Live data from Hacker News

AMD CEO: The Next Challenge Is Energy Efficiency

spectrum.ieee.org

301–310 of 328 posts

Re: AMD CEO: The Next Challenge Is Energy Efficiency

#301
post #83

Earlier quoted context omitted.

I'm not up-to-speed on modern chipsets, but WTF is the Southbridge doing that it needs that much power? Is Thunderbolt going through it or something? I think of the Southbridge as a mostly ignorable part of the chipset (up until something goes horribly wrong).

Single word: PCIe. Lots of fast PCIe lanes eat a lot of energy. A server motherboard contains tons of more PCI devices when compared to a consumer desktop systems, and they are not the cards, but the small units enabling the advanced features in servers, which are embedded on the motherboards themselves. Thunderbolt is just a PCIe encapsulator of some sort, which can also do plethora of other things.

Oh, I assumed that PCIe was on the northbridge. Of course PCIe eats energy.

Re: AMD CEO: The Next Challenge Is Energy Efficiency

#302

Su is just talking up AMD’s strengths, many of which come care of TSMC and for which the original R&D was largely funded by Apple. She is not wrong, and AMD has certainly made large gains in HPC recently, but AMD does not monopolize all possible paths to success here.

5nm? It was also funded by AMD since it is their customer, and by many other customers.

Re: AMD CEO: The Next Challenge Is Energy Efficiency

#303
post #192

Earlier quoted context omitted.

I think you mean "Do not buy Windows computers". Macs are also PCs.

To be fair that whole boat sailed with the "I'm a mac he is PC" series of adverts.

Ah, I'm not in the US so I avoided those ads.

Re: AMD CEO: The Next Challenge Is Energy Efficiency

#304
post #193

Earlier quoted context omitted.

That's interesting to hear. I've not used Windows in over a decade (I primarily work on mac) but I've heard that WSL was quite compromising and not a great experience for people who primarily use Linux tools.

bash (and other shells), coreutils, pipes, git, text/cli utilities etc. work just fine on WSL, I'd call them Linux tools. My in-shell workflow consists of using mainly those + VSCode (with the WSL plugin) + ssh'ing somewhere now and then, and it's entirely sufficient for this purpose. I haven't tried running typical webserver/db services on it though.

I’ve run 120 containers in native docker on it and it worked fine :)

Re: AMD CEO: The Next Challenge Is Energy Efficiency

#305
post #101

Earlier quoted context omitted.

Yes WSL2

WSL2 is great if you're stuck on Windows, but it's still really not there yet. For example: https://github.com/microsoft/WSL/issues/8725

Better to run docker native than docker desktop. Doesn’t leak.

Re: AMD CEO: The Next Challenge Is Energy Efficiency

#306
post #101
post #99

Earlier quoted context omitted.

> The irony being that Microsoft seem to make the best Linux OS for development WSL?

Yes WSL2

Controversially, I have a better development experience on Windows using msys2 + zsh (basically "git-bash" on steroids). I would put that development experience almost on par with MacOS.

WSL2's virtualized workflow just causes too many issues for me. WSL1 was better IMO but it wasn't significantly better than msys2 and also had issues (like you still need remote development tools to mount codebases inside editors) - unless you want to run/develop Linux binaries while on Windows.

For anything that isn't making basic non containerized applications (simple web applications, web servers), Windows is pretty good.

For anything more involved, requires multiple containers/compose/etc, I prefer Linux as it has the tools I need available natively and no gotyas or performance penalties.

That said, credit to Microsoft on WSL2. The auto-scaling hardware provisioning inside the VM has made containerized workflows on Windows much better. To me, it's just not better than running Linux inside VMWare/Hyper-V/VBox and "DIY"ing WSL2 yourself, something I had been doing for years before WSL2 anyway. WSL2 is more fool-proof then hand-rolling a Linux VM, so there is that.

