to move your workloads to AMD if that involves buying new hardware (if on-prem) and retesting everything or having to go to all your software suppliers to check if they support AMD.

Personally I switched to AMD hardware a couple of years ago and haven't looked back, but corporations don't do that.

View on HN

I suspect large customers are going to AMD for a quote, and then taking the quote straight to Intel for them to beat. No transition headache and price benefits with that approach.

But those are on-prem customers, what about cloud customers? In the SOA world, surely greenfield stuff won't be using any of that proprietary Intel software.

Intel's margins are dropping fast, so you're probably right that something is going on.

> Intel said its gross margin, the percentage of revenue remaining after deducting the cost of production, was 55.2%, down more than five percentage points from the same period in 2020. This is a key indicator of the strength of its manufacturing and product pricing. Intel has historically delivered margins above 60%.

https://www.msn.com/en-us/news/technology/intel-falls-most-i...

If AMD are being used to drive down prices, it has a good overall effect though doesn’t it?

And if AMD get the feeling they are being used this way, offering quotes with no margin will make for painful days at Intel.

Intel’s margins are almost certainly higher than AMDs.

Probably they were 4 years ago, but I'm not so sure now. Intel's processors require about 2x as much silicon per core (14nm vs 7nm), so if they are priced similar per core, Intel is probably getting squeezed.

The silicon costs are a very tiny portion of their cost of goods sold.

Electricity, water and even just plain packaging of the cpu will likely cost more than the difference in silicon use between 14 and 7 nm

1. The cost difference between 14mm2 and 7mm2 is a whole new fab at twice the density (assuming transistor count remains the same). Asset costs really matter to Intel.

2. If the constraint on the number of packages you can sell is the number of chips you can produce, then the packaging cost of the chip is not so relevant (assuming packaging is not a constraint on production). If you can halve the chip area on the same production node, you can double production of packages, which can make a huge difference to profits (assuming Intel is a high margin business with high demand and that demand elasticity is in their favour etcetera).

Disclaimer: I am not in the industry, but what you say just seems wrong without even arguing that the cost of the silicon for Intel dominates packaging costs.

Intels 14nm fabs and tooling are now fully amortized and yields are going to be very much higher than 7nm/5nm nodes.

193i steppers are also dirt cheap, and most of their old fabs can also be modified for 14nm/10nm production, whereas EUV tools are 180 tonne behemoths that require overhead cranes and/or physical disassembly of the plant to move.

I was surprised that Intel hasn’t seemed to have bought more of the EUV machines from after they helped fund their development. Do you have any insights?

You're assuming that the cost of a fab (machinery, labour, whatever) is proportional to wafer size rather than transistor count. That seems pretty dubious.

The difference isn't in material usage. Larger dice translate into lower yield per wafer and less throughput for a fab.

Own fab vs outsourced fab. It's significant difference how they calculate the cost. Reducing die size is more serious for AMD but not so much for Intel.

I think you’ve hit the nail on the head here. I also think you may have answered your own question.

Cloud customers may be taking volume quotes to Intel from AMD to see what they can/will do for them on price. I don’t see why they wouldn’t do that, what with their (way out of the normal range) buying power.

Cloud is skipping amd and going straight to ARM.

I suspect AMD is getting a bigger chunk of a smaller pie as ARM makes headwind.

AWS is the only large scale provider with ARM. There are some smaller scale and VPS options but those are more for personal or small biz workloads or low overhead things.

We moved stuff to AMD cores on Google Cloud and saw a roughly 15% reduction in utilization (and thus cost) for the same work load. Those are just Zen2 too, not Zen3 yet.

Who is currently offering ARM VPSes to small businesses? I only knew about Graviton

I stand corrected... I no longer see any. For a while Scaleway and a few others were offering them but I checked and they're gone. They were too early IMHO.

Who except for Amazon is doing this currently? Or are you forecasting a long term trend.

Google would not be talking about it, but they've been rolling their own silicon for more than a decade.

Google was using POWER9 internally a couple years ago. Probably not cost efficient right now.

Exactly. I moved some workloads to AMD, but with Graviton2 out now I’m going straight to it where I can.

Reply on news.ycombinator.com