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

Nvidia to Acquire Mellanox for $6.9B

nvidianews.nvidia.com

Post reply on HN
Print view View original

221–229 of 229 posts

Re: Nvidia to Acquire Mellanox for $6.9B

#221

Post by orbifold »

dnautics wrote: post #118
orbifold wrote: post #63

Earlier quoted context omitted.

My GTX 1080 works flawlessly with Linux, as has any other NVIDIA graphics card I've ever owned (GTX 680, 480). The only time I tried an AMD card it was a complete dumpster fire, nothing worked (the open source driver at the time sucked and the proprietary driver wouldn't install properly). I bought the AMD card based on the myth that AMD has better linux support...

Try doing a PCI passthrough to virtualizes that 1080 onto a VM. Then try the same thing with AMD. Sure, the AMD drivers are relatively terrible, but at least they are open source and don't kick you in the knees when you are doing something you ought to be able to do.

Well this is not a supported feature even on Windows (AFAIK), so I'm not sure why this is an issue for you? See here for a list of supported graphics cards https://docs.nvidia.com/grid/latest/grid-vgpu-user-guide/ind.... Notice how there is no distinction between Windows and Linux. The 1080 is supported perfectly by nvidia-docker though.

Re: Nvidia to Acquire Mellanox for $6.9B

#222

Post by TheValeyard »

penagwin wrote: post #3

I only know of Mellanox for their 10 gigabit ethernet cards. Does anybody know if this is a good or bad thing for Mellanox?

I use Mellanox ConnectX-5 Dual QSFP+ 100Gbit Ethernet cards in my OpenStack private cloud at my business. Mellanox has been instrumental in running flash Ceph arrays by pushing the speed envelope beyond what the Intel/Cisco's of the world are doing.

Mellanox has also embraced bringing RDMA to things like Ceph and working with the broader vendor ecosystem like Red Hat for using this in production.

I hope Nvidia doesn't taint the good reputation of this company.

Re: Nvidia to Acquire Mellanox for $6.9B

#223

Post by dnautics »

orbifold wrote: post #221
dnautics wrote: post #118

Earlier quoted context omitted.

Try doing a PCI passthrough to virtualizes that 1080 onto a VM. Then try the same thing with AMD. Sure, the AMD drivers are relatively terrible, but at least they are open source and don't kick you in the knees when you are doing something you ought to be able to do.

Well this is not a supported feature even on Windows (AFAIK), so I'm not sure why this is an issue for you? See here for a list of supported graphics cards https://docs.nvidia.com/grid/latest/grid-vgpu-user-guide/ind... . Notice how there is no distinction between Windows and Linux. The 1080 is supported perfectly by nvidia-docker though.

There's no reason the hardware can't do it, the driver actively attempts to see if you're on a VM and refuses to talk to the hardware if so. Google "Nvidia error code 43" if you're curious, and some people like to virtualize a video game on their Linux dev PC.

Re: Nvidia to Acquire Mellanox for $6.9B

#224

Post by candiodari »

shaklee3 wrote: post #159
intorio wrote: post #127

Earlier quoted context omitted.

At least for a commodity chip like Broadcom Tomahawk (100G), the latency is 500ns with L3 enabled, and 300ns if only L2 is enabled. Compared to the Mellanox SB7700 at 90ns ethernet has some catching up to do if latency is the end goal. Ethernet tooling for HPC has a ways to go, but I suspect in the future it will be more competitive. Especially if specialty fabric vendors cut down on R&D. CLOS fabric designs seem to…

You're comparing apples and oranges. Mellanox has Ethernet switches with 300ns L3 latencies -- far lower than their broadcom counterparts. So it's not an Ethernet limitation, but a broadcom limitation.

Do keep in mind that having 300ns L3 latencies comes with it's own set of problems. Even at 10Gbit 300ns is not enough to get a packet through an electrical connection. Plus they also have some 90ns latency products.

That means that these switches, while fast, cannot check the packets for correctness (they don't have the full packet). That they will have "aborted" packets. That in some important ways these networks have the problems of the "half-duplex" networks of old.

Broadcom focuses on features for packet transmission. That means these Mellanox switches are pretty much restricted to situations where you want to have a set of servers on a single network segment and nothing else (not even an upstream connection). If that's exactly what you need, great. But mostly you're going to need more.

Re: Nvidia to Acquire Mellanox for $6.9B

#225

Post by wmf »

jabl wrote: post #215
wmf wrote: post #210

Earlier quoted context omitted.

You really don't need EVPN+VXLAN though. (And if you do need it I recommend finding a way to not need it.)

