Intel’s legacy is eroding • The Register

Comment Intel CEO Pat Gelsinger’s carefully assembled house of cards is collapsing around him. And it’s not so surprising when you look at the hand that was dealt to him.

For those who haven’t been following Intel’s product roadmap so closely, here’s a quick recap of the current situation. Society is stuck in a 10nm aging process; its desktop processors are ridiculously power hungry; its dedicated GPUs aren’t particularly competitive, even when their drivers are working; its next families of processors are hopelessly overdue; and it just killed off Optane, which was arguably its most promising development in recent memory.

Of course, none of this is Gelsinger’s fault. All of these projects were well underway long before he agreed to return to Intel early last year. Even the decision to suspend Optane had more to do with Micron’s withdrawal from the arrangement and a lack of demand than anything else.

But the fact remains that these are his problems now, and they seem to be catching up with him. Worse still, investors are no doubt losing patience with the x86 giant after the company’s disastrous second quarter saw Intel post a net loss of $454 million for the three-month period.

In response to the dismal results and even more pessimistic outlook, Gelsinger issued this statement: “The results for this quarter fell short of the standards we have set for the company and our shareholders,” he said. “We must and we will do better.”

But how can a company like Intel do better when it faces falling demand across nearly every business unit and the stiffest competition from rival chip giants in more than a decade?

Perhaps the most important question for Gelsinger is how do you inspire confidence in a company that can’t ship a product in time to save their life?

Intel is falling behind

Intel is a massive organization filled with brilliant scientists, electrical engineers, and software developers. And yet, despite all that talent, the company is falling behind rival AMD and facing a growing threat from Arm processor rivals, which have gained a foothold in the lucrative cloud segment in recent years.

Intel’s Sapphire Rapids Xeon Scalable processors are arguably its most ambitious chips. In a single generation, Intel committed to not only deliver DDR5, PCIe 5.0, and CXL to the data center market, but that it would use a multi-die chip architecture that would close the core count gap with AMD. .

Based on the performance we’ve seen from Ice Lake, the chip would have been a formidable challenger for AMD’s Milan-X had it been delivered on time.

However, it seems trying to perfect what took AMD years to complete wasn’t as easy as Intel thought it would be. And from mid-2021, the company started pushing the launch back first to Q1 2022, then 1H 2022, then 2H. And this week it was revealed in NDA documents obtained by Igor’s Lab that the chip may not be ready until the first quarter of 2023.

Also, the docs indicate that Intel struggled to get the thing to work, with the chip yet to ship being in its 12th – twelfth stage – which I’m sure is rather atypical for a company at the level of Intel. This means that the server’s processor has already gone through a dozen revisions, from A0 to E5, and still hasn’t launched.

Considering what’s at stake, it’s understandable that Intel goes to such lengths to make this work properly.

However, it’s hard to ignore the absence of Sapphire Rapid as companies like AMD relentlessly push forward. The House of Zen has enjoyed bumper sales of Epyc server processors in the absence of Sapphire Rapids, and is expected to release its Ryzen 7000-series desktop, fourth-generation Epyc data center processors and GPUs this year. RDNA-3.

If the reports of Sapphire Rapids’ latest delays are true, this is going to create an even worse comparison, especially among customers who have been putting off their server refresh in anticipation of DDR5 and PCIe 5.0 compatible systems.

But it’s not just Sapphire Rapids that Intel seems to be struggling with. A recent report from TrendForce indicates that the GPU tile for its upcoming Meteor Lake processors may experience more delays.

The firm reports that the GPU tile has already been pushed back due to “process verification and design issues”, but the reasons for the latest delay remain unknown.

Speaking of long-delayed products that only recently started shipping, Intel’s dedicated Arc GPUs recently became a point of contention over the chipmaker’s insistence on shipping them with unfinished software and drivers and not functional.

It doesn’t help that the cards fare much the same as Sapphire Rapids from a competitive standpoint, with gaming stalwarts AMD and Nvidia slated to release their latest-gen graphics cards this fall, and Intel coming in with a first generation product.

Again, Intel faces a rather unfavorable comparison. Benchmarks indicate mediocre performance, putting it on par with low-to-mid level cards from Nvidia and AMD that are two years old. It’s not surprising – or even so damning – for Intel’s first significant attempt to break into the GPU market. And we know that many gamers don’t need the latest and greatest silicon to enjoy their games; the lower and middle tiers are within their budget and will keep them happy. But that arguably underscores Intel’s failure to deliver.

If its Arc GPUs had shipped even a year ago, amid the worst of the chip shortage and booming crypto mining craze, the comparison would have been much different. All Intel should have done at that point was get a low-end, reasonably priced GPU from its factories that people could actually buy. But Intel, as usual, missed its moment.

Intel 4 can’t afford another delay

Intel’s GPUs have at least one advantage over the rest of its lineup. Intel only has to worry about designing the chips and writing the software; the actual manufacturing is in the hands of TSMC.

That’s not the case for the rest of its product line, which is stuck on a 10nm process that looks increasingly dated with each quarter, and Intel’s brazen decision to rename the process in Intel 7 does not fool anyone.

Much of Gelsinger’s rise to CEO can be attributed to Intel’s failure to bring a working 7nm process to market on time – now called Intel 4. And over the past year and a half elapsed since taking office, Gelsinger has managed to deflect the conversation away from his long drawn out process.

Intel can say top to bottom that Intel 4 is set to ship in 2023 with Meteor Lake, but the company’s track record suggests that’s wishful thinking.

Gelsinger didn’t make it easy on this. Opening Intel’s factories to contract manufacturing was a bold move, but who will buy your chips if your process technology never arrives on time? If Intel wants to steal shares from TSMC for making cutting-edge chips, it needs a competitive process. As it stands, it doesn’t seem to have any.

Intel must break the scheme

To put it bluntly, Intel has managed to squander its lead over its rivals and its huge reserves of talent on overly ambitious projects that have repeatedly arrived late to the party.

And yet, it’s a company that thinks it deserves the lion’s share of the $52 billion in chip subsidies passed by Congress last month.

If Intel is to succeed, it must break with this tendency to deliver too little too late, which means developing competitive silicon in time for a change, even if it means giving up features of questionable value or relevance. ®

About Donnie R. Losey

Check Also

BeReal photo-sharing app explores paid features to avoid ads

Photo-sharing app BeReal is exploring adding in-app payments for additional features to avoid Instagram-like advertising, …