Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Intel announces retirement of Pat Gelsinger

intel.com

Post reply on HN
Print view View original

291–300 of 777 posts

Re: Intel announces retirement of Pat Gelsinger

#291

Post by Cumpiler69 »

bhouston wrote: post #172
Cumpiler69 wrote: post #128

Earlier quoted context omitted.

>they could have a line of ARM CPUs now to compete with the onslaught in all major CPU markets. Disagree. Selling ARM chips with high profit margins is tough. There's too much competition from the likes of Samsung, MediaTek and until the US ban, Hi Silicon(Huawei). ARM chips are a race to the bottom in terms of price with a market dominated by companies from Asia. There's no guarantee Intel could have had a competiti…

> There's too much competition from the likes of Samsung, MediaTek and until the US ban, Hi Silicon(Huawei). ARM chips are a race to the bottom in terms of price with a market dominated by companies from Asia Yes, without proprietary technology, margins are slim. But Qualcomm has somewhat succeeded in this area anyhow. I guess they took the ARM base but innovated on top of it in order to command higher margins.

>But Qualcomm has somewhat succeeded in this area anyhow

It's wasn't just anyhow. Qualcomm succeed in the mobile SoC space because they also had the best modems in the industry (name comes form Quality Communications after all). And also the best mobile GPU IP they bought from ATI.

Re: Intel announces retirement of Pat Gelsinger

#292

Post by Wytwwww »

hedora wrote: post #281
shrubble wrote: post #267

Here are 2 things I have noticed that seem obvious weaknesses: 1. Looking at the CyberMonday and BlackFriday deals, I see that the 12900K and 14900* series CPUs are what is being offered on the Intel side. Meanwhile AMD has released newer chips. So Intel has issues with either yield, pricing or other barriers to adoption of their latest. 2. The ARC GPUs are way behind; it seems obvious to me that a way to increase us…

For that to work, they need a software stack to compete with CUDA. Nvidia is so far ahead that a single manufacturer won’t be able to compete for developers. Instead, the only chance the rest of the AI GPU market has is to build some portable open source thing and use it to gang up on Nvidia. That means bigger GPU players (AMD, hyperscalers) will need to be involved.

> For that to work, they need a software stack to compete with CUDA.

Doesn't Intel have pretty decent support in PyTorch? It's not like most people working on/with AI use CUDA directly.

And especially for stuff like LLMs software is hardly the main or even a significant barrier.

Re: Intel announces retirement of Pat Gelsinger

#293

Post by shrubble »

hedora wrote: post #281
shrubble wrote: post #267

Here are 2 things I have noticed that seem obvious weaknesses: 1. Looking at the CyberMonday and BlackFriday deals, I see that the 12900K and 14900* series CPUs are what is being offered on the Intel side. Meanwhile AMD has released newer chips. So Intel has issues with either yield, pricing or other barriers to adoption of their latest. 2. The ARC GPUs are way behind; it seems obvious to me that a way to increase us…

For that to work, they need a software stack to compete with CUDA. Nvidia is so far ahead that a single manufacturer won’t be able to compete for developers. Instead, the only chance the rest of the AI GPU market has is to build some portable open source thing and use it to gang up on Nvidia. That means bigger GPU players (AMD, hyperscalers) will need to be involved.

I am not a PyTorch user, but there is already Intel-supplied ARC acceleration for PyTorch: https://www.intel.com/content/www/us/en/developer/articles/t...

Having half the number of GPUs in a workstation/local server setup to have same amount of VRAM might make up for whatever slowdown there would be if you had to use less-optimized code. For instance running or training a model that required 192GB of VRAM would take 4x48GB VRAM but 8x24GB VRAM GPUs.

Re: Intel announces retirement of Pat Gelsinger

#294

Post by vlan0 »

shermantanktop wrote: post #248
vlan0 wrote: post #235

Earlier quoted context omitted.

>I feel like this is a mistake. Pat's strategy is aggressive but what the company needs. He's also 63. Has plenty of money to survive the rest of his life. Has eight grandchildren. There's so much more to life than business. What's to say he doesn't want to simply enjoy life with more connection and community to loved ones around him?

That would be a healthy, balanced and long-term-oriented approach. But those who get to the level of CEO are subjected to intense forces that select against those traits. I don’t know much about this guy but it’s reasonable to assume that any C-level exec will hold on to the position for dear life until they are forced out.

Here's more info on Pat. He is not your average CEO. He wrote this book almost 20 years ago. Why do you resort to conversation about assumptions?

The Juggling Act: Bringing Balance to Your Faith, Family, and Work

https://www.amazon.com/Juggling-Act-Bringing-Balance-Family/...

Re: Intel announces retirement of Pat Gelsinger

#295

Post by rwmj »

dagmx wrote: post #115
rwmj wrote: post #80

Earlier quoted context omitted.

Don't forget their "brilliant" strategy of suing their own customers.