You mean you have separate subnets for each leaf switch, and then BGP or such for multipath routing between the leaf and spines? Sure, but what about subnet-level services like DHCP & PXE? Sounds cumbersome if you have to replicate that across all your leaf switches? Or maybe you could do one "provisioning and admin" VLAN that spans the entire cluster and which uses spanning tree, and then the high-performance RDMA s…

Yes, a subnet per rack is a best practice. Often people DHCP & PXE over the 1G out-of-band network which is dumb L2.

In most ASICs everything is the same latency since packets go through the whole pipeline whether they use all the functionality or not. Anyway, the latency of plain routing would have to be equal or faster than VXLAN encap + routing.

Re: Nvidia to Acquire Mellanox for $6.9B

#226

Post by finkin1 »

praneshp wrote: post #156
finkin1 wrote: post #129

Earlier quoted context omitted.

WhatsApp ~$22B acquisition by Facebook. LinkedIn ~$26B acquisition by Microsoft. Those are two that I can think of from memory.

Ah, I see. Not only do you have to have made solid technology, it should not be delivered through an app or website. I didn't realize those were the goalposts. ( https://engineering.linkedin.com/open-source as as starting point)

I'm not agreeing with OP's assumption that there isn't any tech in apps or websites. I'm just giving you examples so that you can stop wasting your time with the least charitable interpretation of OP's point and actually respond to what he might have meant.

Re: Nvidia to Acquire Mellanox for $6.9B

#227

Post by wmf »

candiodari wrote: post #224
shaklee3 wrote: post #159

Earlier quoted context omitted.

You're comparing apples and oranges. Mellanox has Ethernet switches with 300ns L3 latencies -- far lower than their broadcom counterparts. So it's not an Ethernet limitation, but a broadcom limitation.

Do keep in mind that having 300ns L3 latencies comes with it's own set of problems. Even at 10Gbit 300ns is not enough to get a packet through an electrical connection. Plus they also have some 90ns latency products. That means that these switches, while fast, cannot check the packets for correctness (they don't have the full packet). That they will have "aborted" packets. That in some important ways these networks h…

If you have so many CRC errors that cut-through bothers you, you might want to investigate why your cabling is so damaged.

Your information may be old; Mellanox has pretty much the same feature set as Broadcom now.

Re: Nvidia to Acquire Mellanox for $6.9B

#228

Post by magduf »

angry_octet wrote: post #195
bewo001 wrote: post #115

Earlier quoted context omitted.

While I agree about Intel's market force, they have a much better open source software story than nvidia. Nvidia is as closed as the other big player, Broadcom.

There is a lot of heavily patent/trade secret encumbered IP in both graphics and compute drivers, making open source extremely difficult. Above that layer I've found Nvidia extremely open. Intel is far worse to deal with, and additionally engages in anti-competitive architectural wars, preventing other vendors from interfacing with the CPU bus. As a result we have NVLink, and now we will soon have official IB cards w…

>There is a lot of heavily patent/trade secret encumbered IP in both graphics and compute drivers, making open source extremely difficult. Above that layer I've found Nvidia extremely open. >Intel is far worse to deal with

Huh? Nvidia is by far the worst option when it comes to GPUs if you want to run Linux. Both Intel and AMD manage to have excellent open source drivers, while Nvidia's is a proprietary mess that everyone complains about and doesn't work all that great with typical distro update mechanisms.

Re: Nvidia to Acquire Mellanox for $6.9B

#229

Post by sangnoir »

orbifold wrote: post #205
sangnoir wrote: post #142

Earlier quoted context omitted.

> Maybe just maybe this is not the fault of Nvidia but due to the fact that large parts of the Linux ecosystem are fragile, time consuming to configure and break if you look at them in the wrong way. How it works with open source drivers is that you main-line your drivers so that the kernel maintainer maintain the drivers for you, for free. Choosing to keep your drivers closed source means committing to keeping your…

I'm sorry but this is not how it works. No-one maintains your drivers for free. The contributors to drivers in the Linux kernel are usually employed by the companies that make the product. In addition they might have to deal with subsystem maintainers that treat their part of the code base as their personal fiefdom. Just follow the hoops that the AMD developers had to jump through so that their driver components got…

"Maintain" was too broad for what I meant: I should have said they'll future-proof your driver. Any future changes to the kernel won't break your mainlined driver - this is what you get for free, not defect fixes or other types of refactors. What you get for free is your driver working with all future kernel releases at the time of release; unlike Nvidia's drivers which lag kernel releases.
Post reply on HN