Re: AMD CEO: The Next Challenge Is Energy Efficiency

#307
post #83

Earlier quoted context omitted.

I'm not up-to-speed on modern chipsets, but WTF is the Southbridge doing that it needs that much power? Is Thunderbolt going through it or something? I think of the Southbridge as a mostly ignorable part of the chipset (up until something goes horribly wrong).

Single word: PCIe. Lots of fast PCIe lanes eat a lot of energy. A server motherboard contains tons of more PCI devices when compared to a consumer desktop systems, and they are not the cards, but the small units enabling the advanced features in servers, which are embedded on the motherboards themselves. Thunderbolt is just a PCIe encapsulator of some sort, which can also do plethora of other things.

The one contradiction I have to point out here is that server motherboards dont need big southbridges: the cores themselves have gobs of PCIe. 1 and 2P AMD Epyc server cores have 128 lanes of PCIe.

I wish modern chips did a better job of breaking down where power went. It'd be so interesting to know how much power is going to usb controllers, how much is going to PCIe. I'd also hope that they could do things like shut down parts of the chip, if there's no USB or PCIe devices plugged in. But these chips seem to have a pretty high starting place of power consumption. Although maybe it's in part because the first example were flagship motherboards with a whole bunch of extra things peppered across the board - fancy NIC chips, supplementary thunderbolt controllers, sound cards, wifi - so maybe there was just an unusual lot of extra stuff going on. But it has been shocking seeing idle power raise so much on the modern platforms. It feels like there's a lot of room for improvement in power-down.

Re: AMD CEO: The Next Challenge Is Energy Efficiency

#308

This is not a completely hashed-out thought. But I'll share it and see what others think. My impression is that the simplest way to improve energy efficiency is to simplify hardware. Silicon is spent isolating software, etc. Time is spent copying data from kernel space to user space. Shift the burden of correctness to compilers, and use proof-carrying code to convince OSes a binary is safe. Let hardware continue mana…

I doubt it. Special purpose hardware is usually more efficient than a software implementation running on general purpose hardware.

But in this case the job of the hardware is to prevent the software from doing things, and it pays a constant overhead to do so whereas static verification as integrated into a compiler would be a one-time cost.

Re: AMD CEO: The Next Challenge Is Energy Efficiency

#309
post #117

Earlier quoted context omitted.

"And it runs Office better than Linux or Mac" Yup, been doing that for decades. It's called "vendor lock-in". And its unethicality has been discussed comprehensively for decades as well. And when countries tried to move into an open document format - they'll bully the country, using the strong arm of the Uncle Sam, until they're back to MS Office again. So when people are amazed by Bill Gates' charities - I don't. Hi…

I understand that corporations are bastards and built on foundations of the crushed skulls of children and involuntary human sacrifice but Office on the Mac is not a bad product and is improving. But the windows version has just been around longer and had more work done on it. And don’t get me started on LibreOffice - it’s buggy as hell. Even more than Office on a bad day I’ve met Bill. He wouldn’t be out of place am…

I wasn't talking about the product - I was talking about MS using the product to lock the whole world into its own =proprietary= format (so no one could reliably open & process it - and then others got blamed for it, not Microsoft), AND then aggressively attack those who try to escape from its lock, even countries.

Re: AMD CEO: The Next Challenge Is Energy Efficiency

#310

Earlier quoted context omitted.

Been there done that since the era of Powerbooks - and it always causes my laptop's battery to deplete much faster. Because I'm using my laptop to, well, work on the go a lot; this is a showstopper.

From my experience, the newest VMWare versions and Linux kernels are very good at conserving power when the VM is mostly idle. I'm doing on that my 2014MBP, and it didn't cut the endurance to half, but need to re-test it for exact numbers. However, it doesn't appear in "power hungry applications" list unless you continuously compile something or run some service at 100% CPU load. Also, you can limit the resources it…

That's so good to hear. Very well done to everyone involved.
Post reply on HN