Being a customer shouldn’t protect a company from lawsuits. ARM feels they have merit here , just like Qualcomm did when they sued Apple. It’s not that rare in the corporate setting to have suits between companies with otherwise amicable relationships.

The optics can still be terrible. Qualcomm (or more accurately, Nuvia, the company they acquired) produced some stunning chips with almost unheard of battery life for a laptop, and Arm are suing them to use their own inferior designs. They even tried to have end user laptops recalled and destroyed! There's no world where this looks good for Arm.

Re: Intel announces retirement of Pat Gelsinger

#296

Post by pbalcer »

gpapilion wrote: post #137
physicsguy wrote: post #119

Earlier quoted context omitted.

You don't get a CUDA killer without the software infrastructure. Intel finally seem to have got their act together a bit with OneAPI but they've languished for years in this area.

They weren’t interested in creating an open solution. Both intel and AMD have been somewhat short sighted and looked to recreate their own cuda, and the mistrust of each other has prevented them from a solution for both of them.

Disclaimer: I work on this stuff for Intel

At least for Intel, that is just not true. Intel's DPC++ is as open as it gets. It implements a Khronos standard (SYCL), most of the development is happening in public on GitHub, it's permissively licensed, it has a viable backend infrastructure (with implementations for both CUDA and HIP). There's also now a UXL foundation with the goal of creating an "open standard accelerator software ecosystem".

Re: Intel announces retirement of Pat Gelsinger

#297

Post by Wytwwww »

mort96 wrote: post #280
Wytwwww wrote: post #269

Earlier quoted context omitted.

Didn't AMD start making x86 chips in 1982?

That seems correct from some quick Wikipedia reading, but I don't understand what it has to do with anything?

The existence of an imaginary x86 monopoly in the 80s?

Re: Intel announces retirement of Pat Gelsinger

#298

Post by cogman10 »

n144q wrote: post #217
cogman10 wrote: post #145

Earlier quoted context omitted.

> There is so little incentive to move away from x86 right now. > IT departments are not going to stop buying x86 processors until they absolutely are forced to. IT departments are buying arm laptops, Apple's. And there is an incentive to switch, cost. If you are in AWS, you can save a pretty penny by adopting graviton processors. Further, the only thing stopping handhelds from being arm machines is poor x86 emulatio…

Looks like your post is talking about two things -- corporate purchased laptops and AWS instances, which are quite different. About corporate laptops, do you have evidence to show that companies are switching to Macbooks from HP/Dell/ThinkPads?

> corporate purchased laptops and AWS instances, which are quite different.

They are similar. Particularly because developing on a corporate hardware with an ARM processor is a surefire way to figure out if the software you write will have issues with ARM in AWS.

That's pretty much the entire reason x86 took off in the server market in the first place.

> About corporate laptops, do you have evidence to show that companies are switching to Macbooks from HP/Dell/ThinkPads?

Nope. Mostly just anecdotal. My company offers devs the option of either an x86 machine or a mac.

Re: Intel announces retirement of Pat Gelsinger

#299

Post by grecy »

etempleton wrote: post #121
cogman10 wrote: post #107

Earlier quoted context omitted.

> The moat of the two companies together would give the new combined company plenty of time to ramp up their fabs. I'm not convinced of this. Fabs are incredibly expensive businesses. Intel has failed to keep up and AMD spun off their fabs to use TSMC. There is also ARM knocking at the door for general computing. It's already gaining traction in previously x86 dominated markets. The model for US based fabbing has to…

They would have at least 5 years to figure it out before ARM becomes viable on desktop assuming there continues to be movement in that direction. There is so little incentive to move away from x86 right now. The latest Intel mobile processors address the efficiency issues and prove that x86 can be efficient enough for laptops. IT departments are not going to stop buying x86 processors until they absolutely are forced…

> There is so little incentive to move away from x86 right now

Massively lower power consumption and way less waste heat to dispose of.

Literally the two biggest concerns of every data centre on earth.

Re: Intel announces retirement of Pat Gelsinger

#300

Post by cogman10 »

JamesLeonis wrote: post #186
cogman10 wrote: post #107

Earlier quoted context omitted.

> The moat of the two companies together would give the new combined company plenty of time to ramp up their fabs. I'm not convinced of this. Fabs are incredibly expensive businesses. Intel has failed to keep up and AMD spun off their fabs to use TSMC. There is also ARM knocking at the door for general computing. It's already gaining traction in previously x86 dominated markets. The model for US based fabbing has to…

> There is also ARM knocking at the door for general computing. It's already gaining traction in previously x86 dominated markets. I know anecdotes aren't data, but I was talking with a colleague about chips recently and he noticed that converting all of his cloud JVM deployments to ARM machines both improved performance and lowered costs. The costs might not even be the chips themselves, but less power and thermal r…

Yeah, my company is gearing up to do the same. We primarily use the JVM so doing the arm switcharoo only makes sense.
Post reply on HN